TCP/IP協(xié)議棧的心跳機(jī)制

1、問題概述

雖然軟件底層模塊在網(wǎng)絡(luò)恢復(fù)后能自動(dòng)重連上服務(wù)器，但會(huì)議因?yàn)榫W(wǎng)絡(luò)問題已經(jīng)退出，需要重新加入會(huì)議。因?yàn)榭蛻籼厥獾木W(wǎng)絡(luò)運(yùn)行環(huán)境，會(huì)頻繁出現(xiàn)網(wǎng)絡(luò)抖動(dòng)不穩(wěn)定的情況，客戶要求必須要實(shí)現(xiàn)60秒內(nèi)網(wǎng)絡(luò)恢復(fù)后能依然保持在會(huì)議中，保證會(huì)議流程不被中斷。

客戶堅(jiān)持要實(shí)現(xiàn)這個(gè)特殊的功能點(diǎn)，項(xiàng)目已經(jīng)接近尾聲，目前處于客戶試用階段，不實(shí)現(xiàn)該功能，項(xiàng)目無法通過驗(yàn)收，客戶不給錢。

前方同事將當(dāng)前問題及項(xiàng)目進(jìn)展情況向研發(fā)部門領(lǐng)導(dǎo)反饋，研發(fā)部緊急召開討論會(huì)議，商討60秒不掉會(huì)的實(shí)現(xiàn)方案。這里面涉及到兩大類的網(wǎng)絡(luò)連接，一類是傳輸控制信令的TCP連接，另一類是傳輸音視頻碼流的UDP連接。UDP連接的問題不大，主要是TCP連接的斷鏈與重連問題，下面主要討論TCP連接相關(guān)問題。

在出現(xiàn)網(wǎng)絡(luò)不穩(wěn)定掉會(huì)時(shí)，可能是系統(tǒng)TCPIP協(xié)議棧已經(jīng)檢測(cè)到網(wǎng)絡(luò)異常，系統(tǒng)協(xié)議層已經(jīng)將網(wǎng)絡(luò)斷開了；也可能軟件應(yīng)用層的心跳機(jī)制檢測(cè)到網(wǎng)絡(luò)故障，斷開了與服務(wù)器的鏈接。對(duì)于系統(tǒng)TCPIP協(xié)議棧自身檢測(cè)出來的網(wǎng)絡(luò)異常，則可能存在兩種情況，一是TCPIP協(xié)議棧自身的心跳機(jī)制檢測(cè)出來的；二是TCP連接的丟包重傳機(jī)制檢測(cè)出異常。

對(duì)于應(yīng)用層的心跳檢測(cè)機(jī)制，我們可以放大超時(shí)檢測(cè)時(shí)間。本文我們主要討論一下TCPIP協(xié)議棧的TCP連接的心跳、丟包重傳、連接超時(shí)等機(jī)制。在檢測(cè)到網(wǎng)絡(luò)異常后，我們底層可以自動(dòng)發(fā)起重連或者信令發(fā)送觸發(fā)自動(dòng)重連，業(yè)務(wù)模塊將會(huì)議相關(guān)資源保存不釋放，在網(wǎng)絡(luò)恢復(fù)后可以繼續(xù)保持在會(huì)議中，可以繼續(xù)接收到會(huì)議中的音視頻碼流，可以繼續(xù)進(jìn)行會(huì)議中的一些操作！

2、TCPIP協(xié)議棧的心跳機(jī)制

2.1、TCP中的ACK機(jī)制

TCP建鏈時(shí)的三次握手流程如下所示：

之所以說TCP連接是可靠的，首先是發(fā)送數(shù)據(jù)前要建立連接，再就是收到數(shù)據(jù)后都會(huì)給對(duì)方恢復(fù)一個(gè)ACK包，表明我收到你的數(shù)據(jù)包了。對(duì)于數(shù)據(jù)發(fā)送端，如果數(shù)據(jù)發(fā)出去后沒有收到ACK包，則會(huì)觸發(fā)丟包重傳機(jī)制。

