Swoole 源码分析——Server模块之Worker事件循环
swManager_loop
函数 manager
进程管理
-
manager
进程开启的时候,首先要调用onManagerStart
回调 - 添加信号处理函数
swSignal_add
,SIGTERM
用于结束server
,只需要running
设置为 0,manager
会逐个杀死worker
进程;SIGUSR1
用于重载所有的worker
进程;SIGUSR2
用于重载所有的task_worker
进程;SIGIO
用于重启已经关闭了的worker
进程;SIGALRM
用于检测所有的超时请求; - 如果设置了
serv->manager_alarm
,那么就是开启了超时请求的监控,此时需要设置alarm
信号,让manager
进程定时去检查是否有超时的请求。 - 如果
running
为 1,就不断while
循环,杀死或者启动相应的worker
进程,如果running
为 0,那么就关闭所有的worker
进程,调用onManagerStop
函数退出程序。 -
调用
wait
函数,监控已结束的worker
进程- 如果
wait
函数返回异常,很有可能是被信号打断。此时需要先检查ManagerProcess.read_message
,如果是 1,那么说明wait
函数被SIGIO
信号打断,该信号由worker
进程发送,用于告知manager
进程该worker
进程即将关闭。此时,需要manager
进程重新开启worker
进程。 - 如果
ManagerProcess.alarm
为 1,那么说明wait
函数由SIGALRM
信号打断,此时需要检查超时的请求。erv->hooks[SW_SERVER_HOOK_MANAGER_TIMER]
也就是php_swoole_trace_check
是检查慢请求的函数。 - 如果
ManagerProcess.reload_all_worker
为 1,那么wait
函数由SIGUSR1
打断,此时应该重启所有的worker
进程 - 如果
ManagerProcess.reload_task_worker
为 1,那么wait
函数由SIGUSR2
打断,此时应该重启所有的task_worker
进程 - 如果
wait
返回值正常,那么就要从serv->workers
、serv->gs->task_workers
、serv->user_worker
中寻找退出的worker
进程。如果该进程是STOPPED
状态,说明很有可能是调试状态,此时不需要重启,只需要调用tracer
函数
- 如果
static void swManager_signal_handle(int sig)
{
switch (sig)
{
case SIGTERM:
SwooleG.running = 0;
break;
/**
* reload all workers
*/
case SIGUSR1:
if (ManagerProcess.reloading == 0)
{
ManagerProcess.reloading = 1;
ManagerProcess.reload_all_worker = 1;
}
break;
/**
* only reload task workers
*/
case SIGUSR2:
if (ManagerProcess.reloading == 0)
{
ManagerProcess.reloading = 1;
ManagerProcess.reload_task_worker = 1;
}
break;
case SIGIO:
ManagerProcess.read_message = 1;
break;
case SIGALRM:
ManagerProcess.alarm = 1;
break;
default:
#ifdef SIGRTMIN
if (sig == SIGRTMIN)
{
swServer_reopen_log_file(SwooleG.serv);
}
#endif
break;
}
}
static int swManager_loop(swFactory *factory)
{
int pid, new_pid;
int i;
int reload_worker_i = 0;
int reload_worker_num;
int reload_init = 0;
pid_t reload_worker_pid = 0;
int status;
SwooleG.use_signalfd = 0;
SwooleG.use_timerfd = 0;
memset(&ManagerProcess, 0, sizeof(ManagerProcess));
swServer *serv = factory->ptr;
swWorker *reload_workers;
if (serv->hooks[SW_SERVER_HOOK_MANAGER_START])
{
swServer_call_hook(serv, SW_SERVER_HOOK_MANAGER_START, serv);
}
if (serv->onManagerStart)
{
serv->onManagerStart(serv);
}
reload_worker_num = serv->worker_num + serv->task_worker_num;
reload_workers = sw_calloc(reload_worker_num, sizeof(swWorker));
if (reload_workers == NULL)
{
swError("malloc[reload_workers] failed");
return SW_ERR;
}
//for reload
swSignal_add(SIGHUP, NULL);
swSignal_add(SIGTERM, swManager_signal_handle);
swSignal_add(SIGUSR1, swManager_signal_handle);
swSignal_add(SIGUSR2, swManager_signal_handle);
swSignal_add(SIGIO, swManager_signal_handle);
#ifdef SIGRTMIN
swSignal_add(SIGRTMIN, swManager_signal_handle);
#endif
//swSignal_add(SIGINT, swManager_signal_handle);
if (serv->manager_alarm > 0)
{
alarm(serv->manager_alarm);
swSignal_add(SIGALRM, swManager_signal_handle);
}
SwooleG.main_reactor = NULL;
while (SwooleG.running > 0)
{
_wait: pid = wait(&status);
if (ManagerProcess.read_message)
{
swWorkerStopMessage msg;
while (swChannel_pop(serv->message_box, &msg, sizeof(msg)) > 0)
{
if (SwooleG.running == 0)
{
continue;
}
pid_t new_pid = swManager_spawn_worker(factory, msg.worker_id);
if (new_pid > 0)
{
serv->workers[msg.worker_id].pid = new_pid;
}
}
ManagerProcess.read_message = 0;
}
if (pid < 0)
{
if (ManagerProcess.alarm == 1)
{
ManagerProcess.alarm = 0;
alarm(serv->manager_alarm);
if (serv->hooks[SW_SERVER_HOOK_MANAGER_TIMER])
{
swServer_call_hook(serv, SW_SERVER_HOOK_MANAGER_TIMER, serv);
}
}
if (ManagerProcess.reloading == 0)
{
error: if (errno != EINTR)
{
swSysError("wait() failed.");
}
continue;
}
//reload task & event workers
else if (ManagerProcess.reload_all_worker == 1)
{
swNotice("Server is reloading now.");
if (reload_init == 0)
{
reload_init = 1;
memcpy(reload_workers, serv->workers, sizeof(swWorker) * serv->worker_num);
reload_worker_num = serv->worker_num;
if (serv->task_worker_num > 0)
{
