在unix系统中，socket和普通文件一样对待，因为它可以像普通文件一样被读和写，但是它还有一些自己独特的特点，例如，文件的读写位置可以设置，但是socket只能被顺序的读写等等，那么在unix系统中，是如何实现这种方式的呢？
如下图，其中有以下重要数据结构：proc、filedesc、file等，对这些重要数据结构及其之间的关系弄清楚之后，上面的问题自然就有答案了。在本文介绍中，使用的操作系统源码为：4.4bsd-lite版本，该版本是《TCP/IP协议卷2——实现》一书使用的源码，同时该源码相对于目前使用的linux操作系统来更为小巧、简单，更适合学习。
（1）每个进程在OS中都有一个数据结构（struct filedesc）为之对应，通常称该数据结构为PCB（进程控制块），该数据结构中详细定义了控制该进程所需的全部数据，这里只需重点关注其成员：struct filedesc *p_fd，该成员指向一个文件描述信息的数据结构。struct filedesc定义在文件4.4BSD-Lite\sys\sys\filedesc.h中，详细定义如下：
struct proc {
struct proc *p_forw; /* Doubly-linked run/sleep queue. */
struct proc *p_back;
struct proc *p_next; /* Linked list of active procs */
struct proc **p_prev; /* and zombies. */
/* substructures: */
struct pcred *p_cred; /* Process owner's identity. */
struct filedesc *p_fd; /* Ptr to open files structure. */
struct pstats *p_stats; /* Accounting/statistics (PROC ONLY). */
struct plimit *p_limit; /* Process limits. */
struct vmspace *p_vmspace; /* Address space. */
struct sigacts *p_sigacts; /* Signal actions, state (PROC ONLY). */
#define p_ucred p_cred->pc_ucred
#define p_rlimit p_limit->pl_rlimit
int p_flag; /* P_* flags. */
char p_stat; /* S* process status. */
char p_pad1[3];
pid_t p_pid; /* Process identifier. */
struct proc *p_hash; /* Hashed based on p_pid for kill+exit+... */
struct proc *p_pgrpnxt; /* Pointer to next process in process group. */
struct proc *p_pptr; /* Pointer to process structure of parent. */
struct proc *p_osptr; /* Pointer to older sibling processes. */
/* The following fields are all zeroed upon creation in fork. */
#define p_startzero p_ysptr
struct proc *p_ysptr; /* Pointer to younger siblings. */
struct proc *p_cptr; /* Pointer to youngest living child. */
pid_t p_oppid; /* Save parent pid during ptrace. XXX */
int p_dupfd; /* Sideways return value from fdopen. XXX */
/* scheduling */
u_int p_estcpu; /* Time averaged value of p_cpticks. */
int p_cpticks; /* Ticks of cpu time. */
fixpt_t p_pctcpu; /* %cpu for this process during p_swtime */
void *p_wchan; /* Sleep address. */
char *p_wmesg; /* Reason for sleep. */
u_int p_swtime; /* Time swapped in or out. */
u_int p_slptime; /* Time since last blocked. */
struct itimerval p_realtimer; /* Alarm timer. */
struct timeval p_rtime; /* Real time. */
u_quad_t p_uticks; /* Statclock hits in user mode. */
u_quad_t p_sticks; /* Statclock hits in system mode. */
u_quad_t p_iticks; /* Statclock hits processing intr. */
int p_traceflag; /* Kernel trace points. */
struct vnode *p_tracep; /* Trace to vnode. */
int p_siglist; /* Signals arrived but not delivered. */
struct vnode *p_textvp; /* Vnode of executable. */
long p_spare[5]; /* pad to 256, avoid shifting eproc. */
/* End area that is zeroed on creation. */
#define p_endzero p_startcopy
/* The following fields are all copied upon creation in fork. */
#define p_startcopy p_sigmask
sigset_t p_sigmask; /* Current signal mask. */
sigset_t p_sigignore; /* Signals being ignored. */
sigset_t p_sigcatch; /* Signals being caught by user. */
u_char p_priority; /* Process priority. */
u_char p_usrpri; /* User-priority based on p_cpu and p_nice. */
char p_nice; /* Process "nice" value. */
char p_comm[MAXCOMLEN+1];
struct pgrp *p_pgrp; /* Pointer to process group. */
/* End area that is copied on creation. */
#define p_endcopy p_thread
int p_thread; /* Id for this "thread"; Mach glue. XXX */
struct user *p_addr; /* Kernel virtual addr of u-area (PROC ONLY). */
struct mdproc p_md; /* Any machine-dependent fields. */