不管是建鏈時(shí)，還是建鏈后的數(shù)據(jù)收發(fā)時(shí)，都有ACK包，TCPIP協(xié)議棧的心跳包也不例外。

2.2、TCPIP協(xié)議棧的心跳機(jī)制說明

TCPIP協(xié)議棧有個(gè)默認(rèn)的TCP心跳機(jī)制，這個(gè)心跳機(jī)制是和socket套接字（TCP套接字）綁定的，可以對(duì)指定的套接字開啟協(xié)議棧的心跳檢測(cè)機(jī)制。默認(rèn)情況下，協(xié)議棧的心跳機(jī)制對(duì)socket套接字是關(guān)閉的，如果要使用需要人為開啟的。

在Windows中，默認(rèn)是每隔2個(gè)小時(shí)發(fā)一次心跳包，客戶端程序?qū)⑿奶l(fā)給服務(wù)器后，接下來會(huì)有兩種情況：

1）網(wǎng)絡(luò)正常時(shí)：服務(wù)器收到心跳包，會(huì)立即回復(fù)ACK包，客戶端收到ACK包后，再等2個(gè)小時(shí)發(fā)送下一個(gè)心跳包。其中，心跳包發(fā)送時(shí)間間隔時(shí)間keepalivetime，Windows系統(tǒng)中默認(rèn)是2小時(shí)，可配置。如果在2個(gè)小時(shí)的時(shí)間間隔內(nèi)，客戶端和服務(wù)器有數(shù)據(jù)交互，客戶端會(huì)收到服務(wù)器的ACK包，也算作心跳機(jī)制的心跳包，2個(gè)小時(shí)的時(shí)間間隔會(huì)重新計(jì)時(shí)。2）網(wǎng)絡(luò)異常時(shí)：服務(wù)器收不到客戶端發(fā)過去的心跳包，沒法回復(fù)ACK，Windows系統(tǒng)中默認(rèn)的是1秒超時(shí)，1秒后會(huì)重發(fā)心跳包。如果還收不到心跳包的ACK，則1秒后重發(fā)心跳包，如果始終收不到心跳包，則在發(fā)出10個(gè)心跳包就達(dá)到了系統(tǒng)的上限，就認(rèn)為網(wǎng)絡(luò)出故障了，協(xié)議棧就會(huì)直接將連接斷開了。其中，發(fā)出心跳包收不到ACK的超時(shí)時(shí)間稱為keepaliveinterval，Windows系統(tǒng)中默認(rèn)是1秒，可配置；收不到心跳包對(duì)應(yīng)的ACK包的重發(fā)次數(shù)probe，Windows系統(tǒng)是固定的，是固定的10次，不可配置的。

所以TCPIP協(xié)議棧的心跳機(jī)制也能檢測(cè)出網(wǎng)絡(luò)異常，不過在默認(rèn)配置下可能需要很久才能檢測(cè)出來，除非網(wǎng)絡(luò)異常出現(xiàn)在正在發(fā)送心跳包后等待對(duì)端的回應(yīng)時(shí)，這種情況下如果多次重發(fā)心跳包都收不到ACK回應(yīng)，協(xié)議棧就會(huì)判斷網(wǎng)絡(luò)出故障，主動(dòng)將連接關(guān)閉掉。

2.3、修改TCPIP協(xié)議棧的默認(rèn)心跳參數(shù)

TCPIP協(xié)議棧的默認(rèn)心跳機(jī)制的開啟，不是給系統(tǒng)整個(gè)協(xié)議棧開啟心跳監(jiān)測(cè)，而是對(duì)某個(gè)socket套接字開啟。

開啟心跳機(jī)制后，還可以修改心跳的時(shí)間參數(shù)。從代碼上看，先調(diào)用setsockopt給目標(biāo)套接字開啟心跳監(jiān)測(cè)機(jī)制，再調(diào)用WSAIoctl去修改心跳檢測(cè)的默認(rèn)時(shí)間參數(shù)，相關(guān)代碼如下所示：

SOCKET socket;
// ......(中間代碼省略)

int optval = 1;
int nRet = setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE, (const char *)&optval,
sizeof(optval));
if (nRet != 0)
return;

tcp_keepalive alive;
alive.onoff = TRUE;
alive.keepalivetime = 101000;
alive.keepaliveinterval = 21000;

DWORD dwBytesRet = 0;
nRet = WSAIoctl(socket, SIO_KEEPALIVE_VALS, &alive, sizeof(alive), NULL, 0,
&dwBytesRet, NULL, NULL);
if (nRet != 0)
return;