memcpy(reload_workers + serv->worker_num, serv->gs->task_workers.workers,
sizeof(swWorker) * serv->task_worker_num);
reload_worker_num += serv->task_worker_num;
}
ManagerProcess.reload_all_worker = 0;
if (serv->reload_async)
{
for (i = 0; i < serv->worker_num; i++)
{
if (kill(reload_workers[i].pid, SIGTERM) < 0)
{
swSysError("kill(%d, SIGTERM) [%d] failed.", reload_workers[i].pid, i);
}
}
reload_worker_i = serv->worker_num;
}
else
{
reload_worker_i = 0;
}
}
goto kill_worker;
}
//only reload task workers
else if (ManagerProcess.reload_task_worker == 1)
{
if (serv->task_worker_num == 0)
{
swWarn("cannot reload task workers, task workers is not started.");
continue;
}
swNotice("Server is reloading now.");
if (reload_init == 0)
{
memcpy(reload_workers, serv->gs->task_workers.workers, sizeof(swWorker) * serv->task_worker_num);
reload_worker_num = serv->task_worker_num;
reload_worker_i = 0;
reload_init = 1;
ManagerProcess.reload_task_worker = 0;
}
goto kill_worker;
}
else
{
goto error;
}
}
if (SwooleG.running == 1)
{
//event workers
for (i = 0; i < serv->worker_num; i++)
{
//compare PID
if (pid != serv->workers[i].pid)
{
continue;
}
if (WIFSTOPPED(status) && serv->workers[i].tracer)
{
serv->workers[i].tracer(&serv->workers[i]);
serv->workers[i].tracer = NULL;
goto _wait;
}
//Check the process return code and signal
swManager_check_exit_status(serv, i, pid, status);
while (1)
{
new_pid = swManager_spawn_worker(factory, i);
if (new_pid < 0)
{
usleep(100000);
continue;
}
else
{
serv->workers[i].pid = new_pid;
break;
}
}
}
swWorker *exit_worker;
//task worker
if (serv->gs->task_workers.map)
{
exit_worker = swHashMap_find_int(serv->gs->task_workers.map, pid);
if (exit_worker != NULL)
{
if (WIFSTOPPED(status) && exit_worker->tracer)
{
exit_worker->tracer(exit_worker);
exit_worker->tracer = NULL;
goto _wait;
}
swManager_check_exit_status(serv, exit_worker->id, pid, status);
swProcessPool_spawn(&serv->gs->task_workers, exit_worker);
}
}
//user process
if (serv->user_worker_map != NULL)
{
swManager_wait_user_worker(&serv->gs->event_workers, pid, status);
}
if (pid == reload_worker_pid)
{
reload_worker_i++;
}
}
//reload worker
kill_worker: if (ManagerProcess.reloading == 1)
{
//reload finish
if (reload_worker_i >= reload_worker_num)
{
reload_worker_pid = reload_worker_i = reload_init = ManagerProcess.reloading = 0;
continue;
}
reload_worker_pid = reload_workers[reload_worker_i].pid;
if (kill(reload_worker_pid, SIGTERM) < 0)
{
if (errno == ECHILD)
{
reload_worker_i++;
goto kill_worker;
}
swSysError("kill(%d, SIGTERM) [%d] failed.", reload_workers[reload_worker_i].pid, reload_worker_i);
}
}
}
sw_free(reload_workers);
swSignal_none();
//kill all child process
for (i = 0; i < serv->worker_num; i++)
{
swTrace("[Manager]kill worker processor");
kill(serv->workers[i].pid, SIGTERM);
}
//kill and wait task process
if (serv->task_worker_num > 0)
{
swProcessPool_shutdown(&serv->gs->task_workers);
}
//wait child process
for (i = 0; i < serv->worker_num; i++)
{
if (swWaitpid(serv->workers[i].pid, &status, 0) < 0)
{
swSysError("waitpid(%d) failed.", serv->workers[i].pid);
}
}
//kill all user process
if (serv->user_worker_map)
{
swManager_kill_user_worker(serv);
}
if (serv->onManagerStop)
{
serv->onManagerStop(serv);
}
return SW_OK;
}
void php_swoole_trace_check(void *arg)
{
swServer *serv = (swServer *) arg;
uint8_t timeout = serv->request_slowlog_timeout;
int count = serv->worker_num + serv->task_worker_num;
int i = serv->trace_event_worker ? 0 : serv->worker_num;
swWorker *worker;
for (; i < count; i++)
{
worker = swServer_get_worker(serv, i);
swTraceLog(SW_TRACE_SERVER, "trace request, worker#%d, pid=%d. request_time=%d.", i, worker->pid, worker->request_time);
if (!(worker->request_time > 0 && worker->traced == 0 && serv->gs->now - worker->request_time >= timeout))
{
continue;
}
if (ptrace(PTRACE_ATTACH, worker->pid, 0, 0) < 0)
{
swSysError("failed to ptrace(ATTACH, %d) worker#%d,", worker->pid, worker->id);
continue;
}
worker->tracer = trace_request;
worker->traced = 1;
}
}
