1
0
Fork 0
mirror of https://github.com/warmcat/libwebsockets.git synced 2025-03-16 00:00:07 +01:00
libwebsockets/lib/event-libs/libuv/libuv.c

999 lines
23 KiB
C
Raw Normal View History

/*
* libwebsockets - small server side websockets and web server implementation
*
* Copyright (C) 2010-2018 Andy Green <andy@warmcat.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation:
* version 2.1 of the License.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
* MA 02110-1301 USA
*/
#include "core/private.h"
static void
lws_uv_hrtimer_cb(uv_timer_t *timer
#if UV_VERSION_MAJOR == 0
, int status
#endif
)
{
struct lws_context_per_thread *pt = lws_container_of(timer,
struct lws_context_per_thread, uv.hrtimer);
lws_usec_t us;
lws_pt_lock(pt, __func__);
us = __lws_event_service_get_earliest_wake(pt, lws_now_usecs());
if (us)
uv_timer_start(&pt->uv.hrtimer, lws_uv_hrtimer_cb,
LWS_US_TO_MS(us), 0);
lws_pt_unlock(pt);
}
static void
lws_uv_idle(uv_idle_t *handle
#if UV_VERSION_MAJOR == 0
, int status
#endif
)
{
struct lws_context_per_thread *pt = lws_container_of(handle,
struct lws_context_per_thread, uv.idle);
lws_usec_t us;
2018-04-17 11:43:20 +08:00
lws_service_do_ripe_rxflow(pt);
/*
* is there anybody with pending stuff that needs service forcing?
*/
if (!lws_service_adjust_timeout(pt->context, 1, pt->tid)) {
/* -1 timeout means just do forced service */
_lws_plat_service_tsi(pt->context, -1, pt->tid);
/* still somebody left who wants forced service? */
if (!lws_service_adjust_timeout(pt->context, 1, pt->tid))
/* yes... come back again later */
return;
}
/* account for hrtimer */
lws_pt_lock(pt, __func__);
us = __lws_event_service_get_earliest_wake(pt, lws_now_usecs());
if (us)
uv_timer_start(&pt->uv.hrtimer, lws_uv_hrtimer_cb,
LWS_US_TO_MS(us), 0);
lws_pt_unlock(pt);
/* there is nobody who needs service forcing, shut down idle */
uv_idle_stop(handle);
}
static void
lws_io_cb(uv_poll_t *watcher, int status, int revents)
{
struct lws *wsi = (struct lws *)((uv_handle_t *)watcher)->data;
struct lws_context *context = wsi->context;
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
struct lws_pollfd eventfd;
#if defined(WIN32) || defined(_WIN32)
eventfd.fd = watcher->socket;
#else
eventfd.fd = watcher->io_watcher.fd;
#endif
eventfd.events = 0;
eventfd.revents = 0;
if (status < 0) {
2017-09-23 12:55:21 +08:00
/*
* At this point status will be an UV error, like UV_EBADF,
* we treat all errors as LWS_POLLHUP
*
* You might want to return; instead of servicing the fd in
* some cases */
if (status == UV_EAGAIN)
return;
eventfd.events |= LWS_POLLHUP;
eventfd.revents |= LWS_POLLHUP;
} else {
if (revents & UV_READABLE) {
eventfd.events |= LWS_POLLIN;
eventfd.revents |= LWS_POLLIN;
}
if (revents & UV_WRITABLE) {
eventfd.events |= LWS_POLLOUT;
eventfd.revents |= LWS_POLLOUT;
}
}
lws_service_fd_tsi(context, &eventfd, wsi->tsi);
uv_idle_start(&pt->uv.idle, lws_uv_idle);
}
/*
* This does not actually stop the event loop. The reason is we have to pass
* libuv handle closures through its event loop. So this tries to close all
* wsi, and set a flag; when all the wsi closures are finalized then we
* actually stop the libuv event loops.
