Unix下可用的I/O模型有五种：
详见Unix网络编程卷一第六章
select()和poll()在Unix系统中存在时间长，主要优势在于可移植性，主要缺点在于当同时检查大量的文件描述符时性能拓展性不佳。
epoll API的关键优势在于能让应用高效地检查大量的文件描述符，主要缺点是专属于Linux系统的API。
select()首次出现在BSD系统的套接字API中。
select()系统调用的用途：在一段指定的时间内，监听用户感兴趣的文件描述符上的可读、可写和异常事件。
系统调用select()会一直阻塞，直到一个或多个文件描述符集合成为就绪态。
#include <sys/select.h>
#include <sys/time.h>
//若有就绪描述符则返回其数目，若超时则返回0，若出错则返回-1
int select(int maxfdp1, fd_set *readset, fd_set *writeset,
fd_set *exceptset, const struct timeval *timeout);
探究下fd_set的结构
/*typesizes.h*/
#define __FD_SETSIZE 1024
/*select.h*/
typedef long int __fd_mask;
//long int类型共有多少bits
#define __NFDBITS (8 * (int) sizeof (__fd_mask))
typedef struct
{
//long int型数组，数组大小 = 描述符最大数 / long int的位数
//数组大小为 __FD_SETSIZE bits
__fd_mask fds_bits[__FD_SETSIZE / __NFDBITS];
} fd_set;
select()程序示例：
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
#include <stdio.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <fcntl.h>
#include <stdlib.h>
#include <stdarg.h>
static void usageError(const char* progName){
fprintf(stderr, "Usage: %s {timeout | -} fd-num[rw]...\n", progName);
fprintf(stderr, " - means infinite timeout; \n");
fprintf(stderr, " r = monitor for read\n");
fprintf(stderr, " w = monitor for wirite\n\n");
fprintf(stderr, " e.g.: %s - 0rw 1w\n", progName);
exit(1);
}
void cmdLineErr(const char *format, ...)
{
va_list argList;
fflush(stdout); /* Flush any pending stdout */
fprintf(stderr, "Command-line usage error: ");
va_start(argList, format);
vfprintf(stderr, format, argList);
va_end(argList);
fflush(stderr); /* In case stderr is not line-buffered */
exit(EXIT_FAILURE);
}
int main(int argc, char* argv[]){
fd_set readfds, writefds;
int ready, nfds, fd, numRead, j;
struct timeval timeout;
struct timeval *pto;
char buf[10];
if(argc < 2 || strcmp(argv[1], "--help") == 0){
usageError(argv[0]);
}
if(strcmp(argv[1], "-") == 0){
pto = NULL;
}
else{
pto = &timeout;
timeout.tv_sec = strtol(argv[1], NULL, 0);
timeout.tv_usec = 0;
}
nfds = 0;
FD_ZERO(&readfds);
FD_ZERO(&writefds);
for(j = 2; j < argc; j++){
numRead = sscanf(argv[j], "%d%2[rw]", &fd, buf);
if(numRead != 2){
usageError(argv[0]);
}
if(fd >= FD_SETSIZE){
cmdLineErr("file descriptor exceeds limit (%d)\n", FD_SETSIZE);
}
if(fd >= nfds){
nfds = fd + 1;
}
if(strchr(buf, 'r') != NULL){
FD_SET(fd, &readfds);
}
if(strchr(buf, 'w') != NULL){
FD_SET(fd, &writefds);
}
}
ready = select(nfds, &readfds, &writefds, NULL, pto);
if(ready == -1){
printf("errExit(select)");
exit(1);
}
printf("ready = %d\n", ready);
for(fd = 0; fd < nfds; fd++){
printf("%d: %s%s\n",fd, FD_ISSET(fd, &readfds) ? "r" : "",
FD_ISSET(fd, &writefds) ? "w" : "");
}
if(pto != NULL){
printf("timeout after select(): %ld.%03ld\n",
(long) timeout.tv_sec, (long) timeout.tv_usec / 1000);
}
exit(0);
}
select处理正常数据和带外数据：
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
#include <stdio.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <fcntl.h>
#include <stdlib.h>
int main(int argc, char* argv[]){
if(argc <= 2){
printf("usage: %s ip_adress port_number\n", basename(argv[0]));
return 1;
}
const char* ip = argv[1];
int port = atoi(argv[2]);
int ret = 0;
struct sockaddr_in address;
bzero(&address, sizeof(address));