static void swManager_check_exit_status(swServer *serv, int worker_id, pid_t pid, int status)
{
if (status != 0)
{
swWarn("worker#%d abnormal exit, status=%d, signal=%d", worker_id, WEXITSTATUS(status), WTERMSIG(status));
if (serv->onWorkerError != NULL)
{
serv->onWorkerError(serv, worker_id, pid, WEXITSTATUS(status), WTERMSIG(status));
}
}
}
swWorker_loop
函数 worker
事件循环
-
worker
进程的事件循环和reactor
线程类似,都是创建reactor
对象,然后调用SwooleG.main_reactor->wait
函数进行事件循环,不同的是worker
进程监控的是pipe_worker
这个socket
。 - 如果
worker
的dispatch_mode
是stream
,reactor
还要监听serv->stream_fd
,以便可以更加高效的消费reactor
线程发送的数据 -
swServer_worker_init
函数用于初始化worker
进程,swWorker_onStart
用于调用回调函数,swWorker_onStop
用于停止worker
进程
int swWorker_loop(swFactory *factory, int worker_id)
{
swServer *serv = factory->ptr;
#ifndef SW_WORKER_USE_SIGNALFD
SwooleG.use_signalfd = 0;
#elif defined(HAVE_SIGNALFD)
SwooleG.use_signalfd = 1;
#endif
//timerfd
#ifdef HAVE_TIMERFD
SwooleG.use_timerfd = 1;
#endif
//worker_id
SwooleWG.id = worker_id;
SwooleG.pid = getpid();
swWorker *worker = swServer_get_worker(serv, worker_id);
swServer_worker_init(serv, worker);
SwooleG.main_reactor = sw_malloc(sizeof(swReactor));
if (SwooleG.main_reactor == NULL)
{
swError("[Worker] malloc for reactor failed.");
return SW_ERR;
}
if (swReactor_create(SwooleG.main_reactor, SW_REACTOR_MAXEVENTS) < 0)
{
swError("[Worker] create worker_reactor failed.");
return SW_ERR;
}
worker->status = SW_WORKER_IDLE;
int pipe_worker = worker->pipe_worker;
swSetNonBlock(pipe_worker);
SwooleG.main_reactor->ptr = serv;
SwooleG.main_reactor->add(SwooleG.main_reactor, pipe_worker, SW_FD_PIPE | SW_EVENT_READ);
SwooleG.main_reactor->setHandle(SwooleG.main_reactor, SW_FD_PIPE, swWorker_onPipeReceive);
SwooleG.main_reactor->setHandle(SwooleG.main_reactor, SW_FD_WRITE, swReactor_onWrite);
/**
* set pipe buffer size
*/
int i;
swConnection *pipe_socket;
for (i = 0; i < serv->worker_num + serv->task_worker_num; i++)
{
worker = swServer_get_worker(serv, i);
pipe_socket = swReactor_get(SwooleG.main_reactor, worker->pipe_master);
pipe_socket->buffer_size = SW_MAX_INT;
pipe_socket = swReactor_get(SwooleG.main_reactor, worker->pipe_worker);
pipe_socket->buffer_size = SW_MAX_INT;
}
if (serv->dispatch_mode == SW_DISPATCH_STREAM)
{
SwooleG.main_reactor->add(SwooleG.main_reactor, serv->stream_fd, SW_FD_LISTEN | SW_EVENT_READ);
SwooleG.main_reactor->setHandle(SwooleG.main_reactor, SW_FD_LISTEN, swWorker_onStreamAccept);
SwooleG.main_reactor->setHandle(SwooleG.main_reactor, SW_FD_STREAM, swWorker_onStreamRead);
swStream_set_protocol(&serv->stream_protocol);
serv->stream_protocol.package_max_length = SW_MAX_INT;
serv->stream_protocol.onPackage = swWorker_onStreamPackage;
serv->buffer_pool = swLinkedList_new(0, NULL);
}
swWorker_onStart(serv);
#ifdef HAVE_SIGNALFD
if (SwooleG.use_signalfd)
{
swSignalfd_setup(SwooleG.main_reactor);
}
#endif
//main loop
SwooleG.main_reactor->wait(SwooleG.main_reactor, NULL);
//clear pipe buffer
swWorker_clean();
//worker shutdown
swWorker_onStop(serv);
return SW_OK;
}
swServer_worker_init
初始化函数
- 与
reactor
线程一样,首先如果设置了CPU
亲和度的话,要将worker
进程绑定到特定的CPU
上,指定CPU
的方法仍然是SwooleWG.id % serv->cpu_affinity_available_num
,这样可以保证对应的reactor
线程和worker
进程在同一个CPU
核上 -
swWorker_signal_init
用于设置worker
进程的信号处理函数:SIGTERM
信号用于关闭当前worker
进程;SIGALRM
代表定时任务。 -
buffer_input
用于存储来源于reactor
线程发送的数据,是一个serv->reactor_num + serv->dgram_port_num
大小的数组。
void swWorker_signal_init(void)
{
swSignal_clear();
/**
* use user settings
*/
SwooleG.use_signalfd = SwooleG.enable_signalfd;
swSignal_add(SIGHUP, NULL);
swSignal_add(SIGPIPE, NULL);
swSignal_add(SIGUSR1, NULL);
swSignal_add(SIGUSR2, NULL);
swSignal_add(SIGTERM, swWorker_signal_handler);
swSignal_add(SIGALRM, swSystemTimer_signal_handler);
//for test
swSignal_add(SIGVTALRM, swWorker_signal_handler);
#ifdef SIGRTMIN
swSignal_add(SIGRTMIN, swWorker_signal_handler);
#endif
}
int swServer_worker_init(swServer *serv, swWorker *worker)
{
#ifdef HAVE_CPU_AFFINITY
if (serv->open_cpu_affinity)
{
cpu_set_t cpu_set;
CPU_ZERO(&cpu_set);
if (serv->cpu_affinity_available_num)
