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/core/context.c

1749 lines
44 KiB
C
Raw Permalink Normal View History

refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
/*
* libwebsockets - small server side websockets and web server implementation
*
rtos diet: make basic auth support a config option on by default Remove wrapper content for < tls v1.2
2020-02-28 09:29:25 +00:00
* Copyright (C) 2010 - 2020 Andy Green <andy@warmcat.com>
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
*
license: switch LGPLv2.1+SLE parts to MIT
2019-08-14 10:44:14 +01:00
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
*
license: switch LGPLv2.1+SLE parts to MIT
2019-08-14 10:44:14 +01:00
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
*
license: switch LGPLv2.1+SLE parts to MIT
2019-08-14 10:44:14 +01:00
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
*/
private.h: rename to contain dir Having unique private header names is a requirement of a particular platform build system it's desirable to work with
2019-08-15 10:49:52 +01:00
#include "private-lib-core.h"
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
#ifndef LWS_BUILD_HASH
#define LWS_BUILD_HASH "unknown-build-hash"
#endif
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
static const char *library_version = LWS_LIBRARY_VERSION;
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
context: properly check for sys/resource.h Instead of hardcoding where sys/resource.h is available, check for it using cmake.
2020-11-01 11:24:31 +01:00
#if defined(LWS_HAVE_SYS_RESOURCE_H)
context: if linux and fd_limit_per_thread set try to configure by setrlimit
2020-05-06 07:41:16 +01:00
/* for setrlimit */
#include <sys/resource.h>
#endif
lws_validity: unified connection validity tracking Refactor everything around ping / pong handling in ws and h2, so there is instead a protocol-independent validity lws_sul tracking how long it has been since the last exchange that confirms the operation of the network connection in both directions. Clean out periodic role callback and replace the last two role users with discrete lws_sul for each pt.
2019-09-18 13:09:32 +01:00
#if defined(LWS_WITH_NETWORK)
/* in ms */
static uint32_t default_backoff_table[] = { 1000, 3000, 9000, 17000 };
#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
/**
* lws_get_library_version: get version and git hash library built from
*
* returns a const char * to a string like "1.1 178d78c"
* representing the library version followed by the git head hash it
* was built from
*/
cleaning
2020-01-02 08:32:23 +00:00
const char *
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
lws_get_library_version(void)
{
return library_version;
}
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
#if defined(LWS_WITH_STATS)
static void
lws_sul_stats_cb(lws_sorted_usec_list_t *sul)
{
struct lws_context_per_thread *pt = lws_container_of(sul,
struct lws_context_per_thread, sul_stats);
lws_stats_log_dump(pt->context);
sul: multiple timer domains Adapt the pt sul owner list to be an array, and define two different lists, one that acts like before and is the default for existing users, and another that has the ability to cooperate with systemwide suspend to restrict the interval spent suspended so that it will wake in time for the earliest thing on this wake-suspend sul list. Clean the api a bit and add lws_sul_cancel() that only needs the sul as the argument. Add a flag for client creation info to indicate that this client connection is important enough that, eg, validity checking it to detect silently dead connections should go on the wake-suspend sul list. That flag is exposed in secure streams policy so it can be added to a streamtype with "swake_validity": true Deprecate out the old vhost timer stuff that predates sul. Add a flag LWS_WITH_DEPRECATED_THINGS in cmake so users can get it back temporarily before it will be removed in a v4.2. Adapt all remaining in-tree users of it to use explicit suls.
2020-05-28 12:48:17 +01:00
__lws_sul_insert_us(&pt->pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&pt->sul_stats, 10 * LWS_US_PER_SEC);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
}
#endif
#if defined(LWS_WITH_PEER_LIMITS)
static void
lws_sul_peer_limits_cb(lws_sorted_usec_list_t *sul)
{
struct lws_context_per_thread *pt = lws_container_of(sul,
struct lws_context_per_thread, sul_peer_limits);
lws_peer_cull_peer_wait_list(pt->context);
sul: multiple timer domains Adapt the pt sul owner list to be an array, and define two different lists, one that acts like before and is the default for existing users, and another that has the ability to cooperate with systemwide suspend to restrict the interval spent suspended so that it will wake in time for the earliest thing on this wake-suspend sul list. Clean the api a bit and add lws_sul_cancel() that only needs the sul as the argument. Add a flag for client creation info to indicate that this client connection is important enough that, eg, validity checking it to detect silently dead connections should go on the wake-suspend sul list. That flag is exposed in secure streams policy so it can be added to a streamtype with "swake_validity": true Deprecate out the old vhost timer stuff that predates sul. Add a flag LWS_WITH_DEPRECATED_THINGS in cmake so users can get it back temporarily before it will be removed in a v4.2. Adapt all remaining in-tree users of it to use explicit suls.
2020-05-28 12:48:17 +01:00
__lws_sul_insert_us(&pt->pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&pt->sul_peer_limits, 10 * LWS_US_PER_SEC);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
}
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#if defined(LWS_WITH_NETWORK)
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
static const char * system_state_names[] = {
"undef",
"CONTEXT_CREATED",
"INITIALIZED",
"IFACE_COLDPLUG",
"DHCP",
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
"CPD_PRE_TIME",
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
"TIME_VALID",
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
"CPD_POST_TIME",
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
"POLICY_VALID",
"REGISTERED",
"AUTH1",
"AUTH2",
"OPERATIONAL",
"POLICY_INVALID"
};
context: option to disable system state management
2020-06-23 13:19:30 +01:00
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
/*
* Handle provoking protocol init when we pass through the right system state
*/
static int
lws_state_notify_protocol_init(struct lws_state_manager *mgr,
struct lws_state_notify_link *link, int current,
int target)
{
struct lws_context *context = lws_container_of(mgr, struct lws_context,
mgr_system);
coverity: 62313: static blob cant be null but coverity needs convincing
2020-08-18 11:37:54 +01:00
#if defined(LWS_WITH_SECURE_STREAMS) && defined(LWS_WITH_SECURE_STREAMS_SYS_AUTH_API_AMAZON_COM)
lws_system_blob_t *ab0, *ab1;
#endif
lws_system: helpers for attaching to existing event loop from other threads In the case code is composed into a single process, but it isn't monolithic in the sense it's made up of modular "applications" that are written separate projects, provide a way for the "applications" to request a callback from the lws event loop thread context safely. From the callback the applications can set up their operations on the lws event loop and drop their own thread. Since it requires system-specific locking to be threadsafe, provide a non-threadsafe helper and then indirect the actual usage through a user-defined lws_system ops function pointer that wraps the unsafe api with the system locking to make it safe.
2019-12-31 15:24:58 +00:00
int n;
/*
* Deal with any attachments that were waiting for the right state
* to come along
*/
for (n = 0; n < context->count_threads; n++)
lws_system_do_attach(&context->pt[n]);
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
if (target == LWS_SYSTATE_DHCP) {
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
/*
* Don't let it past here until at least one iface has been
* configured for operation with DHCP
*/
if (!lws_dhcpc_status(context, NULL))
return 1;
}
#endif
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
#if defined(LWS_WITH_SYS_NTPCLIENT)
if (target == LWS_SYSTATE_TIME_VALID &&
lws_now_secs() < 1594017754) /* 06:42 Mon Jul 6 2020 UTC */ {
lws_ntpc_trigger(context);
return 1;
}
#endif
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_WITH_SECURE_STREAMS) && defined(LWS_WITH_SECURE_STREAMS_SYS_AUTH_API_AMAZON_COM)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
/*
* Skip this if we are running something without the policy for it
ss auth: dont attempt auth if necessary blob unavailable
2020-05-01 14:27:19 -07:00
*
* If root token is empty, skip too.
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
*/
coverity: 62313: static blob cant be null but coverity needs convincing
2020-08-18 11:37:54 +01:00
ab0 = lws_system_get_blob(context, LWS_SYSBLOB_TYPE_AUTH, 0);
ab1 = lws_system_get_blob(context, LWS_SYSBLOB_TYPE_AUTH, 1);
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
if (target == LWS_SYSTATE_AUTH1 &&
coverity: 62313: static blob cant be null but coverity needs convincing
2020-08-18 11:37:54 +01:00
context->pss_policies && ab0 && ab1 &&
!lws_system_blob_get_size(ab0) &&
lws_system_blob_get_size(ab1)) {
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
lwsl_info("%s: AUTH1 state triggering api.amazon.com auth\n", __func__);
/*
* Start trying to acquire it if it's not already in progress
* returns nonzero if we determine it's not needed
*/
if (!lws_ss_sys_auth_api_amazon_com(context))
return 1;
}
#endif
#if defined(LWS_WITH_SECURE_STREAMS)
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
#if defined(LWS_WITH_DRIVERS)
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
/*
* See if we should do the SS Captive Portal Detection
*/
if (target == LWS_SYSTATE_CPD_PRE_TIME) {
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
if (lws_system_cpd_state_get(context) == LWS_CPD_INTERNET_OK)
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
return 0; /* allow it */
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
/*
* Don't allow it to move past here until we get an IP and
* CPD passes, driven by SMD
*/
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
return 1;
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
}
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
#endif
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#if !defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
/*
* Skip this if we are running something without the policy for it
*/
if (target == LWS_SYSTATE_POLICY_VALID &&
context->pss_policies && !context->policy_updated) {
/*
* Start trying to acquire it if it's not already in progress
* returns nonzero if we determine it's not needed
*/
if (!lws_ss_sys_fetch_policy(context))
return 1;
}
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#endif
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#endif
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
/* protocol part */
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
if (context->protocol_init_done)
return 0;
if (target != LWS_SYSTATE_POLICY_VALID)
return 0;
lwsl_info("%s: doing protocol init on POLICY_VALID\n", __func__);
coverity: 62211: use lws_protocol_init retcode in context init
2020-08-18 11:40:15 +01:00
return lws_protocol_init(context);
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
}
static void
lws_context_creation_completion_cb(lws_sorted_usec_list_t *sul)
{
struct lws_context *context = lws_container_of(sul, struct lws_context,
sul_system_state);
/* if nothing is there to intercept anything, go all the way */
lws_state_transition_steps(&context->mgr_system,
LWS_SYSTATE_OPERATIONAL);
}
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif /* WITH_SYS_STATE */
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
#if defined(LWS_WITH_SYS_SMD)
static int
lws_system_smd_cb(void *opaque, lws_smd_class_t _class, lws_usec_t timestamp,
void *buf, size_t len)
{
struct lws_context *cx = (struct lws_context *)opaque;
if (_class != LWSSMDCL_NETWORK)
return 0;
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
/* something external requested CPD check */
if (!lws_json_simple_strcmp(buf, len, "\"trigger\":", "cpdcheck"))
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
lws_system_cpd_start(cx);
lws_netdev: use lws_settings Perform the AP selection using lws_settings and a generic scan state machine
2020-07-02 15:38:53 +01:00
else
/*
* IP acquisition on any interface triggers captive portal
* check on default route
*/
if (!lws_json_simple_strcmp(buf, len, "\"type\":", "ipacq"))
lws_system_cpd_start(cx);
#if defined(LWS_WITH_SYS_NTPCLIENT)
/*
* Captive portal detect showing internet workable triggers NTP Client
*/
if (!lws_json_simple_strcmp(buf, len, "\"type\":", "cps") &&
!lws_json_simple_strcmp(buf, len, "\"result\":", "OK") &&
lws_now_secs() < 1594017754) /* 06:42 Mon Jul 6 2020 UTC */
lws_ntpc_trigger(cx);
#endif
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
/*
* Any network interface linkup triggers DHCP
*/
if (!lws_json_simple_strcmp(buf, len, "\"type\":", "linkup"))
lws_ntpc_trigger(cx);
#endif
#if defined(LWS_WITH_DRIVERS) && defined(LWS_WITH_NETWORK)
lws_netdev_smd_cb(opaque, _class, timestamp, buf, len);
#endif
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
return 0;
}
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#endif
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
#endif /* NETWORK */
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
#if !defined(LWS_WITH_NO_LOGS)
static const char * const opts_str =
#if defined(LWS_WITH_NETWORK)
"NET "
#else
"NoNET "
#endif
#if defined(LWS_WITH_CLIENT)
"CLI "
#endif
#if defined(LWS_WITH_SERVER)
"SRV "
#endif
#if defined(LWS_ROLE_H1)
"H1 "
#endif
#if defined(LWS_ROLE_H2)
"H2 "
#endif
#if defined(LWS_ROLE_WS)
"WS "
#endif
#if defined(LWS_ROLE_MQTT)
"MQTT "
#endif
#if defined(LWS_WITH_SECURE_STREAMS) && !defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
"SS-JSON-POL "
#endif
#if defined(LWS_WITH_SECURE_STREAMS) && defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
"SS-STATIC-POL "
#endif
#if defined(LWS_WITH_SECURE_STREAMS_PROXY_API)
"SSPROX "
#endif
#if defined(LWS_WITH_SYS_ASYNC_DNS)
"ASYNC_DNS "
#endif
#if defined(LWS_WITH_SYS_NTPCLIENT)
"NTPCLIENT "
#endif
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
"DHCP_CLIENT "
#endif
;
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#if defined(LWS_WITH_EVLIB_PLUGINS) && defined(LWS_WITH_EVENT_LIBS)
static const struct lws_evlib_map {
uint64_t flag;
const char *name;
} map[] = {
{ LWS_SERVER_OPTION_LIBUV, "evlib_uv" },
{ LWS_SERVER_OPTION_LIBEVENT, "evlib_event" },
{ LWS_SERVER_OPTION_GLIB, "evlib_glib" },
{ LWS_SERVER_OPTION_LIBEV, "evlib_ev" },
};
static const char * const dlist[] = {
LWS_INSTALL_LIBDIR,
NULL
};
#endif
cleaning
2020-01-02 08:32:23 +00:00
struct lws_context *
context info struct: make lws usage all const
2018-04-27 08:27:16 +08:00
lws_create_context(const struct lws_context_creation_info *info)
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
{
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_context *context = NULL;
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
#if !defined(LWS_WITH_NO_LOGS)
const char *s = "IPv6-absent";
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_FILE_OPS)
fops-zip
2017-03-03 12:38:10 +08:00
struct lws_plat_file_ops *prev;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#endif
coverity 155648 low dead code daemonize disabled Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-06 11:00:36 +08:00
#ifndef LWS_NO_DAEMONIZE
daemonize: use pid_t After report from Vitaly Shevtsov https://libwebsockets.org/pipermail/libwebsockets/2019-January/007787.html
2019-01-23 18:06:32 +08:00
pid_t pid_daemon = get_daemonize_pid();
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#endif
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
int count_threads = 1;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
uint8_t *u;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
android fix rlimit https://github.com/warmcat/libwebsockets/issues/488
2016-04-08 16:02:59 +08:00
#if defined(__ANDROID__)
struct rlimit rt;
#endif
ss: static policy: dynamic vhost instantiation Presently a vh is allocated per trust store at policy parsing-time, this is no problem on a linux-class device or if you decide you need a dynamic policy for functionality reasons. However if you're in a constrained enough situation that the static policy makes sense, in the case your trust stores do not have 100% duty cycle, ie, are anyway always in use, the currently-unused vhosts and their x.509 stack are sitting there taking up heap for no immediate benefit. This patch modifies behaviour in ..._STATIC_POLICY_ONLY so that vhosts and associated x.509 tls contexts are not instantiated until a secure stream using them is created; they are refcounted, and when the last logical secure stream using a vhost is destroyed, the vhost and its tls context is also destroyed. If another ss connection is created that wants to use the trust store, the vhost and x.509 context is regenerated again as needed. Currently the refcounting is by ss, it's also possible to move the refcounting to be by connection. The choice is between the delay to generate the vh being visisble at logical ss creation-time, or at connection-time. It's anyway not preferable to have ss instantiated and taking up space with no associated connection or connection attempt underway. NB you will need to reprocess any static policies after this patch so they conform to the trust_store changes.