address.sin_family = AF_INET;
inet_pton(AF_INET, ip, &address.sin_addr);
address.sin_port = htons(port);
int listenfd = socket(PF_INET, SOCK_STREAM, 0);
assert(listenfd >= 0);
ret = bind(listenfd, (struct sockaddr*)&address, sizeof(address));
assert(ret != -1);
ret = listen(listenfd, 5);
assert(ret != -1);
struct sockaddr_in client_address;
socklen_t client_addrlength = sizeof(client_address);
int connfd = accept(listenfd, (struct sockaddr*)&client_address, &client_addrlength);
if(connfd < 0){
printf("error is: %d\n", errno);
close(listenfd);
}
char buf[1024];
fd_set read_fds;
fd_set exception_fds;
FD_ZERO(&read_fds);
FD_ZERO(&exception_fds);
while(true){
memset(buf, '\0', sizeof(buf));
FD_SET(connfd, &read_fds);
FD_SET(connfd, &exception_fds);
ret = select(connfd + 1, &read_fds, NULL, &exception_fds, NULL);
if(ret < 0){
printf("selection failure\n");
break;
}
if(FD_ISSET(connfd, &read_fds)){
ret = recv(connfd, buf, sizeof(buf)-1, 0);
if(ret <= 0){
break;
}
printf("get %d bytes of normal data: %s\n", ret, buf);
}
else if(FD_ISSET(connfd, &exception_fds)){
ret = recv(connfd, buf, sizeof(buf)-1, MSG_OOB);
if(ret <= 0){
break;
}
printf("get %d bytes of oob data: %s\n", ret, buf);
}
}
close(connfd);
close(listenfd);
return 0;
}
poll函数起源于SVR3，最初局限于流设备，SVR4取消了这种限制，允许poll工作在任何描述符上。
poll提供的功能与select类似，不过在处理流设备时，它能够提供额外的信息。
#include <poll.h>
struct pollfd{
int fd;
short events; //指定要测试的条件
short revents;//返回描述符的状态
}
//若有就绪描述符返回其数目，超时返回0，出错返回-1
int poll(struct pollfd *fdarray, unsigned long nfds, int timeout);
select()同poll()返回正整数的区别：如果一个文件描述符在返回的集合中出现了不止一次，系统调用select()会将同一个文件描述符计数多次。而系统调用poll()返回的是就绪态文件描述符个数，且一个文件描述符只会统计一次，就算在相应的revents字段中设定了多个位掩码也是如此。
poll示例程序：
#include <time.h>
#include <poll.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
#include <stdio.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <fcntl.h>
#include <stdlib.h>
#include <stdarg.h>
static void usageError(const char* progName){
fprintf(stderr, "Usage: %s {timeout | -} fd-num[rw]...\n", progName);
fprintf(stderr, " - means infinite timeout; \n");
fprintf(stderr, " r = monitor for read\n");
fprintf(stderr, " w = monitor for wirite\n\n");
fprintf(stderr, " e.g.: %s - 0rw 1w\n", progName);
exit(1);
}
int main(int argc, char* argv[]){
int numPipes, j, ready, randPipe, numWrites;
int (*pfds)[2];//指向数组的指针
struct pollfd *pollFd;
if(argc < 2 || strcmp(argv[1], "--help") == 0){
printf("%s num-pipes [num-writes]\n", argv[0]);
exit(1);
}
numPipes = strtol(argv[1], NULL, 10);
pfds = (int (*)[2])calloc(numPipes, sizeof(int [2]));
if(pfds == NULL){
printf("error malloc");
exit(1);
}
pollFd = (pollfd*)calloc(numPipes, sizeof(struct pollfd));
if(pollFd == NULL){
printf("error malloc");
exit(1);
}
for(j = 0; j < numPipes; j++){
if(pipe(pfds[j]) == -1){
printf("error pipe %d", j);
exit(1);
}
}
numWrites = (argc > 2) ? strtol(argv[2], NULL, 10) : 1;
srandom((int)time(NULL));
for(j = 0; j < numWrites; j++){
randPipe = random() % numPipes;
printf("Writing to fd: %3d (read fd: %3d)\n",
pfds[randPipe][1], pfds[randPipe][0]);
if (write(pfds[randPipe][1], "a", 1) == -1){
printf("write %d", pfds[randPipe][1]);
exit(1);
}
}
for(j = 0; j < numPipes; j++){
pollFd[j].fd = pfds[j][0];