上面的代碼可以看到，先調(diào)用setsockopt函數(shù)，傳入SO_KEEPALIVE參數(shù)，打開TCP連接的心跳開關(guān)，此時(shí)心跳參數(shù)使用系統(tǒng)默認(rèn)的心跳參數(shù)值。緊接著，調(diào)用WSAIoCtrl函數(shù)，傳入SIO_KEEPALIVE_VALS參數(shù)，同時(shí)將設(shè)置好時(shí)間值的心跳參數(shù)結(jié)構(gòu)體傳進(jìn)去。

下面對(duì)心跳參數(shù)結(jié)構(gòu)體tcp_keepalive做個(gè)詳細(xì)的說明：（以Windows系統(tǒng)為例）

1）keepalivetime：默認(rèn)2小時(shí)發(fā)送一次心跳保活包，比如發(fā)送第1個(gè)保活包之后，間隔2個(gè)小時(shí)后再發(fā)起下一個(gè)?；畎?。如果這期間有數(shù)據(jù)交互，也算是有效的保活包，這個(gè)時(shí)間段就不再發(fā)送?；畎?，發(fā)送下個(gè)?；畎臅r(shí)間間隔會(huì)從收發(fā)的最后一條數(shù)據(jù)的時(shí)刻開始重新從0計(jì)時(shí)。2）keepaliveinterval：發(fā)送?；畎?，沒有收到對(duì)端的ack的超時(shí)時(shí)間默認(rèn)為1秒。假設(shè)和對(duì)端的網(wǎng)絡(luò)出問題了，給對(duì)端發(fā)送第1個(gè)?；畎?，1秒內(nèi)沒有收到對(duì)端的ack，則發(fā)第2個(gè)保活包，1秒內(nèi)沒有收到對(duì)端的保活包，再發(fā)送下一個(gè)?；畎?，.....，直到發(fā)送第10個(gè)保活包后，1秒鐘還沒收到ack回應(yīng)，則達(dá)到發(fā)送10次保活包的探測(cè)次數(shù)上限，則認(rèn)為網(wǎng)絡(luò)出問題了。3）probe探測(cè)次數(shù)：Windows系統(tǒng)上的探測(cè)次數(shù)被固定為10次，不可修改。

MSDN上對(duì)心跳機(jī)制檢測(cè)出的網(wǎng)絡(luò)異常的說明如下：

If a connection is dropped as the result of keep-alives the error code WSAENETRESET is returned to any calls in progress on the socket, and any subsequent calls will fail with WSAENOTCONN.

因?yàn)楸；畲螖?shù)達(dá)到上限導(dǎo)致連接被丟棄掉，所有正在調(diào)用中的套接字接口會(huì)返回WSAENETRESET錯(cuò)誤碼，后續(xù)的套接字api函數(shù)的調(diào)用都會(huì)返回WSAENOTCONN。

3、libwebsockets開源庫中的心跳機(jī)制使用的就是TCPIP協(xié)議棧的心跳機(jī)制

我們的產(chǎn)品之前在使用websocket時(shí)，就遇到?jīng)]有設(shè)置心跳機(jī)制導(dǎo)致TCP長(zhǎng)連接被網(wǎng)絡(luò)設(shè)備無故釋放的問題。

我們客戶端程序在登錄時(shí)，會(huì)去連接某業(yè)務(wù)的注冊(cè)服務(wù)器，建立的是websocket長(zhǎng)連接。這個(gè)長(zhǎng)連接一直保持著，只有使用該業(yè)務(wù)模塊的業(yè)務(wù)時(shí)才會(huì)使用到該連接，在該連接上進(jìn)行數(shù)據(jù)交互。軟件登錄后，如果一直沒有操作該業(yè)務(wù)模塊的業(yè)務(wù)，這個(gè)長(zhǎng)連接會(huì)一直處于閑置狀態(tài)，即這個(gè)連接上沒有數(shù)據(jù)交互。