2020-07-20 07:28:28 +01:00
size_t
#if defined(LWS_PLAT_FREERTOS)
/* smaller default, can set in info->pt_serv_buf_size */
s1 = 2048,
#else
s1 = 4096,
#endif
size = sizeof(struct lws_context);
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
int n, lpf = info->fd_limit_per_thread;
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
const lws_plugin_evlib_t *plev = NULL;
#if defined(LWS_WITH_EVLIB_PLUGINS) && defined(LWS_WITH_EVENT_LIBS)
struct lws_plugin *evlib_plugin_list = NULL;
#endif
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
#if defined(LWS_WITH_LIBUV)
char fatal_exit_defer = 0;
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
if (lpf) {
lws_system: ntpclient
2019-09-22 07:25:58 -07:00
lpf+= 2;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#if defined(LWS_WITH_SYS_ASYNC_DNS)
lpf++;
#endif
#if defined(LWS_WITH_SYS_NTPCLIENT)
lpf++;
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#endif
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
lpf++;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#endif
}
android fix rlimit https://github.com/warmcat/libwebsockets/issues/488
2016-04-08 16:02:59 +08:00
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
lwsl_notice("LWS: %s, loglevel %d\n", library_version, log_level);
refactor role ops This only refactors internal architecture and representations, the user api is unaffected.
2018-04-11 13:39:42 +08:00
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
#if defined(LWS_WITH_IPV6) && !defined(LWS_WITH_NO_LOGS)
LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT to default to runtime ssl disable https://github.com/warmcat/libwebsockets/issues/468 Adds lws_check_opt() to regularize multibit flag checking. There's a new context creation flag LWS_SERVER_OPTION_DO_SSL_GLOBAL_INIT, this is included automatically if you give any other SSL-related option flag. If you give no SSL-related option flag, nor this one directly, then even though SSL support may be compiled in, it is never initialized nor used for the whole lifetime of the lws context. Conversely in order to prepare the context to use SSL, even though, eg, you are not listening on SSL but will use SSL client connections later, you can give this flag explicitly to make sure SSL is initialized. Signed-off-by: Andy Green <andy@warmcat.com>
2016-03-23 09:22:11 +08:00
if (!lws_check_opt(info->options, LWS_SERVER_OPTION_DISABLE_IPV6))
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
s = "IPV6-on";
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
else
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
s = "IPV6-off";
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
#endif
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
#if defined(LWS_WITH_STATS)
logging: move most everything during startup to info
2017-09-19 17:30:06 +08:00
lwsl_info(" LWS_WITH_STATS : on\n");
LWS_WITH_STATS
2017-05-07 10:02:03 +08:00
#endif
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
lwsl_notice("%s%s\n", opts_str, s);
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
if (lws_plat_context_early_init())
return NULL;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_NETWORK)
if (info->count_threads)
count_threads = info->count_threads;
if (count_threads > LWS_MAX_SMP)
count_threads = LWS_MAX_SMP;
if (info->pt_serv_buf_size)
s1 = info->pt_serv_buf_size;
client: use block parse and buflist With http, the protocol doesn't indicate where the headers end and the next transaction or body begin. Until now, we handled that for client header response parsing by reading from the tls buffer bytewise. This modernizes the code to read in up to 256-byte chunks and parse the chunks in one hit (the parse API is already set up for doing this elsewhere). Now we have a generic input buflist, adapt the parser loop to go through that and arrange that any leftovers are placed on there.
2019-08-27 06:06:13 +01:00
/* pt fakewsi and the pt serv buf allocations ride after the context */
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
size += count_threads * s1;
#if !defined(LWS_PLAT_FREERTOS)
size += (count_threads * sizeof(struct lws));
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#endif /* network */
#if defined(LWS_WITH_POLL)
{
extern const lws_plugin_evlib_t evlib_poll;
plev = &evlib_poll;
}
#endif
#if defined(LWS_WITH_EVLIB_PLUGINS) && defined(LWS_WITH_EVENT_LIBS)
/*
* New style dynamically loaded event lib support
*
* We have to pick and load the event lib plugin before we allocate
* the context object, so we can overallocate it correctly
*/
lwsl_info("%s: ev lib path %s\n", __func__, LWS_INSTALL_LIBDIR);
for (n = 0; n < (int)LWS_ARRAY_SIZE(map); n++) {
if (!lws_check_opt(info->options, map[n].flag))
continue;
if (lws_plugins_init(&evlib_plugin_list,
dlist, "lws_evlib_plugin",
map[n].name, NULL, NULL)) {
lwsl_err("%s: failed to load %s\n", __func__,
map[n].name);
goto bail;
}
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
#if defined(LWS_WITH_LIBUV)
if (!n) /* libuv */
fatal_exit_defer = !!info->foreign_loops;
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
if (!evlib_plugin_list) {
lwsl_err("%s: unable to load evlib plugin %s\n",
__func__, map[n].name);
goto bail;
}
plev = (const lws_plugin_evlib_t *)evlib_plugin_list->hdr;
break;
}
#else
#if defined(LWS_WITH_EVENT_LIBS)
/*
* set the context event loops ops struct
*
* after this, all event_loop actions use the generic ops
*/
/*
* oldstyle built-in event lib support
*
* We have composed them into the libwebsockets lib itself, we can
* just pick the ops we want and done
*/
#if defined(LWS_WITH_LIBUV)
if (lws_check_opt(info->options, LWS_SERVER_OPTION_LIBUV)) {
extern const lws_plugin_evlib_t evlib_uv;
plev = &evlib_uv;
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
fatal_exit_defer = !!info->foreign_loops;
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
}
#endif
#if defined(LWS_WITH_LIBEVENT)
if (lws_check_opt(info->options, LWS_SERVER_OPTION_LIBEVENT)) {
extern const lws_plugin_evlib_t evlib_event;
plev = &evlib_event;
}
#endif
#if defined(LWS_WITH_GLIB)
if (lws_check_opt(info->options, LWS_SERVER_OPTION_GLIB)) {
extern const lws_plugin_evlib_t evlib_glib;
plev = &evlib_glib;
}
#endif
#if defined(LWS_WITH_LIBEV)
if (lws_check_opt(info->options, LWS_SERVER_OPTION_LIBEV)) {
extern const lws_plugin_evlib_t evlib_ev;
plev = &evlib_ev;
}
#endif
#endif /* with event libs */
#endif /* not with ev plugins */
if (!plev)
goto fail_event_libs;
#if defined(LWS_WITH_NETWORK)
coverity: plugins fixes
2020-08-30 09:30:36 +01:00
size += (size_t)plev->ops->evlib_size_ctx /* the ctx evlib priv */ +
(count_threads * (size_t)plev->ops->evlib_size_pt) /* the pt evlib priv */;
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
lwsl_info("Event loop: %s\n", plev->ops->name);
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#endif
context = lws_zalloc(size, "context");
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
if (!context) {
lwsl_err("No memory for websocket context\n");
return NULL;
}
generic-sessions update Generic sessions has been overdue some love to align it with the progress in the rest of lws. 1) Strict Content Security Policy 2) http2 compatibility 3) fixes and additions for use in a separate process via unix domain socket 4) work on ws and http proxying in lws 5) add minimal example
2019-04-05 21:13:59 +08:00
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#if defined(LWS_WITH_NETWORK)
context->event_loop_ops = plev->ops;
#endif
#if defined(LWS_WITH_EVENT_LIBS)
/* at the very end */
context->evlib_ctx = (uint8_t *)context + size -
plev->ops->evlib_size_ctx;
#endif
#if defined(LWS_WITH_EVLIB_PLUGINS) && defined(LWS_WITH_EVENT_LIBS)
context->evlib_plugin_list = evlib_plugin_list;
#endif
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
#if !defined(LWS_PLAT_FREERTOS)
generic-sessions update Generic sessions has been overdue some love to align it with the progress in the rest of lws. 1) Strict Content Security Policy 2) http2 compatibility 3) fixes and additions for use in a separate process via unix domain socket 4) work on ws and http proxying in lws 5) add minimal example
2019-04-05 21:13:59 +08:00
context->uid = info->uid;
context->gid = info->gid;
context->username = info->username;
context->groupname = info->groupname;
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
#endif
lws_system
2019-08-14 08:59:12 +01:00
context->system_ops = info->system_ops;
windows: clean type warnings There are some minor public api type improvements rather than cast everywhere inside lws and user code to work around them... these changed from int to size_t - lws_buflist_use_segment() return - lws_tokenize_t .len and .token_len - lws_tokenize_cstr() length - lws_get_peer_simple() namelen - lws_get_peer_simple_fd() namelen, int fd -> lws_sockfd_type fd - lws_write_numeric_address() len - lws_sa46_write_numeric_address() len These changes are typically a NOP for user code
2020-01-11 14:04:50 +00:00
context->pt_serv_buf_size = (unsigned int)s1;
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_WITH_UDP)
lws_retry: udp support
2019-09-08 08:08:55 +01:00
context->udp_loss_sim_tx_pc = info->udp_loss_sim_tx_pc;
context->udp_loss_sim_rx_pc = info->udp_loss_sim_rx_pc;
if (context->udp_loss_sim_tx_pc || context->udp_loss_sim_rx_pc)
lwsl_warn("%s: simulating udp loss tx: %d%%, rx: %d%%\n",
__func__, context->udp_loss_sim_tx_pc,
context->udp_loss_sim_rx_pc);
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#endif
lws_retry: udp support
2019-09-08 08:08:55 +01:00
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#if defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
/* directly use the user-provided policy object list */
context->pss_policies = info->pss_policies;
#endif
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#if defined(LWS_WITH_SECURE_STREAMS_PROXY_API)
context->ss_proxy_bind = info->ss_proxy_bind;
context->ss_proxy_port = info->ss_proxy_port;
context->ss_proxy_address = info->ss_proxy_address;
ctest: fixes-and-changes
2020-04-19 08:43:01 +01:00
if (context->ss_proxy_bind && context->ss_proxy_address)
lwsl_notice("%s: using ss proxy bind '%s', port %d, ads '%s'\n",
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
__func__, context->ss_proxy_bind, context->ss_proxy_port,
context->ss_proxy_address);
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_NETWORK)
context->count_threads = count_threads;
detailed latency stats Remove LWS_LATENCY. Add the option LWS_WITH_DETAILED_LATENCY, allowing lws to collect very detailed information on every read and write, and allow the user code to provide a callback to process events.