*/
static void
lws_libuv_stop(struct lws_context *context)
{
struct lws_context_per_thread *pt;
int n, m;
lwsl_err("%s\n", __func__);
if (context->requested_kill) {
lwsl_err("%s: ignoring\n", __func__);
return;
}
context->requested_kill = 1;
m = context->count_threads;
context->being_destroyed = 1;
/*
* Phase 1: start the close of every dynamic uv handle
*/
while (m--) {
pt = &context->pt[m];
if (pt->pipe_wsi) {
uv_poll_stop(pt->pipe_wsi->w_read.uv.pwatcher);
lws_destroy_event_pipe(pt->pipe_wsi);
pt->pipe_wsi = NULL;
}
for (n = 0; (unsigned int)n < context->pt[m].fds_count; n++) {
struct lws *wsi = wsi_from_fd(context, pt->fds[n].fd);
if (!wsi)
continue;
lws_close_free_wsi(wsi,
2018-11-23 08:47:56 +08:00
LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY,
__func__ /* no protocol close */);
n--;
}
}
lwsl_info("%s: started closing all wsi\n", __func__);
/* we cannot have completed... there are at least the cancel pipes */
}
static void
lws_uv_signal_handler(uv_signal_t *watcher, int signum)
{
struct lws_context *context = watcher->data;
if (context->eventlib_signal_cb) {
context->eventlib_signal_cb((void *)watcher, signum);
return;
}
lwsl_err("internal signal handler caught signal %d\n", signum);
lws_libuv_stop(watcher->data);
}
static void
lws_uv_timeout_cb(uv_timer_t *timer
#if UV_VERSION_MAJOR == 0
, int status
#endif
)
{
struct lws_context_per_thread *pt = lws_container_of(timer,
struct lws_context_per_thread, uv.timeout_watcher);
if (pt->context->requested_kill)
return;
lwsl_debug("%s\n", __func__);
lws_service_fd_tsi(pt->context, NULL, pt->tid);
}
static const int sigs[] = { SIGINT, SIGTERM, SIGSEGV, SIGFPE, SIGHUP };
/*
* Closing Phase 2: Close callback for a static UV asset
*/
static void
lws_uv_close_cb_sa(uv_handle_t *handle)
{
struct lws_context *context =
LWS_UV_REFCOUNT_STATIC_HANDLE_TO_CONTEXT(handle);
int n;
lwsl_info("%s: sa left %d: dyn left: %d\n", __func__,
context->count_event_loop_static_asset_handles,
context->count_wsi_allocated);
/* any static assets left? */
if (LWS_UV_REFCOUNT_STATIC_HANDLE_DESTROYED(handle) ||
context->count_wsi_allocated)
return;
/*
* That's it... all wsi were down, and now every
* static asset lws had a UV handle for is down.
*
* Stop the loop so we can get out of here.
*/
for (n = 0; n < context->count_threads; n++) {
struct lws_context_per_thread *pt = &context->pt[n];
if (pt->uv.io_loop && !pt->event_loop_foreign)
uv_stop(pt->uv.io_loop);
}
if (!context->pt[0].event_loop_foreign) {
lwsl_info("%s: calling lws_context_destroy2\n", __func__);
lws_context_destroy2(context);
}
lwsl_info("%s: all done\n", __func__);
}
/*
* These must be called by protocols that want to use libuv objects directly...
*
* .... when the libuv object is created...
*/
LWS_VISIBLE void
lws_libuv_static_refcount_add(uv_handle_t *h, struct lws_context *context)
{
LWS_UV_REFCOUNT_STATIC_HANDLE_NEW(h, context);
}
/*
* ... and in the close callback when the object is closed.