{
CPU_SET(serv->cpu_affinity_available[SwooleWG.id % serv->cpu_affinity_available_num], &cpu_set);
}
else
{
CPU_SET(SwooleWG.id % SW_CPU_NUM, &cpu_set);
}
#ifdef __FreeBSD__
if (cpuset_setaffinity(CPU_LEVEL_WHICH, CPU_WHICH_PID, -1,
sizeof(cpu_set), &cpu_set) < 0)
#else
if (sched_setaffinity(getpid(), sizeof(cpu_set), &cpu_set) < 0)
#endif
{
swSysError("sched_setaffinity() failed.");
}
}
#endif
//signal init
swWorker_signal_init();
SwooleWG.buffer_input = swServer_create_worker_buffer(serv);
if (!SwooleWG.buffer_input)
{
return SW_ERR;
}
if (serv->max_request < 1)
{
SwooleWG.run_always = 1;
}
else
{
SwooleWG.max_request = serv->max_request;
if (SwooleWG.max_request > 10)
{
int n = swoole_system_random(1, SwooleWG.max_request / 2);
if (n > 0)
{
SwooleWG.max_request += n;
}
}
}
worker->start_time = serv->gs->now;
worker->request_time = 0;
worker->request_count = 0;
return SW_OK;
}
swString** swServer_create_worker_buffer(swServer *serv)
{
int i;
int buffer_num;
if (serv->factory_mode == SW_MODE_SINGLE)
{
buffer_num = 1;
}
else
{
buffer_num = serv->reactor_num + serv->dgram_port_num;
}
swString **buffers = sw_malloc(sizeof(swString*) * buffer_num);
if (buffers == NULL)
{
swError("malloc for worker buffer_input failed.");
return NULL;
}
for (i = 0; i < buffer_num; i++)
{
buffers[i] = swString_new(SW_BUFFER_SIZE_BIG);
if (buffers[i] == NULL)
{
swError("worker buffer_input init failed.");
return NULL;
}
}
return buffers;
}
swWorker_onStart
进程启动
-
swWorker_onStart
函数将其他的worker
进程所占内存全部释放 - 设定当前
worker
的状态为SW_WORKER_IDLE
空闲 - 如果用户更改了
worker
进程的用户与组、进行了重定向根目录,那么我们还要调用setgid
、setuid
、chroot
函数进行相应设置
void swWorker_onStart(swServer *serv)
{
/**
* Release other worker process
*/
swWorker *worker;
if (SwooleWG.id >= serv->worker_num)
{
SwooleG.process_type = SW_PROCESS_TASKWORKER;
}
else
{
SwooleG.process_type = SW_PROCESS_WORKER;
}
int is_root = !geteuid();
struct passwd *passwd = NULL;
struct group *group = NULL;
if (is_root)
{
//get group info
if (SwooleG.group)
{
group = getgrnam(SwooleG.group);
if (!group)
{
swWarn("get group [%s] info failed.", SwooleG.group);
}
}
//get user info
if (SwooleG.user)
{
passwd = getpwnam(SwooleG.user);
if (!passwd)
{
swWarn("get user [%s] info failed.", SwooleG.user);
}
}
//chroot
if (SwooleG.chroot)
{
if (0 > chroot(SwooleG.chroot))
{
swSysError("chroot to [%s] failed.", SwooleG.chroot);
}
}
//set process group
if (SwooleG.group && group)
{
if (setgid(group->gr_gid) < 0)
{
swSysError("setgid to [%s] failed.", SwooleG.group);
}
}
//set process user
if (SwooleG.user && passwd)
{
if (setuid(passwd->pw_uid) < 0)
{
swSysError("setuid to [%s] failed.", SwooleG.user);
}
}
}
SwooleWG.worker = swServer_get_worker(serv, SwooleWG.id);
int i;
for (i = 0; i < serv->worker_num + serv->task_worker_num; i++)
{
worker = swServer_get_worker(serv, i);
if (SwooleWG.id == i)
{
continue;
}
else
{
swWorker_free(worker);
}
if (swIsWorker())
{
swSetNonBlock(worker->pipe_master);
}
}
SwooleWG.worker->status = SW_WORKER_IDLE;
sw_shm_protect(serv->session_list, PROT_READ);
if (serv->onWorkerStart)
{
serv->onWorkerStart(serv, SwooleWG.id);
}
}
swWorker_stop
关闭 worker
进程
-
reload_async
设置为 1 后,并不会立刻停止worker
进程,如果reactor
当中还有待监听的事件,reactor
仍然可以继续循环;与此同时,worker
进程设置了一个超时时间,超过时间后,立刻关闭事件循环。 - 为了通知
manager
进程重启worker
进程,需要调用swChannel_push
函数更新message_box
- 从
reactor
中删除对pipe_worker
、stream_fd
的事件监控 -
swWorker_try_to_exit
用于判断当前worker
进程中reactor
是否还有待监听事件,如果没有可以立刻停止进程
static sw_inline int swReactor_remove_read_event(swReactor *reactor, int fd)
{
swConnection *conn = swReactor_get(reactor, fd);
if (conn->events & SW_EVENT_WRITE)
{
conn->events &= (~SW_EVENT_READ);
return reactor->set(reactor, fd, conn->fdtype | conn->events);
}
else
{
return reactor->del(reactor, fd);
}
}
static void swWorker_stop()
{
swWorker *worker = SwooleWG.worker;
swServer *serv = SwooleG.serv;
worker->status = SW_WORKER_BUSY;
/**
* force to end
*/
if (serv->reload_async == 0)
{
SwooleG.running = 0;
SwooleG.main_reactor->running = 0;
return;
}
//The worker process is shutting down now.