2019-08-23 16:10:36 +01:00
#if defined(LWS_WITH_DETAILED_LATENCY)
context->detailed_latency_cb = info->detailed_latency_cb;
context->detailed_latency_filepath = info->detailed_latency_filepath;
context->latencies_fd = -1;
#endif
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_ROLE_WS) && defined(LWS_WITHOUT_EXTENSIONS)
pmd: create_context: warn if info.extensions non-NULL if built without EXTENSIONS
2019-12-01 17:48:05 +00:00
if (info->extensions)
lwsl_warn("%s: LWS_WITHOUT_EXTENSIONS but extensions ptr set\n", __func__);
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#endif /* network */
generic-sessions update Generic sessions has been overdue some love to align it with the progress in the rest of lws. 1) Strict Content Security Policy 2) http2 compatibility 3) fixes and additions for use in a separate process via unix domain socket 4) work on ws and http proxying in lws 5) add minimal example
2019-04-05 21:13:59 +08:00
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#if defined(LWS_WITH_SECURE_STREAMS)
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#if !defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
context->pss_policies_json = info->pss_policies_json;
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#endif
ss: server: h1, h2, ws basic support Add initial support for defining servers using Secure Streams policy and api semantics. Serving h1, h2 and ws should be functional, the new minimal example shows a combined http + SS server with an incrementing ws message shown in the browser over tls, in around 200 lines of user code. NOP out anything to do with plugins, they're not currently used. Update the docs correspondingly.
2020-07-27 10:03:12 +01:00
#if defined(LWS_WITH_SSPLUGINS)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
context->pss_plugins = info->pss_plugins;
ss: server: h1, h2, ws basic support Add initial support for defining servers using Secure Streams policy and api semantics. Serving h1, h2 and ws should be functional, the new minimal example shows a combined http + SS server with an incrementing ws message shown in the browser over tls, in around 200 lines of user code. NOP out anything to do with plugins, they're not currently used. Update the docs correspondingly.
2020-07-27 10:03:12 +01:00
#endif
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#endif
generic-sessions update Generic sessions has been overdue some love to align it with the progress in the rest of lws. 1) Strict Content Security Policy 2) http2 compatibility 3) fixes and additions for use in a separate process via unix domain socket 4) work on ws and http proxying in lws 5) add minimal example
2019-04-05 21:13:59 +08:00
/* if he gave us names, set the uid / gid */
if (lws_plat_drop_app_privileges(context, 0))
goto bail;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_TLS) && defined(LWS_WITH_NETWORK)
refactor: also migrate tls to the ops struct and private.h pattern Several new ops are planned for tls... so better to bite the bullet and clean it out to the same level as roles + event-libs first. Also adds a new travis target "mbedtls" and all the tests except autobahn against mbedtls build.
2018-05-01 12:41:42 +08:00
#if defined(LWS_WITH_MBEDTLS)
context->tls_ops = &tls_ops_mbedtls;
#else
context->tls_ops = &tls_ops_openssl;
#endif
#endif
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 LWS_MAX_SMP > 1
lws_mutex_refcount This creates a "pthread mutex with a reference count" using gcc / clang atomic intrinsics + pthreads. Both pt and context locks are moved to use this, pt already had reference counting but it's new for context.
2018-06-27 07:15:39 +08:00
lws_mutex_refcount_init(&context->mr);
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
#endif
freertos: rename esp32 plat to freertos
2019-08-18 10:35:43 +01:00
#if defined(LWS_PLAT_FREERTOS)
private.h: rename to contain dir Having unique private header names is a requirement of a particular platform build system it's desirable to work with
2019-08-15 10:49:52 +01:00
#if defined(LWS_AMAZON_RTOS)
context->last_free_heap = xPortGetFreeHeapSize();
#else
esp32: report heap changes
2017-09-11 10:23:30 +08:00
context->last_free_heap = esp_get_free_heap_size();
private.h: rename to contain dir Having unique private header names is a requirement of a particular platform build system it's desirable to work with
2019-08-15 10:49:52 +01:00
#endif
esp32: report heap changes
2017-09-11 10:23:30 +08:00
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_FILE_OPS)
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
/* default to just the platform fops implementation */
esp32: align fops member name defines
2017-03-08 11:11:41 +08:00
context->fops_platform.LWS_FOP_OPEN = _lws_plat_file_open;
context->fops_platform.LWS_FOP_CLOSE = _lws_plat_file_close;
context->fops_platform.LWS_FOP_SEEK_CUR = _lws_plat_file_seek_cur;
context->fops_platform.LWS_FOP_READ = _lws_plat_file_read;
context->fops_platform.LWS_FOP_WRITE = _lws_plat_file_write;
fops-zip
2017-03-03 12:38:10 +08:00
context->fops_platform.fi[0].sig = NULL;
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
/*
* arrange a linear linked-list of fops starting from context->fops
*
* platform fops
* [ -> fops_zip (copied into context so .next settable) ]
* [ -> info->fops ]
*/
context->fops = &context->fops_platform;
prev = (struct lws_plat_file_ops *)context->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
fops-zip
2017-03-03 12:38:10 +08:00
#if defined(LWS_WITH_ZIP_FOPS)
/* make a soft copy so we can set .next */
context->fops_zip = fops_zip;
prev->next = &context->fops_zip;
prev = (struct lws_plat_file_ops *)prev->next;
#endif
/* if user provided fops, tack them on the end of the list */
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
if (info->fops)
fops-zip
2017-03-03 12:38:10 +08:00
prev->next = info->fops;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#endif
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
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_SERVER)
Add reject service keywords list
2016-10-13 06:32:57 +08:00
context->reject_service_keywords = info->reject_service_keywords;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#endif
context: external_baggage_free_on_destroy This adds a context creation-time member that points to something that should be freed when the context is destroyed. It's in preparation for context deprecation, when a context might be destroyed asynchronously... a related external with the lifetime of thee context should also be freed at that time. Adapt lwsws to use it with the context "strings" (also used for aligned structs created by the config) allocation.
2016-12-04 07:34:05 +08:00
if (info->external_baggage_free_on_destroy)
context->external_baggage_free_on_destroy =
info->external_baggage_free_on_destroy;
optee: remove build system
2019-01-15 06:59:48 +08:00
#if defined(LWS_WITH_NETWORK)
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
context->time_up = lws_now_usecs();
optee: remove build system
2019-01-15 06:59:48 +08:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
context->pcontext_finalize = info->pcontext;
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
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_WITH_TLS) && defined(LWS_WITH_NETWORK)
cleanups
2017-10-28 07:42:44 +08:00
context->simultaneous_ssl_restriction =
info->simultaneous_ssl_restriction;
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#endif
esp32: separate factory setup
2017-03-16 10:46:31 +08:00
unix plat: add minimal wsi fd map option An lws context usually contains a processwide fd -> wsi lookup table. This allows any possible fd returned by a *nix type OS to be immediately converted to a wsi just by indexing an array of struct lws * the size of the highest possible fd, as found by ulimit -n or similar. This works modestly for Linux type systems where the default ulimit -n for a process is 1024, it means a 4KB or 8KB lookup table for 32-bit or 64-bit systems. However in the case your lws usage is much simpler, like one outgoing client connection and no serving, this represents increasing waste. It's made much worse if the system has a much larger default ulimit -n, eg 1M, the table is occupying 4MB or 8MB, of which you will only use one. Even so, because lws can't be sure the OS won't return a socket fd at any number up to (ulimit -n - 1), it has to allocate the whole lookup table at the moment. This patch looks to see if the context creation info is setting info->fd_limit_per_thread... if it leaves it at the default 0, then everything is as it was before this patch. However if finds that (info->fd_limit_per_thread * actual_number_of_service_threads) where the default number of service threads is 1, is less than the fd limit set by ulimit -n, lws switches to a slower lookup table scheme, which only allocates the requested number of slots. Lookups happen then by iterating the table and comparing rather than indexing the array directly, which is obviously somewhat of a performance hit. However in the case where you know lws will only have a very few wsi maximum, this method can very usefully trade off speed to be able to avoid the allocation sized by ulimit -n. minimal examples for client that can make use of this are also modified by this patch to use the smaller context allocations.
2019-05-17 01:20:07 +01:00
context->options = info->options;
context: if linux and fd_limit_per_thread set try to configure by setrlimit
2020-05-06 07:41:16 +01:00
#if !defined(LWS_PLAT_FREERTOS) && !defined(LWS_PLAT_OPTEE) && !defined(WIN32)
/*
* If asked, try to set the rlimit / ulimit for process sockets / files.
* We read the effective limit in a moment, so we will find out the
* real limit according to system constraints then.
*/
if (info->rlimit_nofile) {
struct rlimit rl;
rl.rlim_cur = rl.rlim_max = info->rlimit_nofile;
setrlimit(RLIMIT_NOFILE, &rl);
}
#endif
coverity 155648 low dead code daemonize disabled Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-06 11:00:36 +08:00
#ifndef LWS_NO_DAEMONIZE
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 (pid_daemon) {
context->started_with_parent = pid_daemon;
daemonize: use pid_t After report from Vitaly Shevtsov https://libwebsockets.org/pipermail/libwebsockets/2019-January/007787.html
2019-01-23 18:06:32 +08:00
lwsl_info(" Started with daemon pid %u\n", (unsigned int)pid_daemon);
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
}
mbed3 plat Signed-off-by: Andy Green <andy.green@linaro.org>
2015-11-02 20:34:12 +08:00
#endif
android fix rlimit https://github.com/warmcat/libwebsockets/issues/488
2016-04-08 16:02:59 +08:00
#if defined(__ANDROID__)
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
n = getrlimit(RLIMIT_NOFILE, &rt);
if (n == -1) {
lwsl_err("Get RLIMIT_NOFILE failed!\n");
rtos: clean warnings where uint32_t is an unsigned long
2019-06-07 11:11:46 +01:00
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto free_context_fail;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
}
context->max_fds = rt.rlim_cur;
android fix rlimit https://github.com/warmcat/libwebsockets/issues/488
2016-04-08 16:02:59 +08:00
#else
esp32: shift to support latest esp-idf Esp-idf has an improved but still kind of abused cmake- based build system now. If we see ESP_PLATFORM coming as a cmake var, we can know we are being built from inside the esp-idf config system. Leave the existing esp32 arrangements alone but triggered off ESP_PLATFORM, adapt to use the cross toolchain file and various quirks automatically. In this way you can build lws a part of your project in a much cleaner way. Prepare a minimal esp32 test app for use in Sai Adapt .sai.json to build for esp32
2020-05-14 21:28:48 +01:00
#if defined(WIN32) || defined(_WIN32) || defined(LWS_AMAZON_RTOS) || defined(LWS_ESP_PLATFORM)
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
context->max_fds = getdtablesize();
rtos: clean warnings where uint32_t is an unsigned long
2019-06-07 11:11:46 +01:00
#else
osx: work around some big sur machines getting MAX_LONG ulimit -n
2020-11-16 17:10:57 +00:00
{
long l = sysconf(_SC_OPEN_MAX);
context->max_fds = 2560;
if (l > 10000000)
lwsl_warn("%s: unreasonable ulimit -n workaround\n",
__func__);
else
if (l != -1l)
context->max_fds = (int)l;
}
rtos: clean warnings where uint32_t is an unsigned long
2019-06-07 11:11:46 +01:00
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
if (context->max_fds < 0) {
lwsl_err("%s: problem getting process max files\n",
__func__);
coverity: fixes
2019-06-08 10:29:08 +01:00
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto free_context_fail;
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
}
cleaning
2020-01-02 08:32:23 +00:00
#endif
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
unix plat: add minimal wsi fd map option An lws context usually contains a processwide fd -> wsi lookup table. This allows any possible fd returned by a *nix type OS to be immediately converted to a wsi just by indexing an array of struct lws * the size of the highest possible fd, as found by ulimit -n or similar. This works modestly for Linux type systems where the default ulimit -n for a process is 1024, it means a 4KB or 8KB lookup table for 32-bit or 64-bit systems. However in the case your lws usage is much simpler, like one outgoing client connection and no serving, this represents increasing waste. It's made much worse if the system has a much larger default ulimit -n, eg 1M, the table is occupying 4MB or 8MB, of which you will only use one. Even so, because lws can't be sure the OS won't return a socket fd at any number up to (ulimit -n - 1), it has to allocate the whole lookup table at the moment. This patch looks to see if the context creation info is setting info->fd_limit_per_thread... if it leaves it at the default 0, then everything is as it was before this patch. However if finds that (info->fd_limit_per_thread * actual_number_of_service_threads) where the default number of service threads is 1, is less than the fd limit set by ulimit -n, lws switches to a slower lookup table scheme, which only allocates the requested number of slots. Lookups happen then by iterating the table and comparing rather than indexing the array directly, which is obviously somewhat of a performance hit. However in the case where you know lws will only have a very few wsi maximum, this method can very usefully trade off speed to be able to avoid the allocation sized by ulimit -n. minimal examples for client that can make use of this are also modified by this patch to use the smaller context allocations.