*/
LWS_VISIBLE void
lws_libuv_static_refcount_del(uv_handle_t *h)
{
lws_uv_close_cb_sa(h);
}
static void lws_uv_close_cb(uv_handle_t *handle)
{
}
static void lws_uv_walk_cb(uv_handle_t *handle, void *arg)
{
if (!uv_is_closing(handle))
uv_close(handle, lws_uv_close_cb);
}
LWS_VISIBLE void
lws_close_all_handles_in_loop(uv_loop_t *loop)
{
uv_walk(loop, lws_uv_walk_cb, NULL);
}
LWS_VISIBLE void
lws_libuv_stop_without_kill(const struct lws_context *context, int tsi)
{
if (context->pt[tsi].uv.io_loop)
uv_stop(context->pt[tsi].uv.io_loop);
}
LWS_VISIBLE uv_loop_t *
lws_uv_getloop(struct lws_context *context, int tsi)
{
if (context->pt[tsi].uv.io_loop)
return context->pt[tsi].uv.io_loop;
return NULL;
}
int
lws_libuv_check_watcher_active(struct lws *wsi)
{
uv_handle_t *h = (uv_handle_t *)wsi->w_read.uv.pwatcher;
if (!h)
return 0;
return uv_is_active(h);
}
#if defined(LWS_WITH_PLUGINS) && (UV_VERSION_MAJOR > 0)
int
lws_uv_plugins_init(struct lws_context *context, const char * const *d)
{
struct lws_plugin_capability lcaps;
struct lws_plugin *plugin;
lws_plugin_init_func initfunc;
int m, ret = 0;
void *v;
uv_dirent_t dent;
uv_fs_t req;
char path[256];
uv_lib_t lib;
int pofs = 0;
#if defined(__MINGW32__) || !defined(WIN32)
pofs = 3;
#endif
lib.errmsg = NULL;
lib.handle = NULL;
uv_loop_init(&context->uv.loop);
lwsl_notice(" Plugins:\n");
while (d && *d) {
lwsl_notice(" Scanning %s\n", *d);
m =uv_fs_scandir(&context->uv.loop, &req, *d, 0, NULL);
if (m < 1) {
lwsl_err("Scandir on %s failed\n", *d);
return 1;
}
while (uv_fs_scandir_next(&req, &dent) != UV_EOF) {
if (strlen(dent.name) < 7)
continue;
lwsl_notice(" %s\n", dent.name);
2017-09-23 12:55:21 +08:00
lws_snprintf(path, sizeof(path) - 1, "%s/%s", *d,
dent.name);
if (uv_dlopen(path, &lib)) {
uv_dlerror(&lib);
lwsl_err("Error loading DSO: %s\n", lib.errmsg);
uv_dlclose(&lib);
goto bail;
}
/* we could open it, can we get his init function? */
#if !defined(WIN32) && !defined(__MINGW32__)
m = lws_snprintf(path, sizeof(path) - 1, "init_%s",
dent.name + pofs /* snip lib... */);
path[m - 3] = '\0'; /* snip the .so */
#else
m = lws_snprintf(path, sizeof(path) - 1, "init_%s",
dent.name + pofs);
path[m - 4] = '\0'; /* snip the .dll */
#endif
if (uv_dlsym(&lib, path, &v)) {
uv_dlerror(&lib);
lwsl_err("Failed to get %s on %s: %s", path,
dent.name, lib.errmsg);
uv_dlclose(&lib);
goto bail;
}
initfunc = (lws_plugin_init_func)v;
lcaps.api_magic = LWS_PLUGIN_API_MAGIC;
m = initfunc(context, &lcaps);
if (m) {
2017-09-23 12:55:21 +08:00
lwsl_err("Init %s failed %d\n", dent.name, m);
goto skip;
}
2017-10-04 07:10:39 +08:00
plugin = lws_malloc(sizeof(*plugin), "plugin");
if (!plugin) {
uv_dlclose(&lib);
lwsl_err("OOM\n");
goto bail;
}
plugin->list = context->plugin_list;
context->plugin_list = plugin;
2018-03-12 09:28:26 +08:00
lws_strncpy(plugin->name, dent.name, sizeof(plugin->name));
plugin->lib = lib;
plugin->caps = lcaps;
context->plugin_protocol_count += lcaps.count_protocols;
context->plugin_extension_count += lcaps.count_extensions;
continue;
skip:
uv_dlclose(&lib);
}
bail:
uv_fs_req_cleanup(&req);
d++;
}
return ret;
}
int
lws_uv_plugins_destroy(struct lws_context *context)
{
struct lws_plugin *plugin = context->plugin_list, *p;
lws_plugin_destroy_func func;
char path[256];
2017-09-23 12:55:21 +08:00
int pofs = 0;
void *v;
int m;
#if defined(__MINGW32__) || !defined(WIN32)