if (SwooleWG.wait_exit)
{
return;
}
//remove read event
if (worker->pipe_worker)
{
swReactor_remove_read_event(SwooleG.main_reactor, worker->pipe_worker);
}
if (serv->stream_fd > 0)
{
SwooleG.main_reactor->del(SwooleG.main_reactor, serv->stream_fd);
close(serv->stream_fd);
serv->stream_fd = 0;
}
if (serv->onWorkerStop)
{
serv->onWorkerStop(serv, SwooleWG.id);
serv->onWorkerStop = NULL;
}
swWorkerStopMessage msg;
msg.pid = SwooleG.pid;
msg.worker_id = SwooleWG.id;
//send message to manager
if (swChannel_push(SwooleG.serv->message_box, &msg, sizeof(msg)) < 0)
{
SwooleG.running = 0;
}
else
{
kill(serv->gs->manager_pid, SIGIO);
}
try_to_exit: SwooleWG.wait_exit = 1;
if (SwooleG.timer.fd == 0)
{
swTimer_init(serv->max_wait_time * 1000);
}
SwooleG.timer.add(&SwooleG.timer, serv->max_wait_time * 1000, 0, NULL, swWorker_onTimeout);
swWorker_try_to_exit();
}
static void swWorker_onTimeout(swTimer *timer, swTimer_node *tnode)
{
SwooleG.running = 0;
SwooleG.main_reactor->running = 0;
swoole_error_log(SW_LOG_WARNING, SW_ERROR_SERVER_WORKER_EXIT_TIMEOUT, "worker exit timeout, forced to terminate.");
}
void swWorker_try_to_exit()
{
swServer *serv = SwooleG.serv;
int expect_event_num = SwooleG.use_signalfd ? 1 : 0;
uint8_t call_worker_exit_func = 0;
while (1)
{
if (SwooleG.main_reactor->event_num == expect_event_num)
{
SwooleG.main_reactor->running = 0;
SwooleG.running = 0;
}
else
{
if (serv->onWorkerExit && call_worker_exit_func == 0)
{
serv->onWorkerExit(serv, SwooleWG.id);
call_worker_exit_func = 1;
continue;
}
}
break;
}
}
swWorker_onPipeReceive
接受数据
接受数据的时候,如果类型是 SW_EVENT_PACKAGE_START
,说明后续还有数据,需要将数据合并在一起接受。
static int swWorker_onPipeReceive(swReactor *reactor, swEvent *event)
{
swEventData task;
swServer *serv = reactor->ptr;
swFactory *factory = &serv->factory;
int ret;
read_from_pipe:
if (read(event->fd, &task, sizeof(task)) > 0)
{
ret = swWorker_onTask(factory, &task);
/**
* Big package
*/
if (task.info.type == SW_EVENT_PACKAGE_START)
{
//no data
if (ret < 0 && errno == EAGAIN)
{
return SW_OK;
}
else if (ret > 0)
{
goto read_from_pipe;
}
}
return ret;
}
return SW_ERR;
}
swWorker_onTask
函数处理数据
-
worker
接受的消息数据类型有多种,最常用的是SW_EVENT_TCP
、SW_EVENT_PACKAGE
、SW_EVENT_PACKAGE_START
、SW_EVENT_PACKAGE_END
- 如果数据类型是
SW_EVENT_TCP
、SW_EVENT_PACKAGE
,首先要调用swWorker_discard_data
函数观察数据是否有效,接着利用onReceive
函数接受ring_buff
数据并且调用回调函数 - 如果数据是
SW_EVENT_PACKAGE_START
、SW_EVENT_PACKAGE_END
,会将数据存储在SwooleWG.buffer_input
中去。最后调用serv->onReceive
-
SW_EVENT_CONNECT
事件由接受连接时触发 -
SW_EVENT_BUFFER_FULL
、SW_EVENT_BUFFER_EMPTY
事件是连接中客户端数据过多worker
无法及时消费触发 -
SW_EVENT_FINISH
由task_worker
完成任务触发 -
SW_EVENT_PIPE_MESSAGE
由发送任务给task_worker
触发
int swWorker_onTask(swFactory *factory, swEventData *task)
{
swServer *serv = factory->ptr;
swString *package = NULL;
swDgramPacket *header;
#ifdef SW_USE_OPENSSL
swConnection *conn;
#endif
factory->last_from_id = task->info.from_id;
serv->last_session_id = task->info.fd;
swWorker *worker = SwooleWG.worker;
//worker busy
worker->status = SW_WORKER_BUSY;
switch (task->info.type)
{
//no buffer
case SW_EVENT_TCP:
//ringbuffer shm package
case SW_EVENT_PACKAGE:
//discard data
if (swWorker_discard_data(serv, task) == SW_TRUE)
{
break;
}
do_task:
{
worker->request_time = serv->gs->now;
#ifdef SW_BUFFER_RECV_TIME
serv->last_receive_usec = task->info.time;
#endif
serv->onReceive(serv, task);
worker->request_time = 0;
#ifdef SW_BUFFER_RECV_TIME
serv->last_receive_usec = 0;
#endif
worker->traced = 0;
worker->request_count++;
sw_atomic_fetch_add(&serv->stats->request_count, 1);
}
if (task->info.type == SW_EVENT_PACKAGE_END)
{
package->length = 0;
}
break;
//chunk package
case SW_EVENT_PACKAGE_START:
case SW_EVENT_PACKAGE_END:
//discard data
if (swWorker_discard_data(serv, task) == SW_TRUE)
{
break;
}
package = swWorker_get_buffer(serv, task->info.from_id);
if (task->info.len > 0)
{
//merge data to package buffer
swString_append_ptr(package, task->data, task->info.len);
}
//package end
if (task->info.type == SW_EVENT_PACKAGE_END)
{
goto do_task;
}
break;
case SW_EVENT_CLOSE:
factory->end(factory, task->info.fd);
break;
case SW_EVENT_CONNECT:
if (serv->onConnect)
{
serv->onConnect(serv, &task->info);
}
break;
case SW_EVENT_BUFFER_FULL:
if (serv->onBufferFull)
{
serv->onBufferFull(serv, &task->info);
}
break;
case SW_EVENT_BUFFER_EMPTY:
if (serv->onBufferEmpty)
{
serv->onBufferEmpty(serv, &task->info);
}
break;
case SW_EVENT_FINISH:
serv->onFinish(serv, task);
break;
case SW_EVENT_PIPE_MESSAGE:
serv->onPipeMessage(serv, task);
break;
default:
swWarn("[Worker] error event[type=%d]", (int )task->info.type);
break;
}
//worker idle
worker->status = SW_WORKER_IDLE;
//maximum number of requests, process will exit.