2019-05-17 01:20:07 +01:00
/*
* deal with any max_fds override, if it's reducing (setting it to
* more than ulimit -n is meaningless). The platform init will
* figure out what if this is something it can deal with.
*/
if (info->fd_limit_per_thread) {
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
int mf = lpf * context->count_threads;
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
unix plat: add minimal wsi fd map option An lws context usually contains a processwide fd -> wsi lookup table. This allows any possible fd returned by a *nix type OS to be immediately converted to a wsi just by indexing an array of struct lws * the size of the highest possible fd, as found by ulimit -n or similar. This works modestly for Linux type systems where the default ulimit -n for a process is 1024, it means a 4KB or 8KB lookup table for 32-bit or 64-bit systems. However in the case your lws usage is much simpler, like one outgoing client connection and no serving, this represents increasing waste. It's made much worse if the system has a much larger default ulimit -n, eg 1M, the table is occupying 4MB or 8MB, of which you will only use one. Even so, because lws can't be sure the OS won't return a socket fd at any number up to (ulimit -n - 1), it has to allocate the whole lookup table at the moment. This patch looks to see if the context creation info is setting info->fd_limit_per_thread... if it leaves it at the default 0, then everything is as it was before this patch. However if finds that (info->fd_limit_per_thread * actual_number_of_service_threads) where the default number of service threads is 1, is less than the fd limit set by ulimit -n, lws switches to a slower lookup table scheme, which only allocates the requested number of slots. Lookups happen then by iterating the table and comparing rather than indexing the array directly, which is obviously somewhat of a performance hit. However in the case where you know lws will only have a very few wsi maximum, this method can very usefully trade off speed to be able to avoid the allocation sized by ulimit -n. minimal examples for client that can make use of this are also modified by this patch to use the smaller context allocations.
2019-05-17 01:20:07 +01:00
if (mf < context->max_fds) {
context->max_fds_unrelated_to_ulimit = 1;
context->max_fds = mf;
}
}
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
content_info: make members conditional
2020-06-17 11:13:01 +01:00
context->token_limits = info->token_limits;
event-lib-glib Adds support for glib event library.
2020-02-07 11:39:32 +00:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_TLS) && defined(LWS_WITH_NETWORK)
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
time(&context->tls.last_cert_check_s);
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
if (info->alpn)
refactor: also migrate tls to the ops struct and private.h pattern Several new ops are planned for tls... so better to bite the bullet and clean it out to the same level as roles + event-libs first. Also adds a new travis target "mbedtls" and all the tests except autobahn against mbedtls build.
2018-05-01 12:41:42 +08:00
context->tls.alpn_default = info->alpn;
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
else {
refactor: also migrate tls to the ops struct and private.h pattern Several new ops are planned for tls... so better to bite the bullet and clean it out to the same level as roles + event-libs first. Also adds a new travis target "mbedtls" and all the tests except autobahn against mbedtls build.
2018-05-01 12:41:42 +08:00
char *p = context->tls.alpn_discovered, first = 1;
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
cmake: allow setting LWS_ROLE_WS You can build lws without support for ws, with -DLWS_ROLE_WS=0. This is thanks to the role ops isolating all the ws-specific business in the ws role. Also retire more test apps replaced by minmal-examples.
2018-04-25 08:42:18 +08:00
LWS_FOR_EVERY_AVAILABLE_ROLE_START(ar) {
if (ar->alpn) {
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
if (!first)
*p++ = ',';
refactor: also migrate tls to the ops struct and private.h pattern Several new ops are planned for tls... so better to bite the bullet and clean it out to the same level as roles + event-libs first. Also adds a new travis target "mbedtls" and all the tests except autobahn against mbedtls build.
2018-05-01 12:41:42 +08:00
p += lws_snprintf(p,
context->tls.alpn_discovered +
sizeof(context->tls.alpn_discovered) -
2 - p, "%s", ar->alpn);
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
first = 0;
}
cmake: allow setting LWS_ROLE_WS You can build lws without support for ws, with -DLWS_ROLE_WS=0. This is thanks to the role ops isolating all the ws-specific business in the ws role. Also retire more test apps replaced by minmal-examples.
2018-04-25 08:42:18 +08:00
} LWS_FOR_EVERY_AVAILABLE_ROLE_END;
refactor: also migrate tls to the ops struct and private.h pattern Several new ops are planned for tls... so better to bite the bullet and clean it out to the same level as roles + event-libs first. Also adds a new travis target "mbedtls" and all the tests except autobahn against mbedtls build.
2018-05-01 12:41:42 +08:00
context->tls.alpn_default = context->tls.alpn_discovered;
alpn: assemble defaults from roles and allow override Since new roles may be incompatible with http, add support for alpn names at the role struct, automatic generation of the default list of alpn names that servers advertise, and the ability to override the used alpn names per-vhost and per- client connection. This not only lets you modulate visibility or use of h2, but also enables vhosts that only offer non-http roles, as well as restricting http role vhosts to only alpn identifiers related to http roles.
2018-04-12 15:56:38 +08:00
}
#endif
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_WITH_NETWORK)
timeout settable from info This adds an info member that allows the user code to set the library's network action timeout in seconds. If left at the default 0, the build-time default AWAITING_TIMEOUT continues to be used. As suggested https://github.com/warmcat/libwebsockets/issues/427 Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 20:36:02 +08:00
if (info->timeout_secs)
context->timeout_secs = info->timeout_secs;
else
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#endif
client: allow setting overall connect timeout from vh init Add a member to the vh init struct allowing control of the overall connection wait introduced in an earlier patch. Set it to 20s by default. The timeout_secs member controls the individual DNS result connect timeout and is reduced to 5s by default.
2020-05-26 17:05:39 +01:00
context->timeout_secs = 5;
timeout settable from info This adds an info member that allows the user code to set the library's network action timeout in seconds. If left at the default 0, the build-time default AWAITING_TIMEOUT continues to be used. As suggested https://github.com/warmcat/libwebsockets/issues/427 Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 20:36:02 +08:00
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (info->max_http_header_data)
context->max_http_header_data = info->max_http_header_data;
else
add max_http_header_data2 and upgrade internal offsets from short to int https://github.com/warmcat/libwebsockets/issues/550 Signed-off-by: Andy Green <andy@warmcat.com>
2016-06-02 12:32:38 +08:00
if (info->max_http_header_data2)
context->max_http_header_data =
info->max_http_header_data2;
else
context->max_http_header_data = LWS_DEF_HEADER_LEN;
POST: handle http/1 pipelined after body Re-use wsi->preamble_rx to also hold leftover rx after dealing with POST body. Ensure ah->rx is always big enough to cope with what may have been read into the pt->serv_buf. Update the check for forced needed to also accept non-NULL wsi->preamble as well as ah->rxpos != ah->rxlen as indication forced needed. Disable autoservice on ah reset during transaction completed... it may close the wsi underneath us when it sees and processes the pending wsi->preamble_rx recursively otherwise.
2018-01-14 10:18:32 +08:00
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
if (info->max_http_header_pool)
context->max_http_header_pool = info->max_http_header_pool;
else
max_http_header_pool2: migrate context ah pool limit and count to unsigned int
2018-06-19 13:27:54 +08:00
if (info->max_http_header_pool2)
context->max_http_header_pool =
info->max_http_header_pool2;
else
context->max_http_header_pool = context->max_fds;
content_info: make members conditional
2020-06-17 11:13:01 +01:00
#endif
lws_spa: avoid reading past table Didn't check the lenghts of the NUL-terminated header table properly. https://github.com/warmcat/libwebsockets/issues/1859
2020-03-14 07:04:14 +00:00
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (info->fd_limit_per_thread)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
context->fd_limit_per_thread = lpf;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
else
lws_spa: avoid reading past table Didn't check the lenghts of the NUL-terminated header table properly. https://github.com/warmcat/libwebsockets/issues/1859
2020-03-14 07:04:14 +00:00
if (context->count_threads)
context->fd_limit_per_thread = context->max_fds /
context->count_threads;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
#if defined(LWS_WITH_SYS_SMD)
lws_mutex_init(context->smd.lock_messages);
lws_mutex_init(context->smd.lock_peers);
/* lws_system smd participant */
if (!lws_smd_register(context, context, 0, LWSSMDCL_NETWORK,
lws_system_smd_cb)) {
lwsl_err("%s: early smd register failed\n", __func__);
}
/* user smd participant */
if (info->early_smd_cb &&
!lws_smd_register(context, info->early_smd_opaque, 0,
info->early_smd_class_filter,
info->early_smd_cb)) {
lwsl_err("%s: early smd register failed\n", __func__);
}
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
n = 0;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
lws_validity: unified connection validity tracking Refactor everything around ping / pong handling in ws and h2, so there is instead a protocol-independent validity lws_sul tracking how long it has been since the last exchange that confirms the operation of the network connection in both directions. Clean out periodic role callback and replace the last two role users with discrete lws_sul for each pt.
2019-09-18 13:09:32 +01:00
context->default_retry.retry_ms_table = default_backoff_table;
context->default_retry.conceal_count =
context->default_retry.retry_ms_table_count =
LWS_ARRAY_SIZE(default_backoff_table);
context->default_retry.jitter_percent = 20;
context->default_retry.secs_since_valid_ping = 300;
context->default_retry.secs_since_valid_hangup = 310;
context: allow override validity timing
2019-11-20 10:17:36 +00:00
if (info->retry_and_idle_policy &&
info->retry_and_idle_policy->secs_since_valid_ping) {
context->default_retry.secs_since_valid_ping =
info->retry_and_idle_policy->secs_since_valid_ping;
context->default_retry.secs_since_valid_hangup =
info->retry_and_idle_policy->secs_since_valid_hangup;
}
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
/*
* Allocate the per-thread storage for scratchpad buffers,
* and header data pool
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
*/
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
u = (uint8_t *)&context[1];
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
for (n = 0; n < context->count_threads; n++) {
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
context->pt[n].serv_buf = u;
u += context->pt_serv_buf_size;
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
libuv create 1Hz background timeout check Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-27 11:03:27 +08:00
context->pt[n].context = context;
context->pt[n].tid = n;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
#if !defined(LWS_PLAT_FREERTOS)
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
/*
* We overallocated for a fakewsi (can't compose it in the
* pt because size isn't known at that time). point to it
* and zero it down. Fakewsis are needed to make callbacks work
* when the source of the callback is not actually from a wsi
* context.