pofs = 3;
#endif
if (!plugin)
return 0;
while (plugin) {
p = plugin;
#if !defined(WIN32) && !defined(__MINGW32__)
2017-09-23 12:55:21 +08:00
m = lws_snprintf(path, sizeof(path) - 1, "destroy_%s",
plugin->name + pofs);
path[m - 3] = '\0';
#else
2017-09-23 12:55:21 +08:00
m = lws_snprintf(path, sizeof(path) - 1, "destroy_%s",
plugin->name + pofs);
path[m - 4] = '\0';
#endif
if (uv_dlsym(&plugin->lib, path, &v)) {
uv_dlerror(&plugin->lib);
lwsl_err("Failed to get %s on %s: %s", path,
plugin->name, plugin->lib.errmsg);
} else {
func = (lws_plugin_destroy_func)v;
m = func(context);
if (m)
lwsl_err("Destroying %s failed %d\n",
plugin->name, m);
}
uv_dlclose(&p->lib);
plugin = p->list;
p->list = NULL;
free(p);
}
context->plugin_list = NULL;
while (uv_loop_close(&context->uv.loop))
;
return 0;
}
#endif
static int
elops_init_context_uv(struct lws_context *context,
const struct lws_context_creation_info *info)
{
int n;
context->eventlib_signal_cb = info->signal_cb;
for (n = 0; n < context->count_threads; n++)
context->pt[n].w_sigint.context = context;
return 0;
}
static int
elops_destroy_context1_uv(struct lws_context *context)
{
struct lws_context_per_thread *pt;
int n, m = 0;
for (n = 0; n < context->count_threads; n++) {
int budget = 10000;
pt = &context->pt[n];
/* only for internal loops... */
if (!pt->event_loop_foreign) {
while (budget-- && (m = uv_run(pt->uv.io_loop,
UV_RUN_NOWAIT)))
;
if (m)
lwsl_err("%s: tsi %d: not all closed\n",
__func__, n);
}
}
/* call destroy2 if internal loop */
return !context->pt[0].event_loop_foreign;
}
static int
elops_destroy_context2_uv(struct lws_context *context)
{
struct lws_context_per_thread *pt;
int n, internal = 0;
for (n = 0; n < context->count_threads; n++) {
pt = &context->pt[n];
/* only for internal loops... */
if (!pt->event_loop_foreign && pt->uv.io_loop) {
internal = 1;
if (!context->finalize_destroy_after_internal_loops_stopped)
uv_stop(pt->uv.io_loop);
else {
#if UV_VERSION_MAJOR > 0
uv_loop_close(pt->uv.io_loop);
#endif
lws_free_set_NULL(pt->uv.io_loop);
}
}
}
return internal;
}
static int
elops_wsi_logical_close_uv(struct lws *wsi)
{
if (!lws_socket_is_valid(wsi->desc.sockfd))
return 0;
if (wsi->listener || wsi->event_pipe) {
lwsl_debug("%s: %p: %d %d stop listener / pipe poll\n",
__func__, wsi, wsi->listener, wsi->event_pipe);
2019-03-21 09:27:45 +08:00
if (wsi->w_read.uv.pwatcher)
uv_poll_stop(wsi->w_read.uv.pwatcher);
}
lwsl_debug("%s: lws_libuv_closehandle: wsi %p\n", __func__, wsi);
/*
* libuv has to do his own close handle processing asynchronously
*/
lws_libuv_closehandle(wsi);
return 1; /* do not complete the wsi close, uv close cb will do it */
}
static int
elops_check_client_connect_ok_uv(struct lws *wsi)
{
if (lws_libuv_check_watcher_active(wsi)) {
lwsl_warn("Waiting for libuv watcher to close\n");
return 1;
}
return 0;
}
static void
lws_libuv_closewsi_m(uv_handle_t* handle)
{
lws_sockfd_type sockfd = (lws_sockfd_type)(lws_intptr_t)handle->data;
lwsl_debug("%s: sockfd %d\n", __func__, sockfd);
compatible_close(sockfd);
lws_free(handle);
}
static void
elops_close_handle_manually_uv(struct lws *wsi)
{
uv_handle_t *h = (uv_handle_t *)wsi->w_read.uv.pwatcher;
lwsl_debug("%s: lws_libuv_closehandle: wsi %p\n", __func__, wsi);
/*
* the "manual" variant only closes the handle itself and the
* related fd. handle->data is the fd.