if (!SwooleWG.run_always && worker->request_count >= SwooleWG.max_request)
{
swWorker_stop();
}
return SW_OK;
}
swWorker_discard_data
验证数据有效性
swServer_connection_verify
函数利用 task->info.fd
这个 sessionid
来验证连接的有效性,如果连接已经关闭,或者已经被删除,那么就要抛弃当前数据
static sw_inline int swWorker_discard_data(swServer *serv, swEventData *task)
{
int fd = task->info.fd;
//check connection
swConnection *conn = swServer_connection_verify(serv, task->info.fd);
if (conn == NULL)
{
if (serv->disable_notify && !serv->discard_timeout_request)
{
return SW_FALSE;
}
goto discard_data;
}
else
{
if (conn->closed)
{
goto discard_data;
}
else
{
return SW_FALSE;
}
}
discard_data:
#ifdef SW_USE_RINGBUFFER
if (task->info.type == SW_EVENT_PACKAGE)
{
swPackage package;
memcpy(&package, task->data, sizeof(package));
swReactorThread *thread = swServer_get_thread(SwooleG.serv, task->info.from_id);
thread->buffer_input->free(thread->buffer_input, package.data);
swoole_error_log(SW_LOG_WARNING, SW_ERROR_SESSION_DISCARD_TIMEOUT_DATA, "[1]received the wrong data[%d bytes] from socket#%d", package.length, fd);
}
else
#endif
{
swoole_error_log(SW_LOG_WARNING, SW_ERROR_SESSION_DISCARD_TIMEOUT_DATA, "[1]received the wrong data[%d bytes] from socket#%d", task->info.len, fd);
}
return SW_TRUE;
}
static sw_inline swConnection *swServer_connection_verify(swServer *serv, int session_id)
{
swConnection *conn = swServer_connection_verify_no_ssl(serv, session_id);
return conn;
}
static sw_inline swConnection *swServer_connection_verify_no_ssl(swServer *serv, int session_id)
{
swSession *session = swServer_get_session(serv, session_id);
int fd = session->fd;
swConnection *conn = swServer_connection_get(serv, fd);
if (!conn || conn->active == 0)
{
return NULL;
}
if (session->id != session_id || conn->session_id != session_id)
{
return NULL;
}
return conn;
}
php_swoole_onReceive
回调函数
该回调函数首先要调用 php_swoole_get_recv_data
获取数据,然后 sw_call_user_function_fast
执行 PHP
的回调函数
int php_swoole_onReceive(swServer *serv, swEventData *req)
{
swFactory *factory = &serv->factory;
zval *zserv = (zval *) serv->ptr2;
zval *zfd;
zval *zfrom_id;
zval *zdata;
zval *retval = NULL;
SWOOLE_GET_TSRMLS;
php_swoole_udp_t udp_info;
swDgramPacket *packet;
SW_MAKE_STD_ZVAL(zfd);
SW_MAKE_STD_ZVAL(zfrom_id);
SW_MAKE_STD_ZVAL(zdata);
{
ZVAL_LONG(zfrom_id, (long ) req->info.from_id);
ZVAL_LONG(zfd, (long ) req->info.fd);
php_swoole_get_recv_data(zdata, req, NULL, 0);
}
{
zval **args[4];
zval *callback = php_swoole_server_get_callback(serv, req->info.from_fd, SW_SERVER_CB_onReceive);
if (callback == NULL || ZVAL_IS_NULL(callback))
{
swoole_php_fatal_error(E_WARNING, "onReceive callback is null.");
return SW_OK;
}
args[0] = &zserv;
args[1] = &zfd;
args[2] = &zfrom_id;
args[3] = &zdata;
zend_fcall_info_cache *fci_cache = php_swoole_server_get_cache(serv, req->info.from_fd, SW_SERVER_CB_onReceive);
if (sw_call_user_function_fast(callback, fci_cache, &retval, 4, args TSRMLS_CC) == FAILURE)
{
swoole_php_fatal_error(E_WARNING, "onReceive handler error.");
}
}
if (EG(exception))
{
zend_exception_error(EG(exception), E_ERROR TSRMLS_CC);
}
sw_zval_ptr_dtor(&zfd);
sw_zval_ptr_dtor(&zfrom_id);
sw_zval_ptr_dtor(&zdata);
if (retval != NULL)
{
sw_zval_ptr_dtor(&retval);
}
return SW_OK;
}
php_swoole_get_recv_data
接受数据
- 如果使用的数据类型是
SW_EVENT_PACKAGE
,那么数据存储在ringBuff
共享内存池中,我们首先要把数据复制到zdata
当中,然后释放共享内存 - 如果数据类型是
SW_EVENT_PACKAGE_END
,那么数据存储在SwooleWG.buffer_input
中
void php_swoole_get_recv_data(zval *zdata, swEventData *req, char *header, uint32_t header_length)
{
char *data_ptr = NULL;
int data_len;
#ifdef SW_USE_RINGBUFFER
swPackage package;
if (req->info.type == SW_EVENT_PACKAGE)
{
memcpy(&package, req->data, sizeof (package));
data_ptr = package.data;
data_len = package.length;
}
#else
if (req->info.type == SW_EVENT_PACKAGE_END)
{
swString *worker_buffer = swWorker_get_buffer(SwooleG.serv, req->info.from_id);
data_ptr = worker_buffer->str;
data_len = worker_buffer->length;
}
#endif
else
{
data_ptr = req->data;
data_len = req->info.len;
}
if (header_length >= data_len)
{
SW_ZVAL_STRING(zdata, "", 1);
}
else
{
SW_ZVAL_STRINGL(zdata, data_ptr + header_length, data_len - header_length, 1);
}
if (header_length > 0)
{
memcpy(header, data_ptr, header_length);
}
#ifdef SW_USE_RINGBUFFER
if (req->info.type == SW_EVENT_PACKAGE)
{
swReactorThread *thread = swServer_get_thread(SwooleG.serv, req->info.from_id);
thread->buffer_input->free(thread->buffer_input, data_ptr);
}
#endif
}
swoole_server->send
向客户端发送数据
worker
进程向客户端发送数据时,会调用 swoole_server->send
函数,该函数会调用 swServer_tcp_send
函数
PHP_METHOD(swoole_server, send)
{
int ret;
zval *zfd;