結(jié)果在某次測(cè)試過程中出現(xiàn)了問題，排查下來發(fā)現(xiàn)，這個(gè)長(zhǎng)連接因?yàn)殚L(zhǎng)時(shí)間沒有數(shù)據(jù)交互，被中間的網(wǎng)絡(luò)設(shè)備關(guān)閉了。后來為了解決這個(gè)問題，我們?cè)诔跏蓟痺ebsocket庫時(shí)設(shè)置心跳參數(shù)，這樣上述websocket長(zhǎng)連接在空閑的時(shí)候能跑一跑心跳包，這樣就能確保該長(zhǎng)連接不會(huì)因?yàn)殚L(zhǎng)時(shí)間沒有跑數(shù)據(jù)被無故關(guān)閉的問題了。

我們?cè)谡{(diào)用lws_create_context接口創(chuàng)建websockets會(huì)話上下文時(shí)，該接口的結(jié)構(gòu)體參數(shù)lws_context_creation_info中，有設(shè)置心跳參數(shù)的字段：

/**

• struct lws_context_creation_info - parameters to create context with
• This is also used to create vhosts.... if LWS_SERVER_OPTION_EXPLICIT_VHOSTS
• is not given, then for backwards compatibility one vhost is created at
• context-creation time using the info from this struct.
• If LWS_SERVER_OPTION_EXPLICIT_VHOSTS is given, then no vhosts are created
• at the same time as the context, they are expected to be created afterwards.
• @port: VHOST: Port to listen on... you can use CONTEXT_PORT_NO_LISTEN to

• suppress listening on any port, that's what you want if you are
  

• not running a websocket server at all but just using it as a
  

• client
  

• @iface: VHOST: NULL to bind the listen socket to all interfaces, or the

• interface name, eg, "eth2"
  

• If options specifies LWS_SERVER_OPTION_UNIX_SOCK, this member is
  

• the pathname of a UNIX domain socket. you can use the UNIX domain
  

• sockets in abstract namespace, by prepending an @ symbole to the
  

• socket name.
  

• @protocols: VHOST: Array of structures listing supported protocols and a protocol-

• specific callback for each one.  The list is ended with an
  

• entry that has a NULL callback pointer.
  

• It's not const because we write the owning_server member
  

• @extensions: VHOST: NULL or array of lws_extension structs listing the

• extensions this context supports.  If you configured with
  

• --without-extensions, you should give NULL here.
  

• @token_limits: CONTEXT: NULL or struct lws_token_limits pointer which is initialized

• with a token length limit for each possible WSI_TOKEN_***
  

• @ssl_cert_filepath: VHOST: If libwebsockets was compiled to use ssl, and you want

• to listen using SSL, set to the filepath to fetch the
  

• server cert from, otherwise NULL for unencrypted
  

• @ssl_private_key_filepath: VHOST: filepath to private key if wanting SSL mode;

• if this is set to NULL but sll_cert_filepath is set, the
  

• OPENSSL_CONTEXT_REQUIRES_PRIVATE_KEY callback is called
  

• to allow setting of the private key directly via openSSL
  

• library calls
  

• @ssl_ca_filepath: VHOST: CA certificate filepath or NULL
• @ssl_cipher_list: VHOST: List of valid ciphers to use (eg,

• "RC4-MD5:RC4-SHA:AES128-SHA:AES256-SHA:HIGH:!DSS:!aNULL"
  

• or you can leave it as NULL to get "DEFAULT"
  

• @http_proxy_address: VHOST: If non-NULL, attempts to proxy via the given address.

• If proxy auth is required, use format
  

• "username:password@server:port"
  

• @http_proxy_port: VHOST: If http_proxy_address was non-NULL, uses this port at

• the address
  

• @gid: CONTEXT: group id to change to after setting listen socket, or -1.
• @uid: CONTEXT: user id to change to after setting listen socket, or -1.
• @options: VHOST + CONTEXT: 0, or LWS_SERVER_OPTION_... bitfields
• @user: CONTEXT: optional user pointer that can be recovered via the context

• pointer using lws_context_user
  

• @ka_time: CONTEXT: 0 for no keepalive, otherwise apply this keepalive timeout to

• all libwebsocket sockets, client or server
  

• @ka_probes: CONTEXT: if ka_time was nonzero, after the timeout expires how many

• times to try to get a response from the peer before giving up
  

• and killing the connection
  

• @ka_interval: CONTEXT: if ka_time was nonzero, how long to wait before each ka_probes

• attempt
  

• @provided_client_ssl_ctx: CONTEXT: If non-null, swap out libwebsockets ssl

• implementation for the one provided by provided_ssl_ctx.
  