*/
h->data = (void *)(lws_intptr_t)wsi->desc.sockfd;
/*
* We take responsibility to close / destroy these now.
* Remove any trace from the wsi.
*/
wsi->desc.sockfd = LWS_SOCK_INVALID;
wsi->w_read.uv.pwatcher = NULL;
wsi->told_event_loop_closed = 1;
uv_close(h, lws_libuv_closewsi_m);
}
static int
elops_accept_uv(struct lws *wsi)
{
struct lws_context_per_thread *pt = &wsi->context->pt[(int)wsi->tsi];
wsi->w_read.context = wsi->context;
wsi->w_read.uv.pwatcher =
lws_malloc(sizeof(*wsi->w_read.uv.pwatcher), "uvh");
if (!wsi->w_read.uv.pwatcher)
return -1;
if (wsi->role_ops->file_handle)
uv_poll_init(pt->uv.io_loop, wsi->w_read.uv.pwatcher,
(int)(long long)wsi->desc.filefd);
else
uv_poll_init_socket(pt->uv.io_loop,
2018-11-23 08:47:56 +08:00
wsi->w_read.uv.pwatcher,
wsi->desc.sockfd);
((uv_handle_t *)wsi->w_read.uv.pwatcher)->data = (void *)wsi;
return 0;
}
static void
elops_io_uv(struct lws *wsi, int flags)
{
struct lws_context_per_thread *pt = &wsi->context->pt[(int)wsi->tsi];
struct lws_io_watcher *w = &wsi->w_read;
int current_events = w->actual_events & (UV_READABLE | UV_WRITABLE);
lwsl_debug("%s: %p: %d\n", __func__, wsi, flags);
/* w->context is set after the loop is initialized */
if (!pt->uv.io_loop || !w->context) {
lwsl_info("%s: no io loop yet\n", __func__);
return;
}
if (!((flags & (LWS_EV_START | LWS_EV_STOP)) &&
(flags & (LWS_EV_READ | LWS_EV_WRITE)))) {
lwsl_err("%s: assert: flags %d", __func__, flags);
assert(0);
}
if (!w->uv.pwatcher || wsi->told_event_loop_closed) {
lwsl_err("%s: no watcher\n", __func__);
return;
}
if (flags & LWS_EV_START) {
if (flags & LWS_EV_WRITE)
current_events |= UV_WRITABLE;
if (flags & LWS_EV_READ)
current_events |= UV_READABLE;
uv_poll_start(w->uv.pwatcher, current_events, lws_io_cb);
} else {
if (flags & LWS_EV_WRITE)
current_events &= ~UV_WRITABLE;
if (flags & LWS_EV_READ)
current_events &= ~UV_READABLE;
if (!(current_events & (UV_READABLE | UV_WRITABLE)))
uv_poll_stop(w->uv.pwatcher);
else
uv_poll_start(w->uv.pwatcher, current_events,
lws_io_cb);
}
w->actual_events = current_events;
}
static int
elops_init_vhost_listen_wsi_uv(struct lws *wsi)
{
struct lws_context_per_thread *pt;
int n;
if (!wsi)
return 0;
if (wsi->w_read.context)
return 0;
pt = &wsi->context->pt[(int)wsi->tsi];
if (!pt->uv.io_loop)
return 0;
wsi->w_read.context = wsi->context;
wsi->w_read.uv.pwatcher =
lws_malloc(sizeof(*wsi->w_read.uv.pwatcher), "uvh");
if (!wsi->w_read.uv.pwatcher)
return -1;
n = uv_poll_init_socket(pt->uv.io_loop, wsi->w_read.uv.pwatcher,
wsi->desc.sockfd);
if (n) {
lwsl_err("uv_poll_init failed %d, sockfd=%p\n", n,
(void *)(lws_intptr_t)wsi->desc.sockfd);
return -1;
}
((uv_handle_t *)wsi->w_read.uv.pwatcher)->data = (void *)wsi;
elops_io_uv(wsi, LWS_EV_START | LWS_EV_READ);
return 0;
}