pollFd[j].events = POLLIN;
}
ready = poll(pollFd, numPipes, -1);
if(ready == -1){
printf("poll error");
exit(1);
}
printf("poll() returned: %d\n", ready);
for(j = 0; j < numPipes; j++){
if(pollFd[j].revents & POLLIN){
printf("Readable: %d %3d\n", j, pollFd[j].fd);
}
}
return 0;
}
epoll API由三组系统调用组成;
epoll_create()创建一个epoll实例epoll_ctl()操作同epoll实例相关联的兴趣列表epoll_wait()返回与epoll相关联的就绪列表中的成员epoll实例：epoll API的核心数据结构，和一个打开的文件描述符相关联。这个文件描述符不用来做IO操作，相反它是内核数据结构的句柄，这些内核数据结构实现了两个目的：
#include <sys/epoll.h>
int epoll_create(int size);
参数size指定我们想要通过epoll实例来检查的描述符个数，不是上限，只是告知内核应该如何为内部数据结构划分初始大小。
函数返回epoll实例的文件描述符，该文件描述符不需要时需要close()。
当所有与epoll实例相关的文件描述符都被关闭时，实例被销毁，相关资源释放。（多个文件描述符可能引用到相同的epoll实例，这是由于调用了fork()或dup()这样的类似函数所致）。
linux2.6.8版以来，size参数被忽略不用。
linux2.6.27以来，Linux支持一个新的系统调用epoll_create1()：
- 去掉了无用的参数size
- 增加了一个可用来修改系统调用行为的flags参数
- flag目前只支持一个标志：EPOLL_CLOEXEC，使内核在新的文件描述符上启动了执行即关闭（close-on-exec）标志（FD_CLOEXEC)
#include <sys/epoll.h>
int epoll_ctl(int epfd, int op, int fd, struct epoll *ev);
成功返回0，失败返回-1并设置errno。
参数fd：指明修改兴趣列表中哪一个文件描述符的设定
参数op：指定需要执行的操作
参数ev：
struct epoll_event{
uint32_t events;//epoll事件，位掩码
epoll_data_t data; //用户数据
}
typedef union epoll_data{
void *ptr;
int fd;
uint32_t u32;
uint64_t u64;
}epoll_data_t;
max_user_watches上限
每个注册到epoll实例上的文件描述符需要占用一小段不能被交换的内核内存空间，因此内核提供了一个接口用来定义每个用户可以注册到epoll实例上的文件描述符总数。
这个上限值可以通过max_user_watches来查看和修改，max_user_watches是专属于Linux系统的/proc/sys/fd/epoll目录下的一个文件。默认上限值根据可用系统内存计算得出。
#include <sys/epoll.h>
int epoll_wait(int epfd, struct epoll_event *evlist, int maxevents, int timeout);
成功返回就绪态的文件描述符的个数，失败返回-1并设置errno
参数evlist指向的结构体数组中返回的是有关就绪态文件描述符的信息。数组evlist的空间由调用者负责申请，所包含的元素个数在参数maxevents中指定。
在数组evlist中每个元素返回的都是单个就绪态文件描述符的信息：
参数timeout用来确定epoll_wait()的阻塞行为：
在多线程程序中，可以在一个线程中使用epoll_ctl()将文件描述符添加到另一个线程中由epoll_wait()所监视的epoll实例的兴趣列表中去。这些对兴趣列表的修改将立刻得到处理，而epoll_wait()调用将返回有关新添加的文件描述符的就绪信息。
epoll事件：除了有一个额外的前缀E外，大多数位掩码的名称同poll中对应的事件掩码名称相同。例外情况：
epoll程序示例：
#include <sys/epoll.h>
#include <fcntl.h>
#include <string.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
#define MAX_BUF 1000
#define MAX_EVENTS 5
int main(int argc, char* argv[]){
int epfd, ready, fd, s, j, numOpenFds;
struct epoll_event ev;
struct epoll_event evlist[MAX_EVENTS];
char buf[MAX_BUF];
if(argc < 2 || strcmp(argv[1], "--help")==0){
printf("usage: %s file...\n", argv[0]);
exit(1);
}
epfd = epoll_create(argc - 1);
if(epfd == -1){
printf("error epoll_create");
exit(1);
}
for(j = 1; j < argc; j++){
fd = open(argv[j], O_RDONLY);
if(fd == -1){
printf("error open");
exit(1);
}
printf("Opened \"%s\" on fd %d\n", argv[j], fd);
ev.events = EPOLLIN;
ev.data.fd = fd;
if(epoll_ctl(epfd, EPOLL_CTL_ADD, fd, &ev) == -1){
printf("error epoll_ctl");
exit(1);
}
}
numOpenFds = argc - 1;
while(numOpenFds > 0){
printf("About to epoll_wait()\n");
ready = epoll_wait(epfd, evlist, MAX_EVENTS, -1);
if(ready == -1){
if(errno == EINTR)continue;
else{
printf("error epoll_wait");
exit(1);
}
}
printf("Ready: %d\n", ready);
for(j = 0; j < ready; j++){
printf(" fd = %d; events: %s%s%s\n", evlist[j].data.fd,
(evlist[j].events & EPOLLIN) ? "EPOLLIN ":"",
(evlist[j].events & EPOLLHUP) ? "EPOLLHUP":"",
(evlist[j].events & EPOLLERR) ? "EPOLLERR":"");
