Mirrors/libwebsockets
1
0
Fork 0
mirror of https://github.com/warmcat/libwebsockets.git synced 2025-03-09 00:00:04 +01:00
Code Issues Releases Wiki Activity
libwebsockets/lib/core-net/wsi-timeout.c

290 lines
8.5 KiB
C
Raw Permalink Normal View History

LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
/*
* libwebsockets - small server side websockets and web server implementation
*
license: switch LGPLv2.1+SLE parts to MIT
2019-08-14 10:44:14 +01:00
* Copyright (C) 2010 - 2019 Andy Green <andy@warmcat.com>
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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:
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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.
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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.
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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"
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
void
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_wsi_remove_from_sul(struct lws *wsi)
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
{
fi: wsi: fault to fake peer hangup after a delay range Example usage, client $ valgrind ./bin/lws-minimal-secure-streams-stress-client -c 4 --budget 40 proxy $ valgrind ./lws-minimal-secure-streams-proxy --fault-injection "ss=mintest/wsi/timedclose(20%),ss=mintest/wsi/timedclose_ms(0..100)"
2021-06-29 13:30:48 +01:00
lws_sul_cancel(&wsi->sul_timeout);
lws_sul_cancel(&wsi->sul_hrtimer);
lws_sul_cancel(&wsi->sul_validity);
#if defined(LWS_WITH_SYS_FAULT_INJECTION)
lws_sul_cancel(&wsi->sul_fault_timedclose);
#endif
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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
/*
* hrtimer
*/
static void
lws_sul_hrtimer_cb(lws_sorted_usec_list_t *sul)
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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
struct lws *wsi = lws_container_of(sul, struct lws, sul_hrtimer);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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 (wsi->a.protocol &&
wsi->a.protocol->callback(wsi, LWS_CALLBACK_TIMER,
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
wsi->user_space, NULL, 0))
__lws_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS,
"hrtimer cb errored");
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
}
void
lws_dll: upgrade all instances to lws_dll2 lws_dll2 removes the downsides of lws_dll and adds new features like a running member count and explicit owner type... it's cleaner and more robust (eg, nodes know their owner, so they can casually switch between list owners and remove themselves without the code knowing the owner). This deprecates lws_dll, but since it's public it allows it to continue to be built for 4.0 release if you give cmake LWS_WITH_DEPRECATED_LWS_DLL. All remaining internal users of lws_dll are migrated to lws_dll2.
2019-08-08 16:58:55 +01:00
__lws_set_timer_usecs(struct lws *wsi, lws_usec_t us)
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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
wsi->sul_hrtimer.cb = lws_sul_hrtimer_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(&pt->pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&wsi->sul_hrtimer, us);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
}
cleaning
2020-01-02 08:32:23 +00:00
void
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
lws_set_timer_usecs(struct lws *wsi, lws_usec_t usecs)
{
__lws_set_timer_usecs(wsi, usecs);
}
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
/*
* wsi timeout
*/
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
static void
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_sul_wsitimeout_cb(lws_sorted_usec_list_t *sul)
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
{
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
struct lws *wsi = lws_container_of(sul, struct lws, sul_timeout);
smp: more lock assertions
2021-04-04 04:06:24 +01:00
struct lws_context *cx = wsi->a.context;
struct lws_context_per_thread *pt = &cx->pt[(int)wsi->tsi];
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
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
/* no need to log normal idle keepalive timeout */
// if (wsi->pending_timeout != PENDING_TIMEOUT_HTTP_KEEPALIVE_IDLE)
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
stats: move to pt and improve presentation
2019-08-10 07:32:32 +01:00
if (wsi->pending_timeout != PENDING_TIMEOUT_USER_OK)
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_info(wsi, "TIMEDOUT WAITING %d, dhdr %d, ah %p, wl %d",
wsi->pending_timeout,
wsi->hdr_parsing_completed, wsi->http.ah,
pt->http.ah_wait_list_length);
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_CGI)
if (wsi->http.cgi)
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_notice(wsi, "CGI timeout: %s", wsi->http.cgi->summary);
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
#else
stats: move to pt and improve presentation
2019-08-10 07:32:32 +01:00
if (wsi->pending_timeout != PENDING_TIMEOUT_USER_OK)
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_info(wsi, "TIMEDOUT WAITING on %d ",
wsi->pending_timeout);
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
/* cgi timeout */
if (wsi->pending_timeout != PENDING_TIMEOUT_HTTP_KEEPALIVE_IDLE)
/*
* Since he failed a timeout, he already had a chance to
* do something and was unable to... that includes
* situations like half closed connections. So process
* this "failed timeout" close as a violent death and
* don't try to do protocol cleanup like flush partials.