*/
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
context->pt[n].fake_wsi = (struct lws *)u;
u += sizeof(struct lws);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
memset(context->pt[n].fake_wsi, 0, sizeof(struct lws));
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
#endif
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#if defined(LWS_WITH_EVENT_LIBS)
context->pt[n].evlib_pt = u;
u += plev->ops->evlib_size_pt;
#endif
refactor: most preparation for -DLWS_ROLE_H1=0
2018-04-27 19:16:50 +08:00
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
refactor: move all ah and http specific vars to http private and conditional inclusion from there
2018-04-27 15:20:56 +08:00
context->pt[n].http.ah_list = NULL;
context->pt[n].http.ah_pool_length = 0;
refactor: most preparation for -DLWS_ROLE_H1=0
2018-04-27 19:16:50 +08:00
#endif
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
lws_pt_mutex_init(&context->pt[n]);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
#if defined(LWS_WITH_SEQUENCER)
lws_seq_pt_init(&context->pt[n]);
#endif
lws_validity: unified connection validity tracking Refactor everything around ping / pong handling in ws and h2, so there is instead a protocol-independent validity lws_sul tracking how long it has been since the last exchange that confirms the operation of the network connection in both directions. Clean out periodic role callback and replace the last two role users with discrete lws_sul for each pt.
2019-09-18 13:09:32 +01:00
LWS_FOR_EVERY_AVAILABLE_ROLE_START(ar) {
if (ar->pt_init_destroy)
ar->pt_init_destroy(context, info,
&context->pt[n], 0);
} LWS_FOR_EVERY_AVAILABLE_ROLE_END;
cgi: call init_destroy even though not on available roles list cgi is a role, but it's not an externally-selectable role... it's only used when the mount indicates it should be. That means it's not in the list of EVERY_AVAILABLE_ROLE... make sure its zombie child reaping timer is started at context creation and stopped at destruction.
2019-10-20 07:12:51 +01:00
#if defined(LWS_WITH_CGI)
role_ops_cgi.pt_init_destroy(context, info, &context->pt[n], 0);
#endif
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
}
complain if zero ka_interval used with positive ka_time https://github.com/warmcat/libwebsockets/issues/308 Signed-off-by: Andy Green <andy.green@linaro.org>
2015-10-16 10:54:04 +08:00
if (!info->ka_interval && info->ka_time > 0) {
lwsl_err("info->ka_interval can't be 0 if ka_time used\n");
return NULL;
}
whitespace trailing mass cleanout Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-14 08:52:03 +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)
/* scale the peer hash table according to the max fds for the process,
* so that the max list depth averages 16. Eg, 1024 fd -> 64,
* 102400 fd -> 6400
*/
unix socket: fixes and improvements Auto-remove any unix socket file already there. Correctly identify if it's in use per-vhost. Make the peer-limits stuff ignore it.
2018-08-02 19:13:53 +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
context->pl_hash_elements =
(context->count_threads * context->fd_limit_per_thread) / 16;
context->pl_hash_table = lws_zalloc(sizeof(struct lws_peer *) *
heap instrumentation
2017-10-04 07:10:39 +08:00
context->pl_hash_elements, "peer limits hash table");
unix socket: fixes and improvements Auto-remove any unix socket file already there. Correctly identify if it's in use per-vhost. Make the peer-limits stuff ignore it.
2018-08-02 19:13:53 +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
context->ip_limit_ah = info->ip_limit_ah;
context->ip_limit_wsi = info->ip_limit_wsi;
PEER_LIMITS: modernize to sa46 and add notification cb There are a bunch of sa46-aware lws apis for formatting, etc sa46 in ./include/libwebsockets/lws-network-helper.h
2020-05-24 19:44:46 +01:00
context->pl_notify_cb = info->pl_notify_cb;
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
#endif
unix plat: add minimal wsi fd map option An lws context usually contains a processwide fd -> wsi lookup table. This allows any possible fd returned by a *nix type OS to be immediately converted to a wsi just by indexing an array of struct lws * the size of the highest possible fd, as found by ulimit -n or similar. This works modestly for Linux type systems where the default ulimit -n for a process is 1024, it means a 4KB or 8KB lookup table for 32-bit or 64-bit systems. However in the case your lws usage is much simpler, like one outgoing client connection and no serving, this represents increasing waste. It's made much worse if the system has a much larger default ulimit -n, eg 1M, the table is occupying 4MB or 8MB, of which you will only use one. Even so, because lws can't be sure the OS won't return a socket fd at any number up to (ulimit -n - 1), it has to allocate the whole lookup table at the moment. This patch looks to see if the context creation info is setting info->fd_limit_per_thread... if it leaves it at the default 0, then everything is as it was before this patch. However if finds that (info->fd_limit_per_thread * actual_number_of_service_threads) where the default number of service threads is 1, is less than the fd limit set by ulimit -n, lws switches to a slower lookup table scheme, which only allocates the requested number of slots. Lookups happen then by iterating the table and comparing rather than indexing the array directly, which is obviously somewhat of a performance hit. However in the case where you know lws will only have a very few wsi maximum, this method can very usefully trade off speed to be able to avoid the allocation sized by ulimit -n. minimal examples for client that can make use of this are also modified by this patch to use the smaller context allocations.
2019-05-17 01:20:07 +01:00
/*
* fds table contains pollfd structs for as many pollfds as we can
* handle... spread across as many service threads as we have going
*/
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
n = sizeof(struct lws_pollfd) * context->count_threads *
context->fd_limit_per_thread;
heap instrumentation
2017-10-04 07:10:39 +08:00
context->pt[0].fds = lws_zalloc(n, "fds table");
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
if (context->pt[0].fds == NULL) {
cleanups after api changes and mbed update Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-04 16:54:12 +08:00
lwsl_err("OOM allocating %d fds\n", context->max_fds);
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
goto bail;
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
}
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
context: focus context init logging Tighten up the logging at info and have a build summary and version info at notice level like this [2020/07/19 07:01:07:5563] N: LWS: 4.0.99-v4.0.0-232-gd602af468, loglevel 1031 [2020/07/19 07:01:07:5567] N: NET IPv6-absent H1 H2 WS MQTT SS-JSON-POL SSPROX ASYNC_DNS
2020-07-19 07:05:06 +01:00
lwsl_info(" ctx: %5luB (%ld ctx + pt(%ld thr x %d)), "
"pt-fds: %d, fdmap: %d\n",
(long)sizeof(struct lws_context) +
(context->count_threads * context->pt_serv_buf_size),
(long)sizeof(struct lws_context),
(long)context->count_threads,
context->pt_serv_buf_size,
context->fd_limit_per_thread, n);
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
lwsl_info(" http: ah_data: %u, ah: %lu, max count %u\n",
context->max_http_header_data,
(long)sizeof(struct allocated_headers),
context->max_http_header_pool);
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_SERVER)
context settable server string Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 13:33:52 +08:00
if (info->server_string) {
context->server_string = info->server_string;
clean build warning on windows Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-04 11:11:15 +08:00
context->server_string_len = (short)
strlen(context->server_string);
context settable server string Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 13:33:52 +08:00
}
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#endif
context settable server string Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-15 13:33:52 +08:00
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
#if LWS_MAX_SMP > 1
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
/* each thread serves his own chunk of fds */
ws: fix coredump of lws_create_context Signed-off-by: t00416110 <tanyifeng1@huawei.com>
2019-01-22 15:45:34 +08:00
for (n = 1; n < (int)context->count_threads; n++)
multithread stability Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-26 20:56:56 +08:00
context->pt[n].fds = context->pt[n - 1].fds +
context->fd_limit_per_thread;
#endif
http header malloc pool implement pool Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-25 12:44:12 +08:00
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
/*
* Past here, we may have added handles to the event lib
* loop and if libuv, have to take care about how to unpick them...
*/
plat combine inits into single lws_plat_init and provide info Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-10 07:14:16 +08:00
if (lws_plat_init(context, info))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
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
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->event_loop_ops->init_context)
if (context->event_loop_ops->init_context(context, info))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->event_loop_ops->init_pt)
for (n = 0; n < context->count_threads; n++) {
void *lp = NULL;
if (info->foreign_loops)
lp = info->foreign_loops[n];
if (context->event_loop_ops->init_pt(context, lp, n))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
}
if (lws_create_event_pipes(context))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +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
lws_context_init_ssl_library(info);
cgi post Signed-off-by: Andy Green <andy@warmcat.com>
2016-04-13 11:42:53 +08:00
context->user_space = info->user;
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
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#if defined(LWS_WITH_SERVER)
http header malloc pool implement pool Signed-off-by: Andy Green <andy.green@linaro.org>
2015-12-25 12:44:12 +08:00
strcpy(context->canonical_hostname, "unknown");
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
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
lws_server_get_canonical_hostname(context, info);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
debloat: remove things from being built by default that should be conditional
2019-08-18 06:29:34 +01:00
#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
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
#if defined(LWS_WITH_STATS)
context->pt[0].sul_stats.cb = lws_sul_stats_cb;
sul: multiple timer domains Adapt the pt sul owner list to be an array, and define two different lists, one that acts like before and is the default for existing users, and another that has the ability to cooperate with systemwide suspend to restrict the interval spent suspended so that it will wake in time for the earliest thing on this wake-suspend sul list. Clean the api a bit and add lws_sul_cancel() that only needs the sul as the argument. Add a flag for client creation info to indicate that this client connection is important enough that, eg, validity checking it to detect silently dead connections should go on the wake-suspend sul list. That flag is exposed in secure streams policy so it can be added to a streamtype with "swake_validity": true Deprecate out the old vhost timer stuff that predates sul. Add a flag LWS_WITH_DEPRECATED_THINGS in cmake so users can get it back temporarily before it will be removed in a v4.2. Adapt all remaining in-tree users of it to use explicit suls.
2020-05-28 12:48:17 +01:00
__lws_sul_insert_us(&context->pt[0].pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&context->pt[0].sul_stats, 10 * LWS_US_PER_SEC);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
#endif
#if defined(LWS_WITH_PEER_LIMITS)
context->pt[0].sul_peer_limits.cb = lws_sul_peer_limits_cb;
sul: multiple timer domains Adapt the pt sul owner list to be an array, and define two different lists, one that acts like before and is the default for existing users, and another that has the ability to cooperate with systemwide suspend to restrict the interval spent suspended so that it will wake in time for the earliest thing on this wake-suspend sul list. Clean the api a bit and add lws_sul_cancel() that only needs the sul as the argument. Add a flag for client creation info to indicate that this client connection is important enough that, eg, validity checking it to detect silently dead connections should go on the wake-suspend sul list. That flag is exposed in secure streams policy so it can be added to a streamtype with "swake_validity": true Deprecate out the old vhost timer stuff that predates sul. Add a flag LWS_WITH_DEPRECATED_THINGS in cmake so users can get it back temporarily before it will be removed in a v4.2. Adapt all remaining in-tree users of it to use explicit suls.
2020-05-28 12:48:17 +01:00
__lws_sul_insert_us(&context->pt[0].pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&context->pt[0].sul_peer_limits, 10 * LWS_US_PER_SEC);
sul: all timed objects use a single pt sul list wsi timeout, wsi hrtimer, sequencer timeout and vh-protocol timer all now participate on a single sorted us list. The whole idea of polling wakes is thrown out, poll waits ignore the timeout field and always use infinite timeouts. Introduce a public api that can schedule its own callback from the event loop with us resolution (usually ms is all the platform can do). Upgrade timeouts and sequencer timeouts to also be able to use us resolution. Introduce a prepared fakewsi in the pt, so we don't have to allocate one on the heap when we need it. Directly handle vh-protocol timer if LWS_MAX_SMP == 1
2019-08-09 10:12:09 +01:00
#endif
capabilities support
2017-06-07 06:10:02 +08:00
#if defined(LWS_HAVE_SYS_CAPABILITY_H) && defined(LWS_HAVE_LIBCAP)
memcpy(context->caps, info->caps, sizeof(context->caps));
context->count_caps = info->count_caps;
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#if defined(LWS_WITH_NETWORK)
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#if defined(LWS_WITH_SYS_ASYNC_DNS) || defined(LWS_WITH_SYS_NTPCLIENT) || \
defined(LWS_WITH_SYS_DHCP_CLIENT)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
{
/*
* system vhost
*/
struct lws_context_creation_info ii;
cleaning
2020-01-02 08:32:23 +00:00
const struct lws_protocols *pp[4];
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
struct lws_vhost *vh;
#if defined(LWS_WITH_SYS_ASYNC_DNS)
extern const struct lws_protocols lws_async_dns_protocol;
#endif
#if defined(LWS_WITH_SYS_NTPCLIENT)
extern const struct lws_protocols lws_system_protocol_ntpc;
#endif
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
extern const struct lws_protocols lws_system_protocol_dhcpc;
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
n = 0;
#if defined(LWS_WITH_SYS_ASYNC_DNS)
pp[n++] = &lws_async_dns_protocol;
#endif
#if defined(LWS_WITH_SYS_NTPCLIENT)
pp[n++] = &lws_system_protocol_ntpc;
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#endif
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
pp[n++] = &lws_system_protocol_dhcpc;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#endif
pp[n] = NULL;
memset(&ii, 0, sizeof(ii));
ii.vhost_name = "system";
ii.pprotocols = pp;
vh = lws_create_vhost(context, &ii);
if (!vh) {
lwsl_err("%s: failed to create system vhost\n",
__func__);
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
}
context->vhost_system = vh;
if (lws_protocol_init_vhost(vh, NULL)) {
lwsl_err("%s: failed to init system vhost\n", __func__);
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
}
#if defined(LWS_WITH_SYS_ASYNC_DNS)
if (lws_async_dns_init(context))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#endif
}
#endif
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
/*
* init the lws_state mgr for the system state
*/
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
context->mgr_system.state_names = system_state_names;
context->mgr_system.name = "system";
context->mgr_system.state = LWS_SYSTATE_CONTEXT_CREATED;
context->mgr_system.parent = context;
context->mgr_system.context = context;
#if defined(LWS_WITH_SYS_SMD)
context->mgr_system.smd_class = LWSSMDCL_SYSTEM_STATE;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
#endif
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
context->protocols_notify.name = "prot_init";
context->protocols_notify.notify_cb = lws_state_notify_protocol_init;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
lws_state_reg_notifier(&context->mgr_system, &context->protocols_notify);
/*
* insert user notifiers here so they can participate with vetoing us
* trying to jump straight to operational, or at least observe us
* reaching 'operational', before we returned from context creation.