static void
elops_run_pt_uv(struct lws_context *context, int tsi)
{
if (context->pt[tsi].uv.io_loop)
uv_run(context->pt[tsi].uv.io_loop, 0);
}
static void
elops_destroy_pt_uv(struct lws_context *context, int tsi)
{
struct lws_context_per_thread *pt = &context->pt[tsi];
int m, ns;
lwsl_info("%s: %d\n", __func__, tsi);
if (!lws_check_opt(context->options, LWS_SERVER_OPTION_LIBUV))
return;
if (!pt->uv.io_loop)
return;
if (pt->event_loop_destroy_processing_done)
return;
pt->event_loop_destroy_processing_done = 1;
if (!pt->event_loop_foreign) {
uv_signal_stop(&pt->w_sigint.uv.watcher);
ns = LWS_ARRAY_SIZE(sigs);
if (lws_check_opt(context->options,
LWS_SERVER_OPTION_UV_NO_SIGSEGV_SIGFPE_SPIN))
ns = 2;
for (m = 0; m < ns; m++) {
uv_signal_stop(&pt->uv.signals[m]);
uv_close((uv_handle_t *)&pt->uv.signals[m],
lws_uv_close_cb_sa);
}
} else
lwsl_debug("%s: not closing pt signals\n", __func__);
uv_timer_stop(&pt->uv.timeout_watcher);
uv_close((uv_handle_t *)&pt->uv.timeout_watcher, lws_uv_close_cb_sa);
uv_timer_stop(&pt->uv.hrtimer);
uv_close((uv_handle_t *)&pt->uv.hrtimer, lws_uv_close_cb_sa);
uv_idle_stop(&pt->uv.idle);
uv_close((uv_handle_t *)&pt->uv.idle, lws_uv_close_cb_sa);
}
/*
* This needs to be called after vhosts have been defined.
*
* If later, after server start, another vhost is added, this must be
* called again to bind the vhost
*/
LWS_VISIBLE int
elops_init_pt_uv(struct lws_context *context, void *_loop, int tsi)
{
struct lws_context_per_thread *pt = &context->pt[tsi];
struct lws_vhost *vh = context->vhost_list;
int status = 0, n, ns, first = 1;
uv_loop_t *loop = (uv_loop_t *)_loop;
if (!pt->uv.io_loop) {
if (!loop) {
loop = lws_malloc(sizeof(*loop), "libuv loop");
if (!loop) {
lwsl_err("OOM\n");
return -1;
}
#if UV_VERSION_MAJOR > 0
uv_loop_init(loop);
#else
lwsl_err("This libuv is too old to work...\n");
return 1;
#endif
pt->event_loop_foreign = 0;
} else {
lwsl_notice(" Using foreign event loop...\n");
pt->event_loop_foreign = 1;
}
pt->uv.io_loop = loop;
uv_idle_init(loop, &pt->uv.idle);
LWS_UV_REFCOUNT_STATIC_HANDLE_NEW(&pt->uv.idle, context);
ns = LWS_ARRAY_SIZE(sigs);
if (lws_check_opt(context->options,
LWS_SERVER_OPTION_UV_NO_SIGSEGV_SIGFPE_SPIN))
ns = 2;
if (!pt->event_loop_foreign) {
assert(ns <= (int)LWS_ARRAY_SIZE(pt->uv.signals));
for (n = 0; n < ns; n++) {
uv_signal_init(loop, &pt->uv.signals[n]);
LWS_UV_REFCOUNT_STATIC_HANDLE_NEW(&pt->uv.signals[n],
context);
pt->uv.signals[n].data = pt->context;
uv_signal_start(&pt->uv.signals[n],
lws_uv_signal_handler, sigs[n]);
}
}
} else
first = 0;
/*
* Initialize the accept wsi read watcher with all the listening sockets
* and register a callback for read operations
*
* We have to do it here because the uv loop(s) are not
* initialized until after context creation.