*/
wsi->socket_is_permanently_unusable = 1;
clean: internally use LWS_WITH_CLIENT and _SERVER Remove some more things in LWS_WITH_SERVER=0 case
2019-08-18 05:04:15 +01:00
#if defined(LWS_WITH_CLIENT)
lws_inform_client_conn_fail There are many places where we have the same code to inform about client connection failure... consolidate it
2019-08-10 09:20:27 +01:00
if (lwsi_state(wsi) == LRS_WAITING_SSL)
lws_inform_client_conn_fail(wsi,
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
(void *)"Timed out waiting SSL", 21);
client: make sure we get CCE if timeout on server reply If facing a captive portal, we may seem to get a tcp level connection okay but find that communication is silently dropped, leading to us timing out in LRS_WAITING_SERVER_REPLY. If so, we need to handle it as a connection fail in order to satisfy at least Captive Portal detection.
2021-04-17 05:39:12 +01:00
if (lwsi_state(wsi) == LRS_WAITING_SERVER_REPLY)
lws_inform_client_conn_fail(wsi,
(void *)"Timed out waiting server reply", 30);
lws_inform_client_conn_fail There are many places where we have the same code to inform about client connection failure... consolidate it
2019-08-10 09:20:27 +01:00
#endif
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
smp: more lock assertions
2021-04-04 04:06:24 +01:00
lws_context_lock(cx, __func__);
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__);
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_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS, "timeout");
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);
smp: more lock assertions
2021-04-04 04:06:24 +01:00
lws_context_unlock(cx);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
}
void
__lws_set_timeout(struct lws *wsi, enum pending_timeout reason, int secs)
{
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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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
wsi->sul_timeout.cb = lws_sul_wsitimeout_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(&pt->pt_sul_owner[LWSSULLI_MISS_IF_SUSPENDED],
&wsi->sul_timeout,
((lws_usec_t)secs) * LWS_US_PER_SEC);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
logs: log contexts
2021-06-18 07:28:23 +01:00
lwsl_wsi_debug(wsi, "%d secs, reason %d\n", secs, reason);
lws_dll: upgrade all instances to lws_dll2 lws_dll2 removes the downsides of lws_dll and adds new features like a running member count and explicit owner type... it's cleaner and more robust (eg, nodes know their owner, so they can casually switch between list owners and remove themselves without the code knowing the owner). This deprecates lws_dll, but since it's public it allows it to continue to be built for 4.0 release if you give cmake LWS_WITH_DEPRECATED_LWS_DLL. All remaining internal users of lws_dll are migrated to lws_dll2.
2019-08-08 16:58:55 +01:00
type comparisons: fixes This is a huge patch that should be a global NOP. For unix type platforms it enables -Wconversion to issue warnings (-> error) for all automatic casts that seem less than ideal but are normally concealed by the toolchain. This is things like passing an int to a size_t argument. Once enabled, I went through all args on my default build (which build most things) and tried to make the removed default cast explicit. With that approach it neither change nor bloat the code, since it compiles to whatever it was doing before, just with the casts made explicit... in a few cases I changed some length args from int to size_t but largely left the causes alone. From now on, new code that is relying on less than ideal casting will complain and nudge me to improve it by warnings.
2020-12-12 06:21:40 +00:00
wsi->pending_timeout = (char)reason;
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
}
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
void
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
lws_set_timeout(struct lws *wsi, enum pending_timeout reason, int secs)
{
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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +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_context_lock(pt->context, __func__);
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_pt_lock(pt, __func__);
lws_dll2_remove(&wsi->sul_timeout.list);
lws_pt_unlock(pt);
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
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 (!secs)
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
goto bail;
lws_http_mark_sse https://github.com/warmcat/libwebsockets/issues/1486
2019-01-30 20:59:56 +08:00
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
if (secs == LWS_TO_KILL_SYNC) {
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_debug(wsi, "TO_KILL_SYNC");
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_context_unlock(pt->context);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
lws_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS,
"to sync kill");
return;
}
if (secs == LWS_TO_KILL_ASYNC)
secs = 0;
mux children: generalize helpers out of h2 implementation This should be a NOP for h2 support and only affects internal apis. But it lets us reuse the working and reliable h2 mux arrangements directly in other protocols later, and share code so building for h2 + new protocols can take advantage of common mux child handling struct and code. Break out common mux handling struct into its own type. Convert all uses of members that used to be in wsi->h2 to wsi->mux Audit all references to the members and break out generic helpers for anything that is useful for other mux-capable protocols to reuse wsi->mux related features.