• Libwebsockets no longer is responsible for freeing the context
  

• if this option is selected.
  

• @max_http_header_data: CONTEXT: The max amount of header payload that can be handled

• in an http request (unrecognized header payload is dropped)
  

• @max_http_header_pool: CONTEXT: The max number of connections with http headers that

• can be processed simultaneously (the corresponding memory is
  

• allocated for the lifetime of the context).  If the pool is
  

• busy new incoming connections must wait for accept until one
  

• becomes free.
  

• @count_threads: CONTEXT: how many contexts to create in an array, 0 = 1
• @fd_limit_per_thread: CONTEXT: nonzero means restrict each service thread to this

• many fds, 0 means the default which is divide the process fd
  

• limit by the number of threads.
  

• @timeout_secs: VHOST: various processes involving network roundtrips in the

• library are protected from hanging forever by timeouts.  If
  

• nonzero, this member lets you set the timeout used in seconds.
  

• Otherwise a default timeout is used.
  

• @ecdh_curve: VHOST: if NULL, defaults to initializing server with "prime256v1"
• @vhost_name: VHOST: name of vhost, must match external DNS name used to

• access the site, like "warmcat.com" as it's used to match
  

• Host: header and / or SNI name for SSL.
  

• @plugin_dirs: CONTEXT: NULL, or NULL-terminated array of directories to

• scan for lws protocol plugins at context creation time
  

• @pvo: VHOST: pointer to optional linked list of per-vhost

• options made accessible to protocols
  

• @keepalive_timeout: VHOST: (default = 0 = 60s) seconds to allow remote

• client to hold on to an idle HTTP/1.1 connection
  

• @log_filepath: VHOST: filepath to append logs to... this is opened before

• any dropping of initial privileges
  

• @mounts: VHOST: optional linked list of mounts for this vhost
• @server_string: CONTEXT: string used in HTTP headers to identify server

• software, if NULL, "libwebsockets".
  

/
struct lws_context_creation_info {
int port; / VH */
const char iface; / VH */
const struct lws_protocols protocols; / VH */
const struct lws_extension extensions; / VH */
const struct lws_token_limits token_limits; / context */
const char ssl_private_key_password; / VH */
const char ssl_cert_filepath; / VH */
const char ssl_private_key_filepath; / VH */
const char ssl_ca_filepath; / VH */
const char ssl_cipher_list; / VH */
const char http_proxy_address; / VH /
unsigned int http_proxy_port; / VH /
int gid; / context /
int uid; / context /
unsigned int options; / VH + context */
void user; / context /
int ka_time; / context /
int ka_probes; / context /
int ka_interval; / context */
#ifdef LWS_OPENSSL_SUPPORT
SSL_CTX provided_client_ssl_ctx; / context /
#else / maintain structure layout either way */
void provided_client_ssl_ctx;
#endif
short max_http_header_data; / context /
short max_http_header_pool; / context /
unsigned int count_threads; / context /
unsigned int fd_limit_per_thread; / context /
unsigned int timeout_secs; / VH */
const char ecdh_curve; / VH */
const char vhost_name; / VH */
const char * const plugin_dirs; / context */
const struct lws_protocol_vhost_options pvo; / VH /
int keepalive_timeout; / VH */
const char log_filepath; / VH */
const struct lws_http_mount mounts; / VH */
const char server_string; / context /
/ Add new things just above here ---^

• This is part of the ABI, don't needlessly break compatibility
• The below is to ensure later library versions with new
• members added above will see 0 (default) even if the app
• was not built against the newer headers.
  */
  void *_unused[8];
  };

其中的ka_time、ka_probes和ka_interval三個(gè)字段就是心跳相關(guān)的設(shè)置參數(shù)。我們初始化websockets上下文的代碼如下：

static lws_context* CreateContext()
{
lws_set_log_level( 0xFF, NULL );
lws_context* plcContext = NULL;

lws_context_creation_info tCreateinfo;
memset(&tCreateinfo, 0, sizeof tCreateinfo);

tCreateinfo.port = CONTEXT_PORT_NO_LISTEN;
tCreateinfo.protocols = protocols;
tCreateinfo.ka_time = LWS_TCP_KEEPALIVE_TIME;
tCreateinfo.ka_interval = LWS_TCP_KEEPALIVE_INTERVAL;
tCreateinfo.ka_probes = LWS_TCP_KEEPALIVE_PROBES;
tCreateinfo.options = LWS_SERVER_OPTION_DISABLE_IPV6;

plcContext = lws_create_context(&tCreateinfo);

return plcContext;