u_short p_xstat; /* Exit status for wait; also stop signal. */
u_short p_acflag; /* Accounting flags. */
struct rusage *p_ru; /* Exit information. XXX */
};
（2）struct filedesc结构体描述了进程打开的所有的文件信息，这里需要重点关注该结构体的两个数组成员：struct file **fd_ofiles和char *fd_ofileflags（见上图），其中：数组fd_ofiles的每个成员对应一个当前进程打开的文件结构体的地址；数组fd_ofileflags的每个成员对应当前进程打开的一个文件的描述符标志，文件描述符的标志采用bit位表示，因此一个打开的文件共有8个bit位表示8中不同的标志，例如标志close-on-exec和标志mapped-from-device。另外，这两个数组的成员是对应的，即：
fd_ofiles[n]对应打开当前进程打开的第n个文件的文件结构体地址；
fd_ofileflags[n]对应当前进程打开的第n个文件的描述符标志；
struct filedesc的定义在文件4.4BSD-Lite\sys\sys\filedesc.h中，其详细定义如下：
struct filedesc {
struct file **fd_ofiles; /* file structures for open files */
char *fd_ofileflags; /* per-process open file flags */
struct vnode *fd_cdir; /* current directory */
struct vnode *fd_rdir; /* root directory */
int fd_nfiles; /* number of open files allocated */
u_short fd_lastfile; /* high-water mark of fd_ofiles */
u_short fd_freefile; /* approx. next free file */
u_short fd_cmask; /* mask for file creation */
u_short fd_refcnt; /* reference count */
};
（3）结构体struct file表示当前进程中一个打开的文件，这里将关注其成员short f_type、struct fileops *f_ops和caddr_t f_data。其中：
short f_type表示当前打开文件的类型，在文件4.4BSD-Lite\usr\src\sys\sys\file.h中定义了成员shortf_type的两种类型：
#define DTYPE_VNODE 1 /*file */
#define DTYPE_SOCKET 2 /*communications endpoint */
例如其值为DTYPE_SOCKET表示当前打开的文件是一个socket，值为DTYPE_VNODE表示一个普通的文件等等；
struct fileops *f_ops定义了5个函数指针，它们将根据具体的文件类型（f_type表示文件类型）指向具体的函数，例如当前打开的文件为socket时（f_type值为DTYPE_SOCKET），这5个函数指针将指向以下5个操作socket的函数：soo_read、soo_write、soo_ioctl、soo_select、soo_close，当打开的文件为普通文件时（f_type值为DTYPE_VNODE），这5个函数指针将指向以下5个操作socket的函数：vn_read、vn_write、vn_ioctl、vn_select、vn_close；
caddr_t f_data对应了该打开文件所对应的数据部分，对应vnode或者socket结构体；这里类型caddr_t 实质上是一个char*，其定义为：typedef char * caddr_t;。
struct file定义在文件4.4BSD-Lite\sys\sys\file.h中，其详细定义如下：
struct file {
struct file *f_filef; /* list of active files */
struct file **f_fileb; /* list of active files */
short f_flag; /* see fcntl.h */
#define DTYPE_VNODE 1 /* file */
#define DTYPE_SOCKET 2 /* communications endpoint */
short f_type; /* descriptor type */
short f_count; /* reference count */
short f_msgcount; /* references from message queue */
struct ucred *f_cred; /* credentials associated with descriptor */
struct fileops {
int (*fo_read) __P((struct file *fp, struct uio *uio,
struct ucred *cred));
int (*fo_write) __P((struct file *fp, struct uio *uio,
struct ucred *cred));
int (*fo_ioctl) __P((struct file *fp, int com,
caddr_t data, struct proc *p));
int (*fo_select) __P((struct file *fp, int which,
struct proc *p));
int (*fo_close) __P((struct file *fp, struct proc *p));
} *f_ops;
off_t f_offset;
caddr_t f_data; /* vnode or socket */
};
（4）caddr_t f_data表示打开文件的实际数据部分，caddr_t的实际定义类型为char*，当打开的文件类型为普通文件时（f_type值为DTYPE_VNODE），f_data指向了一个struct vnode结构体，当打开的文件类型为socket时（f_type值为DTYPE_SOCKET），f_data指向了一个struct socket结构体，这里只需要先关注short so_type和caddr_t so_pcb两个成员变量，其中：short so_type表示socket类型，例如SOCK_DGRAM表示UDP类型，SOCK_STREAM表示TCP类型；caddr_t
so_pcb指向一个协议控制块的双向链表。
socket结构体的定义在文件：4.4BSD-Lite\sys\sys\socketvar.h中，其详细定义为：
struct socket {
short so_type; /* generic type, see socket.h */
short so_options; /* from socket call, see socket.h */
short so_linger; /* time to linger while closing */
short so_state; /* internal state flags SS_*, below */
caddr_t so_pcb; /* protocol control block */
struct protosw *so_proto; /* protocol handle */
/*
* Variables for connection queueing.