2019-12-23 11:31:57 +00:00
// assert(!secs || !wsi->mux_stream_immortal);
if (secs && wsi->mux_stream_immortal)
logs: log contexts
2021-06-18 07:28:23 +01:00
lwsl_wsi_err(wsi, "on immortal stream %d %d", reason, secs);
h2: unify immortal stream tracking across SSE and ws substreams It was already correct but add helpers to isolate and deduplicate processing adding and closing a generically immortal stream. Change the default 31s h2 network connection timeout to be settable by .keepalive_timeout if nonzero. Add a public api allowing a client h2 stream to transition to half-closed LOCAL (by sending a 0-byte DATA with END_STREAM) and mark itself as immortal to create a read-only long-poll stream if the server allows it. Add a vhost server option flag LWS_SERVER_OPTION_VH_H2_HALF_CLOSED_LONG_POLL which allows the vhost to treat half-closed remotes as immortal long poll streams.
2019-09-13 10:33:24 +01:00
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
lws_pt_lock(pt, __func__);
__lws_set_timeout(wsi, reason, secs);
lws_pt_unlock(pt);
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
bail:
lws_context_unlock(pt->context);
LWS_WITH_NETWORK: cmake option for no network code
2019-01-13 06:58:21 +08:00
}
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
void
lws_set_timeout_us(struct lws *wsi, enum pending_timeout reason, lws_usec_t us)
{
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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
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
lws_pt_lock(pt, __func__);
lws_dll2_remove(&wsi->sul_timeout.list);
lws_pt_unlock(pt);
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
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 (!us)
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
return;
lws_pt_lock(pt, __func__);
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],
&wsi->sul_timeout, us);
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_notice(wsi, "%llu us, reason %d",
(unsigned long long)us, reason);
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
type comparisons: fixes This is a huge patch that should be a global NOP. For unix type platforms it enables -Wconversion to issue warnings (-> error) for all automatic casts that seem less than ideal but are normally concealed by the toolchain. This is things like passing an int to a size_t argument. Once enabled, I went through all args on my default build (which build most things) and tried to make the removed default cast explicit. With that approach it neither change nor bloat the code, since it compiles to whatever it was doing before, just with the casts made explicit... in a few cases I changed some length args from int to size_t but largely left the causes alone. From now on, new code that is relying on less than ideal casting will complain and nudge me to improve it by warnings.
2020-12-12 06:21:40 +00:00
wsi->pending_timeout = (char)reason;
unify us sorted waits There are quite a few linked-lists of things that want events after some period. This introduces a type binding an lws_dll2 for the list and a lws_usec_t for the duration. The wsi timeouts, the hrtimer and the sequencer timeouts are converted to use these, also in the common event wait calculation.
2019-08-08 20:05:03 +01:00
lws_pt_unlock(pt);
}
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
static void
lws_validity_cb(lws_sorted_usec_list_t *sul)
{
struct lws *wsi = lws_container_of(sul, struct lws, sul_validity);
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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
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
const lws_retry_bo_t *rbo = wsi->retry_policy;
/* one of either the ping or hangup validity threshold was crossed */
if (wsi->validity_hup) {
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_info(wsi, "validity too old");
validity: fix bad pointer access We cannot access wsi after __lws_close_free_wsi.
2022-05-03 16:47:14 +02:00
struct lws_context *cx = wsi->a.context;
struct lws_context_per_thread *pt = &cx->pt[(int)wsi->tsi];
validity: take context and pt lock before close
2021-08-27 13:08:03 +01:00
validity: fix bad pointer access We cannot access wsi after __lws_close_free_wsi.
2022-05-03 16:47:14 +02:00
lws_context_lock(cx, __func__);
validity: take context and pt lock before close
2021-08-27 13:08:03 +01:00
lws_pt_lock(pt, __func__);
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_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS,
"validity timeout");
validity: take context and pt lock before close
2021-08-27 13:08:03 +01:00
lws_pt_unlock(pt);
validity: fix bad pointer access We cannot access wsi after __lws_close_free_wsi.