*/
lws_state_reg_notifier_list(&context->mgr_system,
info->register_notifier_list);
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
/*
* if he's not saying he'll make his own vhosts later then act
* compatibly and make a default vhost using the data in the info
*/
if (!lws_check_opt(info->options, LWS_SERVER_OPTION_EXPLICIT_VHOSTS))
if (!lws_create_vhost(context, info)) {
lwsl_err("Failed to create default vhost\n");
#if defined(LWS_WITH_PEER_LIMITS)
lws_free_set_NULL(context->pl_hash_table);
#endif
lws_context_create: clean up cancel pipes on fail https://github.com/warmcat/libwebsockets/issues/1765
2019-11-16 16:56:27 +00:00
goto fail_clean_pipes;
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
}
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#if defined(LWS_WITH_SECURE_STREAMS)
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#if !defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
if (context->pss_policies_json) {
/*
* You must create your context with the explicit vhosts flag
* in order to use secure streams
*/
assert(lws_check_opt(info->options,
LWS_SERVER_OPTION_EXPLICIT_VHOSTS));
ss: allow streamtype policy overlays Make the policy load apis public with an extra argument that says if you want the JSON to overlay on an existing policy rather than replace it. Teach the stream type parser stuff to realize it already has an entry for the stream type and to modify that rather than create a second one, allowing overlays to modify stream types. Add --force-portal and --force-no-internet flags to minimal-secure-streams and use the new policy overlay stuff to force the policy for captive portal detection to feel that there is one or that there's no internet.
2020-03-12 14:43:32 +00:00
if (lws_ss_policy_parse_begin(context, 0))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
n = lws_ss_policy_parse(context,
(uint8_t *)context->pss_policies_json,
strlen(context->pss_policies_json));
if (n != LEJP_CONTINUE && n < 0)
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
if (lws_ss_policy_set(context, "hardcoded")) {
lwsl_err("%s: policy set failed\n", __func__);
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
}
} else
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#else
if (context->pss_policies) {
/* user code set the policy objects directly, no parsing step */
if (lws_ss_policy_set(context, "hardcoded")) {
lwsl_err("%s: policy set failed\n", __func__);
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
}
context creation: remove chance to create vhost twice
2020-04-16 14:17:15 +01:00
} //else
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#endif
context creation: remove chance to create vhost twice
2020-04-16 14:17:15 +01:00
// lws_create_vhost(context, info);
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
lws_context_init_extensions(info, context);
lwsl_info(" mem: per-conn: %5lu bytes + protocol rx buf\n",
(unsigned long)sizeof(struct lws));
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
/*
* drop any root privs for this process
* to listen on port < 1023 we would have needed root, but now we are
* listening, we don't want the power for anything 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
if (!lws_check_opt(info->options, LWS_SERVER_OPTION_EXPLICIT_VHOSTS))
permissions: adapt drop permissions plat function to do uid and gid lookup separately
2019-04-21 19:51:03 +01:00
if (lws_plat_drop_app_privileges(context, 1))
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
goto bail_libuv_aware;
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
/*
* We want to move on the syste, state as far as it can go towards
* OPERATIONAL now. But we have to return from here first so the user
* code that called us can set its copy of context, which it may be
* relying on to perform operations triggered by the state change.
*
* We set up a sul to come back immediately and do the state change.
*/
lws_sul_schedule(context, 0, &context->sul_system_state,
lws_context_creation_completion_cb, 1);
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
autobahn fixes This replaces the old test-app for echo with separate client and server minimal versions. The autobahn test script is made more autonomous and tests both client and server.
2018-04-20 10:33:23 +08:00
/* expedite post-context init (eg, protocols) */
lws_cancel_service(context);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
autobahn fixes This replaces the old test-app for echo with separate client and server minimal versions. The autobahn test script is made more autonomous and tests both client and server.
2018-04-20 10:33:23 +08:00
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
return context;
lws_context_create: clean up cancel pipes on fail https://github.com/warmcat/libwebsockets/issues/1765
2019-11-16 16:56:27 +00:00
#if defined(LWS_WITH_NETWORK)
fail_clean_pipes:
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
#if defined(LWS_WITH_LIBUV)
if (fatal_exit_defer) {
lws_context_destroy(context);
return context;
}
#endif
lws_context_create: clean up cancel pipes on fail https://github.com/warmcat/libwebsockets/issues/1765
2019-11-16 16:56:27 +00:00
for (n = 0; n < context->count_threads; n++)
lws_destroy_event_pipe(context->pt[n].pipe_wsi);
lws_free_set_NULL(context->pt[0].fds);
lws_plat_context_late_destroy(context);
lws_free_set_NULL(context);
return NULL;
#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
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_context_destroy(context);
valgrind: context creation error path
2018-03-22 14:54:25 +08:00
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
return NULL;
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
bail_libuv_aware:
lws_context_destroy(context);
#if defined(LWS_WITH_LIBUV)
return fatal_exit_defer ? context : NULL;
#else
return NULL;
#endif
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
fail_event_libs:
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
lwsl_err("Requested event library support not configured\n");
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
libuv: context creation fails via destroy When using a foreign libuv loop, context creation may fail after adding handles to the foreign loop... if so, it can no longer deal with the fatal error by unpicking the created context and returning NULL... it has to brazen it out with a half-baked context that has already started the destroy flow and allow the foreign loop to close out the handles the usual way for libuv. https://github.com/warmcat/libwebsockets/issues/2129
2020-11-28 05:41:26 +00:00
free_context_fail:
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
lws_free(context);
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
return NULL;
}
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
#if defined(LWS_WITH_NETWORK)
int
lws_system_cpd_start(struct lws_context *cx)
{
cx->captive_portal_detect = LWS_CPD_UNKNOWN;
/* if there's a platform implementation, use it */
if (lws_system_get_ops(cx) &&
lws_system_get_ops(cx)->captive_portal_detect_request)
return lws_system_get_ops(cx)->captive_portal_detect_request(cx);
#if defined(LWS_WITH_SECURE_STREAMS)
/*
* Otherwise try to use SS "captive_portal_detect" if that's enabled
*/
return lws_ss_sys_cpd(cx);
#else
return 0;
#endif
}
logs: prioritize WITH_NO_LOGS even in release Really not having any logs makes it difficult to know what is really happening, but if that's you're thing this will align debug and release modes to just have ERR and USER if you give WITH_NO_LOGS
2021-03-15 12:12:16 +00:00
#if (defined(LWS_WITH_SYS_STATE) && defined(LWS_WITH_SYS_SMD)) || !defined(LWS_WITH_NO_LOGS)
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
static const char *cname[] = { "Unknown", "OK", "Captive", "No internet" };
logs: prioritize WITH_NO_LOGS even in release Really not having any logs makes it difficult to know what is really happening, but if that's you're thing this will align debug and release modes to just have ERR and USER if you give WITH_NO_LOGS
2021-03-15 12:12:16 +00:00
#endif
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
void
lws_system_cpd_set(struct lws_context *cx, lws_cpd_result_t result)
{
if (cx->captive_portal_detect != LWS_CPD_UNKNOWN)
return;
logs: prioritize WITH_NO_LOGS even in release Really not having any logs makes it difficult to know what is really happening, but if that's you're thing this will align debug and release modes to just have ERR and USER if you give WITH_NO_LOGS
2021-03-15 12:12:16 +00:00
#if !defined(LWS_WITH_NO_LOGS)
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
lwsl_notice("%s: setting CPD result %s\n", __func__, cname[result]);
logs: prioritize WITH_NO_LOGS even in release Really not having any logs makes it difficult to know what is really happening, but if that's you're thing this will align debug and release modes to just have ERR and USER if you give WITH_NO_LOGS
2021-03-15 12:12:16 +00:00
#endif
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
cx->captive_portal_detect = (uint8_t)result;
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
#if defined(LWS_WITH_SYS_SMD)
lws_smd_msg_printf(cx, LWSSMDCL_NETWORK,
"{\"type\":\"cpd\",\"result\":\"%s\"}",
cname[cx->captive_portal_detect]);
#endif
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
/* if nothing is there to intercept anything, go all the way */
ss: allow url style endpoint addresses The endpoint field in streamtype policy may continue to just be the hostname, like "warmcat.com". But it's also possible now to be a url-formatted string, like, eg, "https://warmcat.com:444/mailman/listinfo" If so (ie, if it contains a : ) then the decoded elements may override if tls is enabled, the endpoint address, the port, and the url path. No ABI change.
2020-03-14 06:56:41 +00:00
if (cx->mgr_system.state != LWS_SYSTATE_POLICY_INVALID)
lws_state_transition_steps(&cx->mgr_system,
LWS_SYSTATE_OPERATIONAL);
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
}
lws_cpd_result_t
lws_system_cpd_state_get(struct lws_context *cx)
{
return (lws_cpd_result_t)cx->captive_portal_detect;
}
#endif
cleaning
2020-01-02 08:32:23 +00:00
int
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
lws_context_is_deprecated(struct lws_context *cx)
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
{
captive portal Implement Captive Portal detection support in lws, with the actual detection happening in platform code hooked up by lws_system_ops_t. Add an implementation using Secure Streams as well, if the policy defines captive_portal_detect streamtype, a SS using that streamtype is used to probe if it's behind a captive portal.
2020-03-11 12:44:01 +00:00
return cx->deprecated;
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
}
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
/*
* When using an event loop, the context destruction is in three separate
* parts. This is to cover both internal and foreign event loops cleanly.
*
* - lws_context_destroy() simply starts a soft close of all wsi and
* related allocations. The event loop continues.
*
* As the closes complete in the event loop, reference counting is used
* to determine when everything is closed. It then calls
* lws_context_destroy2().
*
* - lws_context_destroy2() cleans up the rest of the higher-level logical
* lws pieces like vhosts. If the loop was foreign, it then proceeds to
* lws_context_destroy3(). If it the loop is internal, it stops the
* internal loops and waits for lws_context_destroy() to be called again
* outside the event loop (since we cannot destroy the loop from
* within the loop). That will cause lws_context_destroy3() to run
* directly.
*
* - lws_context_destroy3() destroys any internal event loops and then
* destroys the context itself, setting what was info.pcontext to NULL.
*/
context_destroy: figure out if anything still in event loop
2018-05-02 19:27:29 +08:00
/*
* destroy the actual context itself
*/
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
static void
lws_context_destroy3(struct lws_context *context)
{
struct lws_context **pcontext_finalize = context->pcontext_finalize;
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
int n;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#if defined(LWS_WITH_NETWORK)
event libs: context_destroy sequencing
2020-03-10 19:13:43 +00:00
context->finalize_destroy_after_internal_loops_stopped = 1;
if (context->event_loop_ops->destroy_context2)
context->event_loop_ops->destroy_context2(context);
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
for (n = 0; n < context->count_threads; n++) {
codacy: minor fixes
2018-10-10 13:54:43 +08:00
struct lws_context_per_thread *pt = &context->pt[n];
seq: LWS_WITH_SEQUENCER default-on force off if NO_NETWORK (which is tested in travis)
2019-08-09 09:14:48 +01:00
(void)pt;
#if defined(LWS_WITH_SEQUENCER)
sequencer: upgrade timeout to use us Adapt service loops and event libs to use microsecond waits internally, for hrtimer and sequencer. Reduce granularity according to platform / event lib wait. Add a helper so there's a single place to extend it.