if(evlist[j].events & EPOLLIN){
s = read(evlist[j].data.fd, buf, MAX_BUF);
if(s == -1){
printf("error read");
}
printf(" read %d bytes : %.*s",s,s,buf);
}
else if(evlist[j].events & (EPOLLHUP | EPOLLERR)){
printf(" closing fd %d\n", evlist[j].data.fd);
if(close(evlist[j].data.fd) == -1){
printf("error close");
exit(1);
}
numOpenFds--;
}
}
}
printf("All file descriptors closed; bye\n");
exit(0);
}
ET模式比LT模式触发事件的次数更少：
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
#include <stdio.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <fcntl.h>
#include <stdlib.h>
#include <sys/epoll.h>
#include <pthread.h>
#define MAX_EVENT_NUMBER 1024
#define BUFFER_SIZE 10
int setnonblocking(int fd){
int old_option = fcntl(fd, F_GETFL);
int new_option = old_option | O_NONBLOCK;
fcntl(fd, F_SETFL, new_option);
return old_option;
}
void addfd(int epollfd, int fd, bool enable_et){
epoll_event event;
event.data.fd = fd;
event.events = EPOLLIN;
if(enable_et){
event.events |= EPOLLET;
}
epoll_ctl(epollfd, EPOLL_CTL_ADD, fd, &event);
setnonblocking(fd);
}
void lt(epoll_event *events, int number, int epollfd, int listenfd){
char buf[BUFFER_SIZE];
for(int i = 0; i < number; i++){
int sockfd = events[i].data.fd;
if(sockfd == listenfd){
struct sockaddr_in client_address;
socklen_t client_addrlength = sizeof(client_address);
int connfd = accept(listenfd, (struct sockaddr*)&client_address, &client_addrlength);
addfd(epollfd, connfd, false);
}
else if(events[i].events & EPOLLIN){
printf("event trigger once\n");
memset(buf, '\0', BUFFER_SIZE);
int ret = recv(sockfd, buf, BUFFER_SIZE-1,0);
if(ret <= 0){
close(sockfd);
continue;
}
printf("get %d bytes of content: %s\n", ret, buf);
}
else{
printf("something else happened \n");
}
}
}
void et(epoll_event* events, int number, int epollfd, int listenfd){
char buf[BUFFER_SIZE];
for(int i = 0; i < number; i++){
int sockfd = events[i].data.fd;
if(sockfd == listenfd){
struct sockaddr_in client_address;
socklen_t client_addrlength = sizeof(client_address);
int connfd = accept(listenfd, (struct sockaddr*)&client_address, &client_addrlength);
addfd(epollfd, connfd, true);
}
else if(events[i].events & EPOLLIN){
printf("event trigger once\n");
while(true){
memset(buf, '\0',BUFFER_SIZE);
int ret = recv(sockfd, buf, BUFFER_SIZE-1, 0);
if(ret < 0){
if((errno == EAGAIN) || (errno == EWOULDBLOCK)){
printf("read later\n");
break;
}
close(sockfd);
break;
}
else if(ret == 0){
close(sockfd);
}
else{
printf("get %d bytes of content: %s\n",ret, buf);
}
}
}
else{
printf("something else happend \n");
}
}
}
int main(int argc, char* argv[]){
if(argc <= 2){
printf("usage: %s ip_address port_number\n", basename(argv[0]));
return 1;
}
const char *ip = argv[1];
int port = atoi(argv[2]);
int ret = 0;
struct sockaddr_in address;
bzero(&address, sizeof(address));
address.sin_family = AF_INET;
inet_pton(AF_INET, ip, &address.sin_addr);
address.sin_port = htons(port);
int listenfd = socket(PF_INET, SOCK_STREAM, 0);
assert(listenfd >= 0);
ret = bind(listenfd, (struct sockaddr*)&address, sizeof(address));
assert(ret != -1);
ret = listen(listenfd, 5);
assert(ret != -1);
epoll_event events[MAX_EVENT_NUMBER];
int epollfd = epoll_create(5);