}

通過查閱libwebsockets開源庫代碼得知，此處設(shè)置的心跳使用的就是TCPIP協(xié)議棧的心跳機(jī)制，如下所示：

LWS_VISIBLE int
lws_plat_set_socket_options(struct lws_vhost *vhost, lws_sockfd_type fd)
{
int optval = 1;
int optlen = sizeof(optval);
u_long optl = 1;
DWORD dwBytesRet;
struct tcp_keepalive alive;
int protonbr;
#ifndef _WIN32_WCE
struct protoent *tcp_proto;
#endif

if (vhost->ka_time) {
	/* enable keepalive on this socket */
	// 先調(diào)用setsockopt打開發(fā)送心跳包（設(shè)置）選項(xiàng)
	optval = 1;
	if (setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE,
				     (const char *)&optval, optlen) < 0)
		return 1;

	alive.onoff = TRUE;
	alive.keepalivetime = vhost->ka_time*1000;
	alive.keepaliveinterval = vhost->ka_interval*1000;

	if (WSAIoctl(fd, SIO_KEEPALIVE_VALS, &alive, sizeof(alive),
				      NULL, 0, &dwBytesRet, NULL, NULL))
		return 1;
}

/* Disable Nagle */
optval = 1;

#ifndef _WIN32_WCE
tcp_proto = getprotobyname("TCP");
if (!tcp_proto) {
lwsl_err("getprotobyname() failed with error %dn", LWS_ERRNO);
return 1;
}
protonbr = tcp_proto->p_proto;
#else
protonbr = 6;
#endif

setsockopt(fd, protonbr, TCP_NODELAY, (const char *)&optval, optlen);

/* We are nonblocking... */
ioctlsocket(fd, FIONBIO, &optl);

return 0;

}

4、TCPIP丟包重傳機(jī)制

如果網(wǎng)絡(luò)出故障時(shí)，客戶端與服務(wù)器之間正在進(jìn)行TCP數(shù)據(jù)交互，客戶端給服務(wù)器發(fā)送數(shù)據(jù)包后因?yàn)榫W(wǎng)絡(luò)故障收不到服務(wù)器的ACK包，就會(huì)觸發(fā)客戶端的TCP丟包重傳，丟包重傳機(jī)制也能判斷出網(wǎng)絡(luò)出現(xiàn)異常。

對(duì)于TCP連接，客戶端給服務(wù)器發(fā)送數(shù)據(jù)后沒有收到服務(wù)器的ACK包，會(huì)觸發(fā)丟包重傳。每次重傳的時(shí)間間隔會(huì)加倍，當(dāng)重傳次數(shù)達(dá)到系統(tǒng)上限（Windows默認(rèn)的上限是5次，Linux默認(rèn)的上限是15次）后，協(xié)議棧就認(rèn)為網(wǎng)絡(luò)出故障了，會(huì)直接將對(duì)應(yīng)的連接關(guān)閉了。

所以當(dāng)網(wǎng)絡(luò)出現(xiàn)故障時(shí)有數(shù)據(jù)交互，協(xié)議棧會(huì)在數(shù)十秒內(nèi)檢測(cè)到網(wǎng)路出現(xiàn)異常，就會(huì)直接將連接直接關(guān)閉掉。丟包重傳機(jī)制的詳細(xì)描述如下所示：

對(duì)于丟包重傳機(jī)制，可以通過給PC插拔網(wǎng)線來查看，可以使用wireshark抓包看一下?？焖俨灏尉W(wǎng)線時(shí)（先拔掉網(wǎng)線，等待幾秒鐘再將網(wǎng)線插上），給服務(wù)器發(fā)送的操作指令會(huì)因?yàn)閬G包重傳會(huì)收到數(shù)據(jù)的。

5、使用非阻塞socket和select接口實(shí)現(xiàn)connect連接的超時(shí)控制

5.1、MSDN上對(duì)connect和select接口的說明

對(duì)于tcp套接字，我們需要調(diào)用套接字函數(shù)connect去建立TCP連接。我們先來看看微軟MSDN上對(duì)套接字接口connect的描述：

On a blocking socket, the return value indicates success or failure of the connection attempt.