zval *zdata;
zend_long server_socket = -1;
swServer *serv = swoole_get_object(getThis());
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "zz|l", &zfd, &zdata, &server_socket) == FAILURE)
{
return;
}
char *data;
int length = php_swoole_get_send_data(zdata, &data TSRMLS_CC);
convert: convert_to_long(zfd);
uint32_t fd = (uint32_t) Z_LVAL_P(zfd);
ret = swServer_tcp_send(serv, fd, data, length);
#ifdef SW_COROUTINE
if (ret < 0 && SwooleG.error == SW_ERROR_OUTPUT_BUFFER_OVERFLOW && serv->send_yield)
{
zval_add_ref(zdata);
php_swoole_server_send_yield(serv, fd, zdata, return_value);
}
else
#endif
{
SW_CHECK_RETURN(ret);
}
}
swServer_tcp_send
函数
- 如果使用
stream
模式,那么可以直接向serv->last_stream_fd
发送数据即可 - 如果是普通模式,那么需要打包类型为
SW_EVENT_TCP
的数据,调用finish
函数将数据放入pipe
的缓冲区中 - 注意小数据
_send.length
为 0,大数据_send.length
才会大于 0
int swServer_tcp_send(swServer *serv, int fd, void *data, uint32_t length)
{
swSendData _send;
swFactory *factory = &(serv->factory);
if (unlikely(swIsMaster()))
{
swoole_error_log(SW_LOG_ERROR, SW_ERROR_SERVER_SEND_IN_MASTER,
"can't send data to the connections in master process.");
return SW_ERR;
}
/**
* More than the output buffer
*/
if (length > serv->buffer_output_size)
{
swoole_error_log(SW_LOG_WARNING, SW_ERROR_DATA_LENGTH_TOO_LARGE, "More than the output buffer size[%d], please use the sendfile.", serv->buffer_output_size);
return SW_ERR;
}
else
{
if (fd == serv->last_session_id && serv->last_stream_fd > 0)
{
int _l = htonl(length);
if (SwooleG.main_reactor->write(SwooleG.main_reactor, serv->last_stream_fd, (void *) &_l, sizeof(_l)) < 0)
{
return SW_ERR;
}
if (SwooleG.main_reactor->write(SwooleG.main_reactor, serv->last_stream_fd, data, length) < 0)
{
return SW_ERR;
}
return SW_OK;
}
_send.info.fd = fd;
_send.info.type = SW_EVENT_TCP;
_send.data = data;
if (length >= SW_IPC_MAX_SIZE - sizeof(swDataHead))
{
_send.length = length;
}
else
{
_send.info.len = length;
_send.length = 0;
}
return factory->finish(factory, &_send);
}
return SW_OK;
}
swFactoryProcess_finish
函数
- 首先要验证数据的有效性,利用
swServer_connection_verify_no_ssl
获取到swConnection
对象 -
resp->length
大于 0,那么说明是大数据包,这个时候需要将数据放入worker->send_shm
当中,然后将worker_id
打包到swEventData
对象 - 如果
out_buffer
管道缓存区不是空的,说明管道中有数据未发送完毕(有可能是共享内存数据未发送完毕),那么就利用函数swTaskWorker_large_pack
将数据存放到临时文件中。(猜测防止发送到共享内存时被worker
进程锁锁住) - 如果是小数据包,那么就将数据打包到
swEventData
对象中 -
swWorker_send2reactor
函数将用于将数据发送到reactor
线程
static sw_inline int swWorker_get_send_pipe(swServer *serv, int session_id, int reactor_id)
{
int pipe_index = session_id % serv->reactor_pipe_num;
/**
* pipe_worker_id: The pipe in which worker.
*/
int pipe_worker_id = reactor_id + (pipe_index * serv->reactor_num);
swWorker *worker = swServer_get_worker(serv, pipe_worker_id);
return worker->pipe_worker;
}
static int swFactoryProcess_finish(swFactory *factory, swSendData *resp)
{
int ret, sendn;
swServer *serv = factory->ptr;
int session_id = resp->info.fd;
swConnection *conn;
if (resp->info.type != SW_EVENT_CLOSE)
{
conn = swServer_connection_verify(serv, session_id);
}
else
{
conn = swServer_connection_verify_no_ssl(serv, session_id);
}
if (!conn)
{
swoole_error_log(SW_LOG_NOTICE, SW_ERROR_SESSION_NOT_EXIST, "connection[fd=%d] does not exists.", session_id);
return SW_ERR;
}
else if ((conn->closed || conn->removed) && resp->info.type != SW_EVENT_CLOSE)
{
int _len = resp->length > 0 ? resp->length : resp->info.len;
swoole_error_log(SW_LOG_NOTICE, SW_ERROR_SESSION_CLOSED, "send %d byte failed, because connection[fd=%d] is closed.", _len, session_id);
return SW_ERR;
}
else if (conn->overflow)
{
swoole_error_log(SW_LOG_WARNING, SW_ERROR_OUTPUT_BUFFER_OVERFLOW, "send failed, connection[fd=%d] output buffer has been overflowed.", session_id);
return SW_ERR;
}
swEventData ev_data;
ev_data.info.fd = session_id;
ev_data.info.type = resp->info.type;
swWorker *worker = swServer_get_worker(serv, SwooleWG.id);
/**
* Big response, use shared memory
*/
if (resp->length > 0)
{
if (worker == NULL || worker->send_shm == NULL)
{
goto pack_data;
}
//worker process
if (SwooleG.main_reactor)
{
int _pipe_fd = swWorker_get_send_pipe(serv, session_id, conn->from_id);
swConnection *_pipe_socket = swReactor_get(SwooleG.main_reactor, _pipe_fd);
//cannot use send_shm
if (!swBuffer_empty(_pipe_socket->out_buffer))
{
pack_data:
if (swTaskWorker_large_pack(&ev_data, resp->data, resp->length) < 0)
{
return SW_ERR;
}
ev_data.info.from_fd = SW_RESPONSE_TMPFILE;
goto send_to_reactor_thread;
}
}
swPackage_response response;
response.length = resp->length;
response.worker_id = SwooleWG.id;
ev_data.info.from_fd = SW_RESPONSE_SHM;
ev_data.info.len = sizeof(response);