*/
while (vh) {
if (elops_init_vhost_listen_wsi_uv(vh->lserv_wsi) == -1)
return -1;
vh = vh->vhost_next;
}
if (!first)
return status;
uv_timer_init(pt->uv.io_loop, &pt->uv.timeout_watcher);
LWS_UV_REFCOUNT_STATIC_HANDLE_NEW(&pt->uv.timeout_watcher, context);
uv_timer_start(&pt->uv.timeout_watcher, lws_uv_timeout_cb, 10, 1000);
uv_timer_init(pt->uv.io_loop, &pt->uv.hrtimer);
LWS_UV_REFCOUNT_STATIC_HANDLE_NEW(&pt->uv.hrtimer, context);
return status;
}
static void
lws_libuv_closewsi(uv_handle_t* handle)
{
struct lws *wsi = (struct lws *)handle->data;
struct lws_context *context = lws_get_context(wsi);
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
#if !defined(LWS_WITHOUT_SERVER)
int lspd = 0;
#endif
lwsl_info("%s: %p\n", __func__, wsi);
/*
* We get called back here for every wsi that closes
*/
#if !defined(LWS_WITHOUT_SERVER)
if (wsi->role_ops == &role_ops_listen && wsi->context->deprecated) {
lspd = 1;
context->deprecation_pending_listen_close_count--;
if (!context->deprecation_pending_listen_close_count)
lspd = 2;
}
#endif
lws_pt_lock(pt, __func__);
__lws_close_free_wsi_final(wsi);
lws_pt_unlock(pt);
/* it's our job to close the handle finally */
lws_free(handle);
#if !defined(LWS_WITHOUT_SERVER)
if (lspd == 2 && context->deprecation_cb) {
lwsl_notice("calling deprecation callback\n");
context->deprecation_cb();
}
#endif
lwsl_info("%s: sa left %d: dyn left: %d (rk %d)\n", __func__,
context->count_event_loop_static_asset_handles,
context->count_wsi_allocated, context->requested_kill);
/*
* eventually, we closed all the wsi...
*/
if (context->requested_kill && !context->count_wsi_allocated) {
struct lws_vhost *vh = context->vhost_list;
int m;
/*
* Start Closing Phase 2: close of static handles
*/
lwsl_info("%s: all lws dynamic handles down, closing static\n",
__func__);
for (m = 0; m < context->count_threads; m++)
elops_destroy_pt_uv(context, m);
/* protocols may have initialized libuv objects */
while (vh) {
lws_vhost_destroy1(vh);
vh = vh->vhost_next;
}
if (!context->count_event_loop_static_asset_handles &&
context->pt[0].event_loop_foreign) {
lwsl_info("%s: call lws_context_destroy2\n", __func__);
lws_context_destroy2(context);
}
}
}
void
lws_libuv_closehandle(struct lws *wsi)
{
uv_handle_t* handle;
if (!wsi->w_read.uv.pwatcher)
return;
if (wsi->told_event_loop_closed) {
// assert(0);
return;
}
lwsl_debug("%s: %p\n", __func__, wsi);
wsi->told_event_loop_closed = 1;
/*
* The normal close path attaches the related wsi as the
* handle->data.
*/
handle = (uv_handle_t *)wsi->w_read.uv.pwatcher;
/* ensure we can only do this once */
wsi->w_read.uv.pwatcher = NULL;
uv_close(handle, lws_libuv_closewsi);
}
struct lws_event_loop_ops event_loop_ops_uv = {
/* name */ "libuv",
/* init_context */ elops_init_context_uv,
/* destroy_context1 */ elops_destroy_context1_uv,
/* destroy_context2 */ elops_destroy_context2_uv,
/* init_vhost_listen_wsi */ elops_init_vhost_listen_wsi_uv,
/* init_pt */ elops_init_pt_uv,
/* wsi_logical_close */ elops_wsi_logical_close_uv,
/* check_client_connect_ok */ elops_check_client_connect_ok_uv,
/* close_handle_manually */ elops_close_handle_manually_uv,
/* accept */ elops_accept_uv,
/* io */ elops_io_uv,
/* run_pt */ elops_run_pt_uv,
/* destroy_pt */ elops_destroy_pt_uv,
/* destroy wsi */ NULL,
/* periodic_events_available */ 0,
};