2019-08-08 06:30:14 +01:00
lws_seq_destroy_all_on_pt(pt);
seq: LWS_WITH_SEQUENCER default-on force off if NO_NETWORK (which is tested in travis)
2019-08-09 09:14:48 +01:00
#endif
lws_validity: unified connection validity tracking Refactor everything around ping / pong handling in ws and h2, so there is instead a protocol-independent validity lws_sul tracking how long it has been since the last exchange that confirms the operation of the network connection in both directions. Clean out periodic role callback and replace the last two role users with discrete lws_sul for each pt.
2019-09-18 13:09:32 +01:00
LWS_FOR_EVERY_AVAILABLE_ROLE_START(ar) {
if (ar->pt_init_destroy)
ar->pt_init_destroy(context, NULL, pt, 1);
} LWS_FOR_EVERY_AVAILABLE_ROLE_END;
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
cgi: call init_destroy even though not on available roles list cgi is a role, but it's not an externally-selectable role... it's only used when the mount indicates it should be. That means it's not in the list of EVERY_AVAILABLE_ROLE... make sure its zombie child reaping timer is started at context creation and stopped at destruction.
2019-10-20 07:12:51 +01:00
#if defined(LWS_WITH_CGI)
role_ops_cgi.pt_init_destroy(context, NULL, pt, 1);
#endif
event libs: context_destroy sequencing
2020-03-10 19:13:43 +00:00
#if 0
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
if (context->event_loop_ops->destroy_pt)
context->event_loop_ops->destroy_pt(context, n);
event libs: context_destroy sequencing
2020-03-10 19:13:43 +00:00
#endif
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
while (pt->http.ah_list)
_lws_destroy_ah(pt, pt->http.ah_list);
#endif
}
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
#if defined(LWS_WITH_SYS_SMD)
_lws_smd_destroy(context);
#endif
asynchronous dns for ipv4 and ipv6 This adds the option to have lws do its own dns resolution on the event loop, without blocking. Existing implementations get the name resolution done by the libc, which is blocking. In the case you are opening client connections but need to carefully manage latency, another connection opening and doing the name resolution becomes a big problem. Currently it supports - ipv4 / A records - ipv6 / AAAA records - ipv4-over-ipv6 ::ffff:1.2.3.4 A record promotion for ipv6 - only one server supported over UDP :53 - nameserver discovery on linux, windows, freertos It also has some nice advantages - lws-style paranoid response parsing - random unique tid generation to increase difficulty of poisoning - it's really integrated with the lws event loop, it does not spawn threads or use the libc resolver, and of course no blocking at all - platform-specific server address capturing (from /etc/resolv.conf on linux, windows apis on windows) - it has LRU caching - piggybacking (multiple requests before the first completes go on a list on the first request, not spawn multiple requests) - observes TTL in cache - TTL and timeout use lws_sul timers on the event loop - ipv6 pieces only built if cmake LWS_IPV6 enabled
2019-09-19 06:54:53 +01:00
#if defined(LWS_WITH_SYS_ASYNC_DNS)
lws_async_dns_deinit(&context->async_dns);
#endif
lws_system: dhcpclient Generic lws_system IPv4 DHCP client - netif and route control via lib/plat apis - linux plat pieces implemented - Uses raw ip socket for UDP broadcast and rx - security-aware - usual stuff plus up to 4 x dns server If it's enabled for build, it holds the system state at DHCP until at least one registered interface has acquired a set of IP / mask / router / DNS server It uses PF_PACKET which is Linux-only atm. But those areas are isolated into plat code. TODOs - lease timing and reacquire - plat pieces for other than Linux
2019-09-27 14:32:59 -07:00
#if defined(LWS_WITH_SYS_DHCP_CLIENT)
lws_dhcpc_remove(context, NULL);
#endif
asynchronous dns for ipv4 and ipv6 This adds the option to have lws do its own dns resolution on the event loop, without blocking. Existing implementations get the name resolution done by the libc, which is blocking. In the case you are opening client connections but need to carefully manage latency, another connection opening and doing the name resolution becomes a big problem. Currently it supports - ipv4 / A records - ipv6 / AAAA records - ipv4-over-ipv6 ::ffff:1.2.3.4 A record promotion for ipv6 - only one server supported over UDP :53 - nameserver discovery on linux, windows, freertos It also has some nice advantages - lws-style paranoid response parsing - random unique tid generation to increase difficulty of poisoning - it's really integrated with the lws event loop, it does not spawn threads or use the libc resolver, and of course no blocking at all - platform-specific server address capturing (from /etc/resolv.conf on linux, windows apis on windows) - it has LRU caching - piggybacking (multiple requests before the first completes go on a list on the first request, not spawn multiple requests) - observes TTL in cache - TTL and timeout use lws_sul timers on the event loop - ipv6 pieces only built if cmake LWS_IPV6 enabled
2019-09-19 06:54:53 +01:00
minimal: update copies of test assets to CSP-compatible versions
2018-11-15 17:15:29 +08:00
if (context->pt[0].fds)
lws_free_set_NULL(context->pt[0].fds);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
openssl: add threading apis needed for 1.0.2 and earlier In later OpenSSL, eg 1.1.1, these apis are NOPs in OpenSSL. They are required for normal operation with multithreading on earlier OpenSSL, eg, 1.0.2k
2019-08-02 09:47:23 +01:00
lws_context_deinit_ssl_library(context);
detailed latency stats Remove LWS_LATENCY. Add the option LWS_WITH_DETAILED_LATENCY, allowing lws to collect very detailed information on every read and write, and allow the user code to provide a callback to process events.
2019-08-23 16:10:36 +01:00
#if defined(LWS_WITH_DETAILED_LATENCIES)
if (context->latencies_fd != -1)
compatible_close(context->latencies_fd);
#endif
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
for (n = 0; n < LWS_SYSBLOB_TYPE_COUNT; n++)
lws_system_blob_destroy(
lws_system_get_blob(context, n, 0));
smp: helgrind and sai lock fixes
2020-04-15 20:57:00 +01:00
#if LWS_MAX_SMP > 1
lws_mutex_refcount_destroy(&context->mr);
#endif
event libs: default to building as dynamically loaded plugins Event lib support as it has been isn't scaling well, at the low level libevent and libev headers have a namespace conflict so they can't both be built into the same image, and at the distro level, binding all the event libs to libwebsockets.so makes a bloaty situation for packaging, lws will drag in all the event libs every time. This patch implements the plan discussed here https://github.com/warmcat/libwebsockets/issues/1980 and refactors the event lib support so they are built into isolated plugins and bound at runtime according to what the application says it wants to use. The event lib plugins can be packaged individually so that only the needed sets of support are installed (perhaps none of them if the user code is OK with the default poll() loop). And dependent user code can mark the specific event loop plugin package as required so pieces are added as needed. The eventlib-foreign example is also refactored to build the selected lib support isolated. A readme is added detailing the changes and how to use them. https://libwebsockets.org/git/libwebsockets/tree/READMEs/README.event-libs.md
2020-08-27 15:37:14 +01:00
#if defined(LWS_WITH_EVLIB_PLUGINS) && defined(LWS_WITH_EVENT_LIBS)
if (context->evlib_plugin_list)
lws_plugins_destroy(&context->evlib_plugin_list, NULL, NULL);
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_free(context);
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
lwsl_debug("%s: ctx %p freed\n", __func__, context);
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (pcontext_finalize)
*pcontext_finalize = NULL;
}
context_destroy: figure out if anything still in event loop
2018-05-02 19:27:29 +08:00
/*
* really start destroying things
*/
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
void
lws_context_destroy2(struct lws_context *context)
{
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
struct lws_vhost *vh = NULL, *vh1;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
int n;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
#if defined(LWS_WITH_PEER_LIMITS)
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
uint32_t nu;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
#endif
lwsl_info("%s: ctx %p\n", __func__, context);
smp: context lock around vhost destroy
2018-06-27 07:49:04 +08:00
lws_context_lock(context, "context destroy 2"); /* ------ context { */
refactor: remove more test apps and replace with minimal-examples
2018-04-30 09:16:04 +08:00
context->being_destroyed2 = 1;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
/*
* We're going to trash things like vhost-protocols
* So we need to finish dealing with wsi close that
* might make callbacks first
*/
for (n = 0; n < context->count_threads; n++) {
struct lws_context_per_thread *pt = &context->pt[n];
(void)pt;
#if defined(LWS_WITH_SECURE_STREAMS)
lws_dll2_foreach_safe(&pt->ss_owner, NULL, lws_ss_destroy_dll);
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#if !defined(LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY)
ss: server: h1, h2, ws basic support Add initial support for defining servers using Secure Streams policy and api semantics. Serving h1, h2 and ws should be functional, the new minimal example shows a combined http + SS server with an incrementing ws message shown in the browser over tls, in around 200 lines of user code. NOP out anything to do with plugins, they're not currently used. Update the docs correspondingly.
2020-07-27 10:03:12 +01:00
while (context->server_der_list) {
struct lws_ss_x509 *x = context->server_der_list;
context->server_der_list = x->next;
lws_free((void *)x->ca_der);
}
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
if (context->ac_policy)
lwsac_free(&context->ac_policy);
#endif
ss: add static policy as a build option In some cases devices may be too constrained to handle JSON policies but still want to use SS apis and methodology. This introduces an off-by-default cmake option LWS_WITH_SECURE_STREAMS_STATIC_POLICY_ONLY, if enabled the JSON parsing part is excluded and it's assumed the user code provides its policy as hardcoded policy structs.
2020-03-26 06:48:34 +00:00
#endif
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#if defined(LWS_WITH_SECURE_STREAMS_PROXY_API)
lws_dll2_foreach_safe(&pt->ss_client_owner, NULL, lws_sspc_destroy_dll);
#endif
#if defined(LWS_WITH_SEQUENCER)
lws_seq_destroy_all_on_pt(pt);
#endif
LWS_FOR_EVERY_AVAILABLE_ROLE_START(ar) {
if (ar->pt_init_destroy)
ar->pt_init_destroy(context, NULL, pt, 1);
} LWS_FOR_EVERY_AVAILABLE_ROLE_END;
#if defined(LWS_WITH_CGI)
role_ops_cgi.pt_init_destroy(context, NULL, pt, 1);
#endif
if (context->event_loop_ops->destroy_pt)
context->event_loop_ops->destroy_pt(context, n);
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
while (pt->http.ah_list)
_lws_destroy_ah(pt, pt->http.ah_list);
#endif
}
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
/*
* free all the per-vhost allocations
*/
vh = context->vhost_list;
while (vh) {
vh1 = vh->vhost_next;
smp: context lock around vhost destroy
2018-06-27 07:49:04 +08:00
__lws_vhost_destroy2(vh);
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
vh = vh1;
}
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
lwsl_debug("%p: post vh listl\n", __func__);
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
/* remove ourselves from the pending destruction list */
while (context->vhost_pending_destruction_list)
/* removes itself from list */
smp: context lock around vhost destroy
2018-06-27 07:49:04 +08:00
__lws_vhost_destroy2(context->vhost_pending_destruction_list);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
lwsl_debug("%p: post pdl\n", __func__);
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_stats_log_dump(context);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_ssl_context_destroy(context);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_plat_context_late_destroy(context);
#if defined(LWS_WITH_PEER_LIMITS)
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
for (nu = 0; nu < context->pl_hash_elements; nu++) {
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_start_foreach_llp(struct lws_peer **, peer,
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
context->pl_hash_table[nu]) {
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
struct lws_peer *df = *peer;
*peer = df->next;
lws_free(df);
continue;
} lws_end_foreach_llp(peer, next);
}
lws_free(context->pl_hash_table);
#endif
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
lwsl_debug("%p: baggage\n", __func__);
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->external_baggage_free_on_destroy)
free(context->external_baggage_free_on_destroy);
smp: helgrind and sai lock fixes
2020-04-15 20:57:00 +01:00
lws_context_unlock(context); /* } context ------ */
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->event_loop_ops->destroy_context2)
if (context->event_loop_ops->destroy_context2(context)) {
context->finalize_destroy_after_internal_loops_stopped = 1;
return;
}
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
lwsl_debug("%p: post dc2\n", __func__);
smp: additional locking for libuv With SMP + event lib, extra locking is required when dealing with cross-thread adoption case, and cross-vhost cases like wsi close, we need to hold the pt or context lock. These lock apis are NOPs when LWS_MAX_SMP == 1 which is the default.
