Shulou(Shulou.com)06/01 Report--

This article introduces you how to deeply understand the bind implementation of TCP/IP protocol, the content is very detailed, interested friends can refer to, hope to be helpful to you.

According to the order of socket network programming, we analyze bind functions in this article. We get a socket structure through the socket function. The logic of the bind function is actually relatively simple. It simply binds an address to the socket structure and simply assigns values to some of its fields. Talk is cheap . Show me the code .

Static int sock_bind (int fd, struct sockaddr * umyaddr, int addrlen)

{

Struct socket * sock

Int i

Char address[MAX _ SOCK_ADDR]

Int err

/ / find the corresponding socket through the file descriptor

If (! (sock = sockfd_lookup (fd, NULL)

Return (- ENOTSOCK)

If ((err=move_addr_to_kernel (umyaddr,addrlen,address)) ops- > bind (sock, (struct sockaddr *) address, addrlen))

< 0) { return(i); } return(0); } 主要是两个函数，我们一个个来。 1 sockfd_lookup 通过之前一些文章的分析，我们应该数socket和文件的内存布局比较熟悉了。下面的代码不难理解。就是根据文件描述符从pcb中找到inode节点。因为inode节点里保存了socket结构体，所以最后返回fd对应的socke结构体就行。 // 通过fd找到file结构体，从而找到inode节点，最后找到socket结构体 static inline struct socket *sockfd_lookup(int fd, struct file **pfile) { struct file *file; struct inode *inode; if (fd < 0 || fd >

= NR_OPEN | |! (file = current- > files- > FDD [FD]))

Return NULL

Inode = file- > f_inode

If (! inode | |! inode- > i_sock)

Return NULL

If (pfile)

* pfile = file

Return socki_lookup (inode)

}

/ / inode and socket refer to each other

Inline struct socket * socki_lookup (struct inode * inode)

{

Return & inode- > u.socket_i

}

2 sock- > ops- > bind

When we review the socket article, we can see that there are a series of operation functions in the socket structure. assuming that the protocol suite is ipv4, then the bind function is the inet_bind function (omitting some of the code).

/ / bind an address to socket

Static int inet_bind (struct socket * sock, struct sockaddr * uaddr

Int addr_len)

{

Struct sockaddr_in * addr= (struct sockaddr_in *) uaddr

/ / get the underlying sock structure

Struct sock * sk= (struct sock *) sock- > data, * sk2

Unsigned short snum = 0 / * Stoopid compiler.. This IS ok * /

Int chk_addr_ret

/ / these data of raw protocol are populated by the user.

If (sock- > type! = SOCK_RAW)

{/ / the port has been bound

If (sk- > num! = 0)

Return (- EINVAL)

Snum = ntohs (addr- > sin_port)

/ / if the port is invalid, get a port that is not available for root at random.

If (snum = = 0)

{

Snum = get_new_socknum (sk- > prot, 0)

}

/ / Super user privileges are required for ports less than 1024

If (snum

< PROT_SOCK && !suser()) return(-EACCES); } // 判断ip chk_addr_ret = ip_chk_addr(addr->

Sin_addr.s_addr)

/ / illegal address

If (addr- > sin_addr.s_addr! = 0 & & chk_addr_ret! = IS_MYADDR & & chk_addr_ret! = IS_MULTICAST)

Return (- EADDRNOTAVAIL); / * Source address MUST be ours! * /

/ / record ip

If (chk_addr_ret | | addr- > sin_addr.s_addr = = 0)

Sk- > saddr = addr- > sin_addr.s_addr

If (sock- > type! = SOCK_RAW)

{

/ * Make sure we are allowed to bind here. , /

Cli ()

/ / traverses the hash table. Hash table conflict resolution is the chain address method, which verifies the validity of the bound port.

For (sk2 = sk- > prot- > sock_ Array [snum & (SOCK_ARRAY_SIZE-1)]

Sk2! = NULL; sk2 = sk2- > next)

{

/ / the port has not been bound yet. Check the next one directly.

If (sk2- > num! = snum)

Continue

/ / Port has been used and no reusable flag has been set, such as whether it can be reused in 2msl after disconnection, through the setsockopt function.

If (! sk- > reuse)

{

Sti ()

Return (- EADDRINUSE)

}

/ / Port is the same, but ip is different. Ok, next

If (sk2- > saddr! = sk- > saddr)

Continue; / * socket per slot!-FB * /

/ / both the port and ip are the same. The monitored port cannot be used at the same time

If (! sk2- > reuse | | sk2- > state==TCP_LISTEN)

{

Sti ()

Return (- EADDRINUSE)

}

}

Sti ()

/ / make sure that the sk is not in the sock_array queue

Remove_sock (sk)

/ / Mount to sock_array

Put_sock (snum, sk)

/ / Source port in the tcp header

Sk- > dummy_th.source = ntohs (sk- > num)

Sk- > daddr = 0

Sk- > dummy_th.dest = 0

}

Return (0)

}

The bind function mainly makes a check on the bound ip and port, which is recorded in the sock structure when it is legal. And mount the sock structure into a global hash table.

On how to in-depth understanding of the TCP/IP protocol bind implementation to share here, I hope that the above content can be of some help to you, can learn more knowledge. If you think the article is good, you can share it for more people to see.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.