zoukankan      html  css  js  c++  java
  • 套接字源代码分析

    我是非常喜欢linux内核的,作为世界上最伟大的开源软件(我认为)。随手可得的最新版本号的源码。有那么多大牛在维护与更新。读读它。真的对
    

    我的帮助特别大。零零散散的非常久了。如今想要学习网络,学习网络就须要懂套接字编程。也就是去学习一大堆API的使用方法,可是这样非常easy忘记也没有什么
    价值。我认为最好的学习套接字的方法。就是去读源码。网络协议栈是linux源码中比較庞大的一部分。特别是比較新的版本号,所以能够选择一些低版本号的
    源码来研读。尽管代码一直在迭代,可是核心思想是不会有太大的改变的,特别是函数接口,由于那是标准,所以仅仅要不是特别须要,函数接口是不会改变的。
    接下来的时间里面,我将系统的对linux 1.2.13的源码做读书笔记,可是顺序可能不确定,有可能是先看网络。然后再看内存管理。但我个人认为,读
    操作系统的源码,首先要弄明确的是内存管理。这是操作系统执行的基础,仅仅有弄明确了内存管理,才干去理解其它的部分如进程管理等,还有。我认为有些
    部分能够有选择的读,比方设备管理,我就不打算去读。由于太费时间,须要去了解硬件,不到迫不得已,我决定不去读设备管理部分,文件系统不须要读,当然
    选择一个就能够了,比方能够选择FAT。或者EXT,或者minux,总之要去读,由于这是文件操作的基础,重点是内存管理。特别easy迷失在内核空间与用户空间
    之间。搞明确特别重要。ipc一定要看。文件系统一定要看,网络也要看。


    好了,今天开个头,非常明显今天要选择的部分是网络,也就是套接字,暂且不正确网络协议栈作更深入的研究,我今天仅仅读那些在套机字编程中须要的函数,
    对于每一个函数,先了解功能。然后分析源码。这样会比較深刻。


    须要知道的是。套接字文件实如今BSD层。能够在文件linux/net/sock.c里面看到套接字操作的第一层函数,对于底层调用,我们暂且不深入,由于今天
    的目标是明确函数是用来干嘛的,而不是怎样实现的。假设要想知道一个函数比方send函数是怎样实现的。就须要到INET层去寻找,然后再继续追踪下去,
    由于linux的设计是分层的。所以有一些函数实现特别简单,这是简单的调用下层函数。

    以下是对socket.c这个文件的凝视

    
    /*
     * NET      An implementation of the SOCKET network access protocol.
     *
     * Version: @(#)socket.c    1.1.93  18/02/95
     *
     * Authors: Orest Zborowski, <obz@Kodak.COM>
     *      Ross Biro, <bir7@leland.Stanford.Edu>
     *      Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
     *
     * Fixes:
     *      Anonymous   :   NOTSOCK/BADF cleanup. Error fix in
     *                  shutdown()
     *      Alan Cox    :   verify_area() fixes
     *      Alan Cox    :   Removed DDI
     *      Jonathan Kamens :   SOCK_DGRAM reconnect bug
     *      Alan Cox    :   Moved a load of checks to the very
     *                  top level.
     *      Alan Cox    :   Move address structures to/from user
     *                  mode above the protocol layers.
     *      Rob Janssen :   Allow 0 length sends.
     *      Alan Cox    :   Asynchronous I/O support (cribbed from the
     *                  tty drivers).
     *      Niibe Yutaka    :   Asynchronous I/O for writes (4.4BSD style)
     *      Jeff Uphoff :   Made max number of sockets command-line
     *                  configurable.
     *      Matti Aarnio    :   Made the number of sockets dynamic,
     *                  to be allocated when needed, and mr.
     *                  Uphoff's max is used as max to be
     *                  allowed to allocate.
     *      Linus       :   Argh. removed all the socket allocation
     *                  altogether: it's in the inode now.
     *      Alan Cox    :   Made sock_alloc()/sock_release() public
     *                  for NetROM and future kernel nfsd type
     *                  stuff.
     *
     *
     *      This program is free software; you can redistribute it and/or
     *      modify it under the terms of the GNU General Public License
     *      as published by the Free Software Foundation; either version
     *      2 of the License, or (at your option) any later version.
     *
     *
     *  This module is effectively the top level interface to the BSD socket
     *  paradigm. Because it is very simple it works well for Unix domain sockets,
     *  but requires a whole layer of substructure for the other protocols.
     *
     *  In addition it lacks an effective kernel -> kernel interface to go with
     *  the user one.
     */
    