assert(epollfd != -1);
addfd(epollfd, listenfd, true);
while(true){
int ret = epoll_wait(epollfd, events, MAX_EVENT_NUMBER, -1);
if(ret < 0){
printf("epoll failure\n");
break;
}
lt(events, ret, epollfd, listenfd);
//et(events, ret, epollfd, listenfd);
}
close(listenfd);
return 0;
}
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <assert.h>
#include <stdio.h>
#include <unistd.h>
#include <errno.h>
#include <string.h>
#include <fcntl.h>
#include <stdlib.h>
#include <sys/epoll.h>
#include <pthread.h>
#define MAX_EVENT_NUMBER 1024
#define BUFFER_SIZE 1024
struct fds{
int epollfd;
int sockfd;
};
int setnonblocking(int fd){
int old_option = fcntl(fd, F_GETFL);
int new_option = old_option | O_NONBLOCK;
fcntl(fd, F_SETFL, new_option);
return old_option;
}
void addfd(int epollfd, int fd, bool oneshot){
epoll_event event;
event.data.fd = fd;
event.events = EPOLLIN | EPOLLET;
if(oneshot){
event.events |= EPOLLONESHOT;
}
epoll_ctl(epollfd, EPOLL_CTL_ADD, fd, &event);
setnonblocking(fd);
}
void reset_oneshot(int epollfd, int fd){
epoll_event event;
event.data.fd = fd;
event.events = EPOLLIN | EPOLLET | EPOLLONESHOT;
epoll_ctl(epollfd, EPOLL_CTL_MOD, fd, &event);
}
void *worker(void *arg){
int sockfd = ((fds*)arg)->sockfd;
int epollfd = ((fds*)arg)->epollfd;
printf("start new thread to receive data on fd: %d\n", sockfd);
char buf[BUFFER_SIZE];
memset(buf, '\0', BUFFER_SIZE);
while(1){
int ret = recv(sockfd, buf, BUFFER_SIZE-1, 0);
if(ret == 0){
close(sockfd);
printf("foreiner closed the connection\n");
break;
}
else if(ret < 0){
if(errno == EAGAIN){
reset_oneshot(epollfd, sockfd);
printf("read later\n");
break;
}
}
else{
printf("get content: %s\n", buf);
sleep(5);
}
}
printf("end thread receving data on fd : %d\n", sockfd);
}
int main(int argc, char* argv[]){
if(argc < 2){
printf("usage: %s ip_address port_number\n", basename(argv[0]));
return 1;
}
const char* ip = argv[1];
int port = atoi(argv[2]);
int ret = 0;
struct sockaddr_in address;
bzero(&address, sizeof(address));
address.sin_family = AF_INET;
inet_pton(AF_INET, ip, &address.sin_addr);
address.sin_port = htons(port);
int listenfd = socket(PF_INET, SOCK_STREAM, 0);
assert(listenfd >= 0);
ret = bind(listenfd, (struct sockaddr*)&address, sizeof(address));
assert(ret != -1);
ret = listen(listenfd, 5);
assert(ret != -1);
epoll_event events[MAX_EVENT_NUMBER];
int epollfd = epoll_create(5);
assert(epollfd != -1);
addfd(epollfd, listenfd, false);
while(1){
int ret = epoll_wait(epollfd, events, MAX_EVENT_NUMBER, -1);
if(ret < 0){
printf("epoll failure\n");
break;
}
for(int i = 0; i < ret; i++){
int sockfd = events[i].data.fd;
if(sockfd == listenfd){
struct sockaddr_in client_address;
socklen_t client_addrlength = sizeof(client_address);
int connfd = accept(listenfd, (struct sockaddr*)&client_address, &client_addrlength);
addfd(epollfd, connfd, true);
}
else if(events[i].events & EPOLLIN){
pthread_t thread;
fds fds_for_new_worker;
fds_for_new_worker.epollfd = epollfd;
fds_for_new_worker.sockfd = sockfd;
pthread_create(&thread, NULL, worker, (void*)&fds_for_new_worker);
}
else{
printf("something else happened \n");
}
}
}
close(listenfd);
return 0;
}xxxxxxxxxx c