對(duì)于阻塞式的socket，通過connect的返回值就能確定有沒有連接成功，返回0表示連接成功。

With a nonblocking socket, the connection attempt cannot be completed immediately. In this case, connect will return SOCKET_ERROR, and WSAGetLastError will return WSAEWOULDBLOCK. In this case, there are three possible scenarios:

Use the select function to determine the completion of the connection request by checking to see if the socket is writeable.

對(duì)于非組賽式的socket，connect調(diào)用會(huì)立即返回，但連接操作還沒有完成。connect返回SOCKET_ERROR，對(duì)于非阻塞式socket，返回SOCKET_ERROR并不表示失敗，需要調(diào)用WSAGetLastError獲取connect函數(shù)執(zhí)行后的LastError值，一般此時(shí)WSAGetLastError會(huì)返回WSAEWOULDBLOCK：

表明連接正在進(jìn)行中。可以使用select接口檢測(cè)一下套接字是否可寫（套接字是否在writefds集合中），如果可寫，則表示連接成功。如果套接字在exceptfds集合中，則說明連接出現(xiàn)了異常，如下所示：

5.2、使用非阻塞socket和select實(shí)現(xiàn)連接超時(shí)的控制

對(duì)于阻塞式的socket，在Windows下，如果遠(yuǎn)端的IP和Port不可達(dá)，則會(huì)阻塞75s后返回SOCKET_ERROR，表明連接失敗。所以當(dāng)我們測(cè)試遠(yuǎn)端的IP和Port是否可以連接時(shí)，我們不使用阻塞式的socket，而是使用非阻塞式socket，然后調(diào)用select，通過select添加連接超時(shí)時(shí)間，實(shí)現(xiàn)連接超時(shí)的控制。

select函數(shù)因?yàn)槌瑫r(shí)返回，會(huì)返回0；如果發(fā)生錯(cuò)誤，則返回SOCKET_ERROR，所以判斷時(shí)要判斷select返回值，如果小于等于0，則是連接失敗，立即將套接字關(guān)閉掉。如果select返回值大于0，則該返回值是已經(jīng)準(zhǔn)備就緒的socket個(gè)數(shù)，比如連接成功的socket。我們判斷套接字是否在可寫集合writefds中，如果在該集合中，則表示連接成功。

根據(jù)MSDN上的相關(guān)描述，我們就能大概知道該如何實(shí)現(xiàn)connect的超時(shí)控制了，相關(guān)代碼如下：

bool ConnectDevice( char* pszIP, int nPort )
{
// 創(chuàng)建TCP套接字
SOCKET connSock = socket(AF_INET, SOCK_STREAM, 0);
if (connSock == INVALID_SOCKET)
{
return false;
}

// 填充IP和端口
SOCKADDR_IN devAddr;
memset(&devAddr, 0, sizeof(SOCKADDR_IN));
devAddr.sin_family = AF_INET;
devAddr.sin_port = htons(nPort);
devAddr.sin_addr.s_addr = inet_addr(pszIP);

// 將套接字設(shè)置為非阻塞式的，為下面的select做準(zhǔn)備
unsigned long ulnoblock = 1;
ioctlsocket(connSock, FIONBIO, &ulnoblock);

// 發(fā)起connnect，該接口立即返回
connect(connSock, (sockaddr*)&devAddr, sizeof(devAddr));

FD_SET writefds;
FD_ZERO(&writefds);
FD_SET(connSock, &writefds);

// 設(shè)置連接超時(shí)時(shí)間為1秒
timeval tv;
tv.tv_sec = 1; //超時(shí)1s
tv.tv_usec = 0;
// The select function returns the total number of socket handles that are ready and contained 
// in the fd_set structures, zero if the time limit expired, or SOCKET_ERROR(-1) if an error occurred. 
if (select(0, NULL, &writefds, NULL, &tv) <= 0)
{
	closesocket(connSock);
	return false; //超時(shí)未連接上就退出
}

ulnoblock = 0;
ioctlsocket(connSock, FIONBIO, &ulnoblock);

closesocket(connSock);
return true;

}