    /*
       该文件为BSD socket层
    */
    
    
    #include <linux/config.h>
    #include <linux/signal.h>
    #include <linux/errno.h>
    #include <linux/sched.h>
    #include <linux/mm.h>
    #include <linux/kernel.h>
    #include <linux/major.h>
    #include <linux/stat.h>
    #include <linux/socket.h>
    #include <linux/fcntl.h>
    #include <linux/net.h>
    #include <linux/interrupt.h>
    #include <linux/netdevice.h>
    
    #include <asm/system.h>
    #include <asm/segment.h>
    
    /*
      以下是声明对网络数据提供普通文件操作的接口函数。
      以及相关数据结构的声明。

    file_operations结构定义了普通文件操作函数集合 系统中每一个文件相应一个file结构。file结构中有一个file_operations变量,当使用 write、read等函数对某个文件描写叙述符进行操作时,系统首先依据文件描写叙述符找到其相应的file 结构,然后调用其成员变量的file_operations中相应的函数来完毕请求 */ static int sock_lseek(struct inode *inode, struct file *file, off_t offset, int whence); static int sock_read(struct inode *inode, struct file *file, char *buf, int size); static int sock_write(struct inode *inode, struct file *file, char *buf, int size); static int sock_readdir(struct inode *inode, struct file *file, struct dirent *dirent, int count); static void sock_close(struct inode *inode, struct file *file); static int sock_select(struct inode *inode, struct file *file, int which, select_table *seltable); static int sock_ioctl(struct inode *inode, struct file *file, unsigned int cmd, unsigned long arg); static int sock_fasync(struct inode *inode, struct file *filp, int on); /* * Socket files have a set of 'special' operations as well as the generic file ones. These don't appear * in the operation structures but are done directly via the socketcall() multiplexor. */ /* 【hujian】 定义一个函数集合。这样当我们须要操作read这种操作时。 仅仅须要调用sock_read来完毕就好了。

    可是没有相似open的函数,由于socket函数就相当于完毕了open的功能 */ static struct file_operations socket_file_ops = { sock_lseek, sock_read, sock_write, sock_readdir, sock_select, sock_ioctl, NULL, /* mmap */ NULL, /* no special open code... */ sock_close, NULL, /* no fsync */ sock_fasync }; /* * The protocol list. Each protocol is registered in here. */ /* proto_ops是一个结构,该结构中声明了一系列操作函数,pops数据将在sock_register函数中 被初始化,对于不同的操作域,具有不同的操作函数集合,比方相应于INET域的操作函数集合是 inet_proto_ops操作函数集合。