2022-05-03 16:47:14 +02:00
lws_context_unlock(cx);
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
return;
}
/* schedule a protocol-dependent ping */
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_info(wsi, "scheduling validity check");
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
roles: compress role ops structs role ops are usually only sparsely filled, there are currently 20 function pointers but several roles only fill in two. No single role has more than 14 of the ops. On a 32/64 bit build this part of the ops struct takes a fixed 80 / 160 bytes then. First reduce the type of the callback reason part from uint16_t to uint8_t, this saves 12 bytes unconditionally. Change to a separate function pointer array with a nybble index array, it costs 10 bytes for the index and a pointer to the separate array, for 32-bit the cost is 2 + (4 x ops_used) and for 64-bit 6 + (8 x ops_used) for 2 x ops_used it means 32-bit: 10 vs 80 / 64-bit: 22 vs 160 For a typical system with h1 (9), h2 (14), listen (2), netlink (2), pipe (1), raw_skt (3), ws (12), == 43 ops_used out of 140, it means the .rodata for this reduced from 32-bit: 560 -> 174 (386 byte saving) and 64-bit: 1120 -> 350 (770 byte saving) This doesn't account for the changed function ops calling code, two ways were tried, a preprocessor macro and explicit functions For an x86_64 gcc 10 build with most options, release mode, .text + .rodata before patch: 553282 accessor macro: 552714 (568 byte saving) accessor functions: 553674 (392 bytes worse than without patch) therefore we went with the macros
2020-10-19 13:55:21 +01:00
if (lws_rops_fidx(wsi->role_ops, LWS_ROPS_issue_keepalive))
lws_rops_func_fidx(wsi->role_ops, LWS_ROPS_issue_keepalive).
issue_keepalive(wsi, 0);
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
/*
* We arrange to come back here after the additional ping to hangup time
* and do the hangup, unless we get validated (by, eg, a PONG) and
* reset the timer
*/
assert(rbo->secs_since_valid_hangup > rbo->secs_since_valid_ping);
wsi->validity_hup = 1;
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[!!wsi->conn_validity_wakesuspend],
&wsi->sul_validity,
((uint64_t)rbo->secs_since_valid_hangup -
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
rbo->secs_since_valid_ping) * LWS_US_PER_SEC);
}
/*
* The role calls this back to actually confirm validity on a particular wsi
* (which may not be the original wsi)
*/
void
_lws_validity_confirmed_role(struct lws *wsi)
{
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
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
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
const lws_retry_bo_t *rbo = wsi->retry_policy;
if (!rbo || !rbo->secs_since_valid_hangup)
return;
wsi->validity_hup = 0;
wsi->sul_validity.cb = lws_validity_cb;
wsi->validity_hup = rbo->secs_since_valid_ping >=
rbo->secs_since_valid_hangup;
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
lwsl_wsi_info(wsi, "setting validity timer %ds (hup %d)",
wsi->validity_hup ? rbo->secs_since_valid_hangup :
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
rbo->secs_since_valid_ping,
logs: more internal conversion to log_cx
2021-06-28 05:05:57 +01:00
wsi->validity_hup);
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
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[!!wsi->conn_validity_wakesuspend],
&wsi->sul_validity,
((uint64_t)(wsi->validity_hup ?
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
rbo->secs_since_valid_hangup :
rbo->secs_since_valid_ping)) * LWS_US_PER_SEC);
}
void
lws_validity_confirmed(struct lws *wsi)
{
/*
* This may be a stream inside a muxed network connection... leave it
* to the role to figure out who actually needs to understand their
* validity was confirmed.
*/
ws-over-h2: disable validity checking on encapsulated wsi The h2 nwsi is subject to its own validity checking PINGs, we don't need to also do them to encapsulated ws-over-h2
2019-11-05 04:53:09 +00:00
if (!wsi->h2_stream_carries_ws && /* only if not encapsulated */
roles: compress role ops structs role ops are usually only sparsely filled, there are currently 20 function pointers but several roles only fill in two. No single role has more than 14 of the ops. On a 32/64 bit build this part of the ops struct takes a fixed 80 / 160 bytes then. First reduce the type of the callback reason part from uint16_t to uint8_t, this saves 12 bytes unconditionally. Change to a separate function pointer array with a nybble index array, it costs 10 bytes for the index and a pointer to the separate array, for 32-bit the cost is 2 + (4 x ops_used) and for 64-bit 6 + (8 x ops_used) for 2 x ops_used it means 32-bit: 10 vs 80 / 64-bit: 22 vs 160 For a typical system with h1 (9), h2 (14), listen (2), netlink (2), pipe (1), raw_skt (3), ws (12), == 43 ops_used out of 140, it means the .rodata for this reduced from 32-bit: 560 -> 174 (386 byte saving) and 64-bit: 1120 -> 350 (770 byte saving) This doesn't account for the changed function ops calling code, two ways were tried, a preprocessor macro and explicit functions For an x86_64 gcc 10 build with most options, release mode, .text + .rodata before patch: 553282 accessor macro: 552714 (568 byte saving) accessor functions: 553674 (392 bytes worse than without patch) therefore we went with the macros
2020-10-19 13:55:21 +01:00
wsi->role_ops &&
lws_rops_fidx(wsi->role_ops, LWS_ROPS_issue_keepalive))
lws_rops_func_fidx(wsi->role_ops, LWS_ROPS_issue_keepalive).
issue_keepalive(wsi, 1);
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
}
Reference in a new issue Copy permalink