memcpy(ev_data.data, &response, sizeof(response));
swTrace("[Worker] big response, length=%d|worker_id=%d", response.length, response.worker_id);
worker->lock.lock(&worker->lock);
memcpy(worker->send_shm, resp->data, resp->length);
}
else
{
//copy data
memcpy(ev_data.data, resp->data, resp->info.len);
ev_data.info.len = resp->info.len;
ev_data.info.from_fd = SW_RESPONSE_SMALL;
}
send_to_reactor_thread: ev_data.info.from_id = conn->from_id;
sendn = ev_data.info.len + sizeof(resp->info);
swTrace("[Worker] send: sendn=%d|type=%d|content=%s", sendn, resp->info.type, resp->data);
ret = swWorker_send2reactor(&ev_data, sendn, session_id);
if (ret < 0)
{
swWarn("sendto to reactor failed. Error: %s [%d]", strerror(errno), errno);
}
return ret;
}
swTaskWorker_large_pack
函数
int swTaskWorker_large_pack(swEventData *task, void *data, int data_len)
{
swPackage_task pkg;
bzero(&pkg, sizeof(pkg));
memcpy(pkg.tmpfile, SwooleG.task_tmpdir, SwooleG.task_tmpdir_len);
//create temp file
int tmp_fd = swoole_tmpfile(pkg.tmpfile);
if (tmp_fd < 0)
{
return SW_ERR;
}
//write to file
if (swoole_sync_writefile(tmp_fd, data, data_len) <= 0)
{
swWarn("write to tmpfile failed.");
return SW_ERR;
}
task->info.len = sizeof(swPackage_task);
//use tmp file
swTask_type(task) |= SW_TASK_TMPFILE;
pkg.length = data_len;
memcpy(task->data, &pkg, sizeof(swPackage_task));
close(tmp_fd);
return SW_OK;
}
int swoole_tmpfile(char *filename)
{
#if defined(HAVE_MKOSTEMP) && defined(HAVE_EPOLL)
int tmp_fd = mkostemp(filename, O_WRONLY | O_CREAT);
#else
int tmp_fd = mkstemp(filename);
#endif
if (tmp_fd < 0)
{
swSysError("mkstemp(%s) failed.", filename);
return SW_ERR;
}
else
{
return tmp_fd;
}
}
int swoole_sync_writefile(int fd, void *data, int len)
{
int n = 0;
int count = len, towrite, written = 0;
while (count > 0)
{
towrite = count;
if (towrite > SW_FILE_CHUNK_SIZE)
{
towrite = SW_FILE_CHUNK_SIZE;
}
n = write(fd, data, towrite);
if (n > 0)
{
data += n;
count -= n;
written += n;
}
else if (n == 0)
{
break;
}
else
{
if (errno == EINTR || errno == EAGAIN)
{
continue;
}
swSysError("write(%d, %d) failed.", fd, towrite);
break;
}
}
return written;
}
swWorker_send2reactor
发送数据
-
swWorker_send2reactor
函数专门负责将swEventData
数据发送到pipefd
的缓冲区中,其使用的是main_reactor->write
方法,我们之前在reactor
中已经了解过。 -
swWorker_get_send_pipe
用于计算发送给客户端的pipefd
。我们知道,用户可以在任何worker
中调用swoole_server->send(int $fd, string $data, int $extraData = 0)
向客户端发送数据,但是其中的fd
并不一定是本worker
进程负责的session_id
。我们可以从session_id
中获取到swConnection
,进而获取到reactor_id
线程,但是我们无法确定当前该连接被分配给了那个worker
。因此为了均衡各个worker
,首先计算出平均每个reactor
负责的worker
数量reactor_pipe_num
,然后利用session_id
以取模的方式随机选择其中一个worker
,然后计算出该worker
的id
,进而取出其pipe_worker
serv->reactor_pipe_num = serv->worker_num / serv->reactor_num
static sw_inline int swWorker_get_send_pipe(swServer *serv, int session_id, int reactor_id)
{
int pipe_index = session_id % serv->reactor_pipe_num;
/**
* pipe_worker_id: The pipe in which worker.
*/
int pipe_worker_id = reactor_id + (pipe_index * serv->reactor_num);
swWorker *worker = swServer_get_worker(serv, pipe_worker_id);
return worker->pipe_worker;
}
int swWorker_send2reactor(swEventData *ev_data, size_t sendn, int session_id)
{
int ret;
swServer *serv = SwooleG.serv;
int _pipe_fd = swWorker_get_send_pipe(serv, session_id, ev_data->info.from_id);
if (SwooleG.main_reactor)
{
ret = SwooleG.main_reactor->write(SwooleG.main_reactor, _pipe_fd, ev_data, sendn);
}
else
{
ret = swSocket_write_blocking(_pipe_fd, ev_data, sendn);
}
return ret;
}
swFactoryProcess_end
关闭连接
当用户主动调用 swoole_server->close
函数的时候,就会调用本函数。swFactoryProcess_end
函数主要用于调用 onClose
函数,进而调用 `swFactoryProcess_finish
函数`
static int swFactoryProcess_end(swFactory *factory, int fd)
{
swServer *serv = factory->ptr;
swSendData _send;
swDataHead info;
bzero(&_send, sizeof(_send));
_send.info.fd = fd;
_send.info.len = 0;
_send.info.type = SW_EVENT_CLOSE;
swConnection *conn = swWorker_get_connection(serv, fd);
if (conn == NULL || conn->active == 0)
{
SwooleG.error = SW_ERROR_SESSION_NOT_EXIST;
return SW_ERR;
}
else if (conn->close_force)
{
goto do_close;
}
else if (conn->closing)
{
swoole_error_log(SW_LOG_NOTICE, SW_ERROR_SESSION_CLOSING, "The connection[%d] is closing.", fd);
return SW_ERR;
}
else if (conn->closed)
{
return SW_ERR;
}
else
{
do_close:
conn->closing = 1;
if (serv->onClose != NULL)
{
info.fd = fd;
if (conn->close_actively)
{
info.from_id = -1;
}
else
{
info.from_id = conn->from_id;
}
info.from_fd = conn->from_fd;
serv->onClose(serv, &info);
}
conn->closing = 0;
conn->closed = 1;
conn->close_errno = 0;
return factory->finish(factory, &_send);
}
}