    */ static struct proto_ops *pops[NPROTO]; /* * Statistics counters of the socket lists * 非常easy理解,就是系统正在使用的套接字数量 */ static int sockets_in_use = 0; /* * Support routines. Move socket addresses back and forth across the kernel/user * divide and look after the messy bits. */ #define MAX_SOCK_ADDR 128 /* 108 for Unix domain - 16 for IP, 16 for IPX, about 80 for AX.25 */ /* 【hujian】 以下两个函数将实现内核缓冲区与用户缓冲区之间的数据复制 */ static int move_addr_to_kernel(void *uaddr, int ulen, void *kaddr) { int err; if(ulen<0||ulen>MAX_SOCK_ADDR) return -EINVAL;//地址错误 if(ulen==0) return 0;//不须要赋值数据 if((err=verify_area(VERIFY_READ,uaddr,ulen))<0)//检查该地址区间? return err; /* 调用下层函数来实现数据复制。从用户缓冲区中的uaddr開始赋值ulen长度的 数据到内核缓冲区中的kaddr開始的地方。仅仅须要调用下层函数就能够了 */ memcpy_fromfs(kaddr,uaddr,ulen); return 0; } static int move_addr_to_user(void *kaddr, int klen, void *uaddr, int *ulen) { int err; int len; /*检查地址区间*/ if((err=verify_area(VERIFY_WRITE,ulen,sizeof(*ulen)))<0) return err; len=get_fs_long(ulen); if(len>klen) len=klen; if(len<0 || len> MAX_SOCK_ADDR) return -EINVAL; if(len) { if((err=verify_area(VERIFY_WRITE,uaddr,len))<0) return err; /* memcpy to file system,也就是将内核空间里面的数据拷贝到文件系统上面的某个 地方去,也就是我们想要实现的功能。就是将内核空间中的某段数据拷贝到用户缓冲区中 就能够了。 */ memcpy_tofs(uaddr,kaddr,len); } put_fs_long(len,ulen); return 0; } /* * Obtains the first available file descriptor and sets it up for use. */ /* 【hujian】 获得文件描写叙述符,分配file数据结构(get filedes) 为网络套接字分配一个文件描写叙述字。在socket系统调用sys_socket实现中,内核在分配了inode,socket。sock之后, 会调用该函数来获得一个文件描写叙述符,将这个获得的文件描写叙述符作为socket系统调用的返回值。 函数须要一个inode结构,由于分配文件描写叙述符的同一时候须要一个file结构。file结构中的f_inode字段即指向 这个inode结构。每一个file结构都须要一个inode结构来相应 */ static int get_fd(struct inode *inode) { int fd; struct file *file; /* * Find a file descriptor suitable for return to the user. */ /* get_empty_filp()函数将返回一个空的file结构。内核会在内核里面维护一个files数组 这个数组里面将保存全部file。这个函数将去遍历该数组。找到一个可用的数组元素,然后 变为空然后返回给调用者。 */ file = get_empty_filp(); if (!file) return(-1);//file结构分配失败,非常可能是由于files数组已经没有空的了 /* 这里我们能够看到,每次分配文件描写叙述子都是从最小的可用0文件描写叙述符開始搜索的, 所以该函数分配的文件描写叙述符是系统中当前可用的最小的文件描写叙述符号。