2020-10-01 08:39:25 +01:00
// if (!context->pt[0].event_loop_foreign) {
// int n;
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
for (n = 0; n < context->count_threads; n++)
coverity: cleanup
2018-08-01 06:52:03 +08:00
if (context->pt[n].inside_service) {
minimal-http-server-eventlib-smp
2019-03-21 09:27:45 +08:00
lwsl_debug("%p: bailing as inside service\n", __func__);
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
return;
coverity: cleanup
2018-08-01 06:52:03 +08:00
}
smp: additional locking for libuv With SMP + event lib, extra locking is required when dealing with cross-thread adoption case, and cross-vhost cases like wsi close, we need to hold the pt or context lock. These lock apis are NOPs when LWS_MAX_SMP == 1 which is the default.
2020-10-01 08:39:25 +01:00
// }
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
smp: context lock around vhost destroy
2018-06-27 07:49:04 +08:00
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_context_destroy3(context);
}
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
#if defined(LWS_WITH_NETWORK)
static void
lws_pt_destroy(struct lws_context_per_thread *pt)
{
volatile struct lws_foreign_thread_pollfd *ftp, *next;
volatile struct lws_context_per_thread *vpt;
assert(!pt->is_destroyed);
pt->destroy_self = 0;
vpt = (volatile struct lws_context_per_thread *)pt;
ftp = vpt->foreign_pfd_list;
while (ftp) {
next = ftp->next;
lws_free((void *)ftp);
ftp = next;
}
vpt->foreign_pfd_list = NULL;
smp: lws_mutex_refcount: add assert held helper also additional pt locks shown as needed by that
2020-08-31 08:12:16 +01:00
lws_pt_lock(pt, __func__);
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
if (pt->pipe_wsi)
lws_destroy_event_pipe(pt->pipe_wsi);
smp: lws_mutex_refcount: add assert held helper also additional pt locks shown as needed by that
2020-08-31 08:12:16 +01:00
lws_pt_unlock(pt);
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
pt->pipe_wsi = NULL;
while (pt->fds_count) {
struct lws *wsi = wsi_from_fd(pt->context, pt->fds[0].fd);
if (!wsi)
break;
lws_close_free_wsi(wsi,
LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY,
"ctx destroy"
/* no protocol close */);
}
lws_pt_mutex_destroy(pt);
pt->is_destroyed = 1;
lwsl_info("%s: pt destroyed\n", __func__);
}
#endif
context_destroy: figure out if anything still in event loop
2018-05-02 19:27:29 +08:00
/*
* Begin the context takedown
*/
cleaning
2020-01-02 08:32:23 +00:00
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_context_destroy(struct lws_context *context)
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
{
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
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_vhost *vh = NULL;
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
int m, deferred_pt = 0;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#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
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
if (!context || context->inside_context_destroy)
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
return;
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
context->inside_context_destroy = 1;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->finalize_destroy_after_internal_loops_stopped) {
if (context->event_loop_ops->destroy_context2)
context->event_loop_ops->destroy_context2(context);
lws_context_destroy3(context);
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
/* context is invalid, no need to reset inside flag */
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
return;
}
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#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
if (context->being_destroyed1) {
refactor: change event lib minimal examples to serve as the test apps
2018-04-30 19:17:32 +08:00
if (!context->being_destroyed2) {
lws_context_destroy2(context);
malloc_trim: move out of unix plat and into lwsws https://github.com/warmcat/libwebsockets/issues/1849
2020-02-24 10:06:43 +00:00
return;
refactor: change event lib minimal examples to serve as the test apps
2018-04-30 19:17:32 +08:00
}
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lwsl_info("%s: ctx %p: already being destroyed\n",
clean
2017-09-23 12:55:21 +08:00
__func__, context);
hrtimer: add support for libevent and libev
2018-05-02 18:35:58 +08:00
lws_context_destroy3(context);
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
/* context is invalid, no need to reset inside flag */
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
return;
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
}
logging: move most everything during startup to info
2017-09-19 17:30:06 +08:00
lwsl_info("%s: ctx %p\n", __func__, context);
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
avoid using deallocated things during context dedtroy Signed-off-by: Andy Green <andy.green@linaro.org>
2016-01-18 11:49:41 +08:00
context->being_destroyed = 1;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
lws_state and system state Introduce a generic lws_state object with notification handlers that may be registered in a chain. Implement one of those in the context to manage the "system state". Allow other pieces of lws and user code to register notification handlers on a context list. Handlers can object to or take over responsibility to move forward and retry system state changes if they know that some dependent action must succeed first. For example if the system time is invalid, we cannot move on to a state where anything can do tls until that has been corrected.
2019-09-19 09:48:17 +01:00
lws_state_transition(&context->mgr_system, LWS_SYSTATE_POLICY_INVALID);
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
m = context->count_threads;
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
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
while (m--) {
codacy: minor fixes
2018-10-10 13:54:43 +08:00
struct lws_context_per_thread *pt = &context->pt[m];
bsd: support foreign thread event changes while in poll wait https://github.com/warmcat/libwebsockets/issues/314
2017-11-12 09:16:46 +08:00
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
if (pt->is_destroyed)
continue;
if (pt->inside_lws_service) {
pt->destroy_self = 1;
deferred_pt = 1;
continue;
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
}
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
lws_pt_destroy(pt);
}
if (deferred_pt) {
lwsl_info("%s: waiting for deferred pt close\n", __func__);
lws_cancel_service(context);
goto out;
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
}
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
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
context->being_destroyed1 = 1;
context->requested_kill = 1;
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
/*
* inform all the protocols that they are done and will have no more
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
* callbacks.
*
* We can't free things until after the event loop shuts down.
refactor context related functions into context.c Signed-off-by: Andy Green <andy.green@linaro.org>
2014-04-03 07:29:50 +08:00
*/
context only destroy protocols if init got sent If for some reason we exit before the protocol init action (which is delayed for libuv) we should not send the protocol destroy messages Signed-off-by: Andy Green <andy@warmcat.com>
2016-05-08 17:00:38 +08:00
if (context->protocol_init_done)
vh = context->vhost_list;
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
while (vh) {
vh protocol destroy: dont miss vhosts
2017-10-11 19:42:45 +08:00
struct lws_vhost *vhn = vh->vhost_next;
lws-vhost-destroy Adds a new api lws_vhost_destroy(struct lws_vhost *) which allows dynamic removal of vhosts. The external api calls two parts of internal helpers that get reused for context destroy. The second part is called deferred by 5s... this is to ensure that event library objects composed into structs owned by the vhost all have a chance to complete their close asynchronously. That should happen immediately, but it requires us to return to the event loop first. The vhost being removed is deleted from the context vhost list by the first part, and does not block further removals or creation during the delay for the deferred freeing of the vhost memory. Part 1: - if the vhost owned a listen socket needed by other vhosts listening on same iface + port, the listen socket is first handed off to another vhost so it stays alive - all wsi still open on the vhost are forcibly closed (including any listen socket still attached) - inform all active protocols on the vhost they should destroy themselves - remove vhost from context vhost list (can no longer be found by incoming connections) - add to a "being destroyed" context list and schedule the second part to be called in 5s Part 2: - remove us from the being destroyed list - free all allocations owned by the vhost - zero down the vhost and free the vhost itself In libwebsockets-test-server, you can send it a SIGUSR1 to have it toggle the creation and destruction of a second vhost on port + 1.
2017-07-15 14:37:04 +08:00
lws_vhost_destroy1(vh);
vh protocol destroy: dont miss vhosts
2017-10-11 19:42:45 +08:00
vh = vhn;
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_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#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
lws_plat_context_early_destroy(context);
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_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#if defined(LWS_WITH_NETWORK)
context_destroy: figure out if anything still in event loop
2018-05-02 19:27:29 +08:00
/*
* We face two different needs depending if foreign loop or not.
*
* 1) If foreign loop, we really want to advance the destroy_context()
* past here, and block only for libuv-style async close completion.
*
* 2a) If poll, and we exited by ourselves and are calling a final
* destroy_context() outside of any service already, we want to
* advance all the way in one step.
*
* 2b) If poll, and we are reacting to a SIGINT, service thread(s) may
* be in poll wait or servicing. We can't advance the
* destroy_context() to the point it's freeing things; we have to
* leave that for the final destroy_context() after the service
* thread(s) are finished calling for service.
*/
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
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
if (context->event_loop_ops->destroy_context1) {
context->event_loop_ops->destroy_context1(context);
lws-vhost-destroy Adds a new api lws_vhost_destroy(struct lws_vhost *) which allows dynamic removal of vhosts. The external api calls two parts of internal helpers that get reused for context destroy. The second part is called deferred by 5s... this is to ensure that event library objects composed into structs owned by the vhost all have a chance to complete their close asynchronously. That should happen immediately, but it requires us to return to the event loop first. The vhost being removed is deleted from the context vhost list by the first part, and does not block further removals or creation during the delay for the deferred freeing of the vhost memory. Part 1: - if the vhost owned a listen socket needed by other vhosts listening on same iface + port, the listen socket is first handed off to another vhost so it stays alive - all wsi still open on the vhost are forcibly closed (including any listen socket still attached) - inform all active protocols on the vhost they should destroy themselves - remove vhost from context vhost list (can no longer be found by incoming connections) - add to a "being destroyed" context list and schedule the second part to be called in 5s Part 2: - remove us from the being destroyed list - free all allocations owned by the vhost - zero down the vhost and free the vhost itself In libwebsockets-test-server, you can send it a SIGUSR1 to have it toggle the creation and destruction of a second vhost on port + 1.
2017-07-15 14:37:04 +08:00
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
goto out;
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
}
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
#endif
context: external_baggage_free_on_destroy This adds a context creation-time member that points to something that should be freed when the context is destroyed. It's in preparation for context deprecation, when a context might be destroyed asynchronously... a related external with the lifetime of thee context should also be freed at that time. Adapt lwsws to use it with the context "strings" (also used for aligned structs created by the config) allocation.
2016-12-04 07:34:05 +08:00
freertos: rename esp32 plat to freertos
2019-08-18 10:35:43 +01:00
#if defined(LWS_PLAT_FREERTOS)
rtos: add support for AMAZON_RTOS
2019-06-05 05:04:17 +01:00
#if defined(LWS_AMAZON_RTOS)
context->last_free_heap = xPortGetFreeHeapSize();
#else
context->last_free_heap = esp_get_free_heap_size();
#endif
#endif
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
context->inside_context_destroy = 0;
refactor: apply ops structs to event loop handlers
2018-04-29 10:44:36 +08:00
lws_context_destroy2(context);
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
return;
malloc_trim: move out of unix plat and into lwsws https://github.com/warmcat/libwebsockets/issues/1849
2020-02-24 10:06:43 +00:00
#if defined(LWS_WITH_NETWORK)
event lib: update http client multi to work with it and clean destroy flow Add selectable event lib support to minimal-http-client-multi and clean up context destroy flow so we can use lws_destroy_context() from inside the callback to indicate we want to end the event loop, without using the traditional "interrupted" flag and in a way that works no matter which event loop backend is being used.
2020-01-23 11:12:28 +00:00
out:
context->inside_context_destroy = 0;
malloc_trim: move out of unix plat and into lwsws https://github.com/warmcat/libwebsockets/issues/1849
2020-02-24 10:06:43 +00:00
#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
}
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#if defined(LWS_WITH_SYS_STATE)
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
struct lws_context *
lws_system_context_from_system_mgr(lws_state_manager_t *mgr)
{
#if defined(LWS_WITH_NETWORK)
lws_smd: system message distribution - Add low level system message distibution framework - Add support for local Secure Streams to participate using _lws_smd streamtype - Add apit test and minimal example - Add SS proxy support for _lws_smd See minimal-secure-streams-smd README.md
2020-06-24 20:15:46 +01:00
return mgr->context;
client: secure streams Secure Streams is an optional layer on top of lws that separates policy like endpoint selection and tls cert validation into a device JSON policy document. Code that wants to open a client connection just specifies a streamtype name, and no longer deals with details like the endpoint, the protocol (!) or anything else other than payloads and optionally generic metadata; the JSON policy contains all the details for each streamtype. h1, h2, ws and mqtt client connections are supported. Logical secure streams outlive any particular connection and supports "nailed-up" connectivity regardless of underlying connection stability.
2020-02-29 12:37:24 +00:00
#else
return NULL;
#endif
}
context: option to disable system state management
2020-06-23 13:19:30 +01:00
#endif
Reference in a new issue Copy permalink