* Socket where accepts occur is so_head in all subsidiary sockets.
* If so_head is 0, socket is not related to an accept.
* For head socket so_q0 queues partially completed connections,
* while so_q is a queue of connections ready to be accepted.
* If a connection is aborted and it has so_head set, then
* it has to be pulled out of either so_q0 or so_q.
* We allow connections to queue up based on current queue lengths
* and limit on number of queued connections for this socket.
*/
struct socket *so_head; /* back pointer to accept socket */
struct socket *so_q0; /* queue of partial connections */
struct socket *so_q; /* queue of incoming connections */
short so_q0len; /* partials on so_q0 */
short so_qlen; /* number of connections on so_q */
short so_qlimit; /* max number queued connections */
short so_timeo; /* connection timeout */
u_short so_error; /* error affecting connection */
pid_t so_pgid; /* pgid for signals */
u_long so_oobmark; /* chars to oob mark */
/*
* Variables for socket buffering.
*/
struct sockbuf {
u_long sb_cc; /* actual chars in buffer */
u_long sb_hiwat; /* max actual char count */
u_long sb_mbcnt; /* chars of mbufs used */
u_long sb_mbmax; /* max chars of mbufs to use */
long sb_lowat; /* low water mark */
struct mbuf *sb_mb; /* the mbuf chain */
struct selinfo sb_sel; /* process selecting read/write */
short sb_flags; /* flags, see below */
short sb_timeo; /* timeout for read/write */
} so_rcv, so_snd;
#define SB_MAX (256*1024) /* default for max chars in sockbuf */
#define SB_LOCK 0x01 /* lock on data queue */
#define SB_WANT 0x02 /* someone is waiting to lock */
#define SB_WAIT 0x04 /* someone is waiting for data/space */
#define SB_SEL 0x08 /* someone is selecting */
#define SB_ASYNC 0x10 /* ASYNC I/O, need signals */
#define SB_NOTIFY (SB_WAIT|SB_SEL|SB_ASYNC)
#define SB_NOINTR 0x40 /* operations not interruptible */
caddr_t so_tpcb; /* Wisc. protocol control block XXX */
void (*so_upcall) __P((struct socket *so, caddr_t arg, int waitf));
caddr_t so_upcallarg; /* Arg for above */
};
（5）协议控制块caddr_tso_pcb，在socket结构体中，协议控制块是非常核心的数据结构，采用双向链表方式表示，它包含以下成员：前一个inpcb、后一个inpcb、inpcb双向链表的首节点、当前inpcb对应socket的本地IP、本地端口、远端IP、远端端口、当前inpcb对应的socket结构体的地址等等；
每个socket都有一个协议控制块inpub与之对应：可通过socket的so_pcb成员来访问它，同时协议控制块中也有个成员struct socket *inp_socket用于指向自己所属的socket结构体。
在OS中，每种类型的socket有且只有一个inpcb链表与之对应，例如：所有的TCP的socket的inpcb都在同一个TCP的inpcb双向链表中，所有的UDP的socket的inpcb都在同一个UDP的inpcb双向链表中。
在通过socket接收数据时，OS从网卡驱动中拿到数据后，先搜索inpcb的双向链表，通过比对本地ip地址、本地端口号、远端ip地址、远端端口号找到匹配的inpcb，进而找到对应socket，并将数据保存到socket的接收缓存中。
协议控制块struct inp定义在文件4.4BSD-Lite\sys\netinet\in_pcb.h中，其详细定义为：
struct inpcb {
struct inpcb *inp_next,*inp_prev;
/* pointers to other pcb's */
struct inpcb *inp_head; /* pointer back to chain of inpcb's
for this protocol */
struct in_addr inp_faddr; /* foreign host table entry */
u_short inp_fport; /* foreign port */
struct in_addr inp_laddr; /* local host table entry */
u_short inp_lport; /* local port */
struct socket *inp_socket; /* back pointer to socket */
caddr_t inp_ppcb; /* pointer to per-protocol pcb */
struct route inp_route; /* placeholder for routing entry */
int inp_flags; /* generic IP/datagram flags */
struct ip inp_ip; /* header prototype; should have more */
struct mbuf *inp_options; /* IP options */
struct ip_moptions *inp_moptions; /* IP multicast options */
};