    */ for (fd = 0; fd < NR_OPEN; ++fd) if (!current->files->fd[fd]) break; /* 已经没有办法再分配文件描写叙述符号了,由于已经满了 */ if (fd == NR_OPEN) { file->f_count = 0; return(-1); } FD_CLR(fd, &current->files->close_on_exec); current->files->fd[fd] = file; /* 初始化,所以每当我们获得一个文件描写叙述字时,我们能够对该文件描写叙述符做 read、write等操作,这都是我们在这里进行初始化的结果 */ file->f_op = &socket_file_ops; file->f_mode = 3; /* 设置该文件能够读写,当然这个标志能够由用户改动 */ file->f_flags = O_RDWR; /* 设置这个文件的訪问次数? */ file->f_count = 1; /* 我们应该知道,每一个文件都和一个inode相相应,该inode就是我们须要给本函数的inode 我们将会把用户提供的inode赋值给我们新建的file结构,这样这个file结构就和一个inode 唯一关联了,我们就能够通过控制该file来訪问存于文件系统里面的inode节点了。 */ file->f_inode = inode; /* inode里面也有一个引用次数的计数器,我们须要加1。不然这个file还是不存在的 内核会在某些检查中将该file分配给新请求的进程的。由于仅仅要这个inode的引用次数 为0。内核就认为没有人再用这个inode了。那么能够回收资源了。 */ if (inode) inode->i_count++; file->f_pos = 0; return(fd); } /* * Go from an inode to its socket slot. * * The original socket implementation wasn't very clever, which is * why this exists at all.. * socket结构查询 */ inline struct socket *socki_lookup(struct inode *inode) { return &inode->u.socket_i; } /* * Go from a file number to its socket slot. */ /* 该函数实现的是依据文件描写叙述子来获得该文件描写叙述子的file信息,然后再file信息里面得到我们须要的 inode信息。然后,我们就能够调用我们在上面实现的socki_lookup将我们须要的socket返回了 */ static inline struct socket *sockfd_lookup(int fd, struct file **pfile) { struct file *file; struct inode *inode; /* 检查文件描写叙述字是否合法。而且获得当前进程的文件的文件描写叙述子为fd的文件(file) 我们能够依据file来获得文件的inode,然后我们能够依据inode来获得该inodedui 应得socket。然后将得到的socket返回,这个函数的功能就完毕了 */ if (fd < 0 || fd >= NR_OPEN || !(file = current->files->fd[fd])) return NULL; /* 获得当前文件描写叙述子的inode信息 */ inode = file->f_inode; /* 假设这个inode不合法,或者该inode指向的socket不存在,则返回null */ if (!inode || !inode->i_sock) return NULL; /* 当然,然后用户提供的參数中的file不为空的话,那么就须要将该file指向用户的指针 当用户视图从这个指针里面获取内容的时候,就能够得到结构 */ if (pfile) *pfile = file; /* 可是我们须要的是我们所给定的文件描写叙述字相应的socket。而不是file。所以我们还须要依据inode信息来得到 我们须要的socket,当然这个函数在上面我们已经看过。仅仅须要通过inode就能够得到socket */ return socki_lookup(inode); } /* * Allocate a socket. */ /* 分配一个新的socket结构,同一时候对结构进行一些初始化。 该函数实际上是先得到一个inode,然后对该inode进行一些初始化 */ struct socket *sock_alloc(void) { struct inode * inode; struct socket * sock; /* 得到一个空的inode,和get_empty_filp函数是一样的,内核将会遍历 内核中维护的一个数组inodes(不知道是不是叫这个名字),然后找到一个能够使用的inode 之后(这个inode的引用次数为0则表明这个inode不会再被使用,这个时候这个inode就是处在可 再次被分配的状态的)。将这个inode初始化。 */ inode = get_empty_inode(); if (!inode) return NULL;//已经没有很多其它能够使用的inode了。可能如今文件系统已经满了 /* 初始化这个新的inode,详细看inode的结果定义 /include/linux/fs.h/ */ inode->i_mode = S_IFSOCK; inode->i_sock = 1; /* 该inode的用户id和用户组id 这是inode的创建者,是为了实现文件保护而设置的标志,当然这些标志是能够被改动的 */ inode->i_uid = current->uid; inode->i_gid = current->gid; /* 获得sock,这个sock是新的。我们能够依据sock来找到file,然后能够找到inode 总之仅仅要我们有当中之中的一个,就能够找到我们须要的其它的内容(暂且这么想吧~) */ sock = &inode->u.socket_i; /* 当然的到这个sock之后,我们须要做一些初始化工作啊 */ sock->state = SS_UNCONNECTED; sock->flags = 0; sock->ops = NULL; sock->data = NULL; sock->conn = NULL; sock->iconn = NULL; sock->next = NULL; sock->wait = &inode->i_wait; sock->inode = inode; /* "backlink": we could use pointer arithmetic instead */ sock->fasync_list = NULL; sockets_in_use++; return sock; } /* * Release a socket. */ /* 释放而且回收一个socket */ static inline void sock_release_peer(struct socket *peer) { peer->state = SS_DISCONNECTING; wake_up_interruptible(peer->wait); sock_wake_async(peer, 1); } /* 继续的释放socket */ void sock_release(struct socket *sock) { int oldstate; struct socket *peersock, *nextsock; /* 在释放之前。须要先将该sock的状态设置为未连接 */ if ((oldstate = sock->state) != SS_UNCONNECTED) sock->state = SS_DISCONNECTING; /* * Wake up anyone waiting for connections. */ for (peersock = sock->iconn; peersock; peersock = nextsock) { nextsock = peersock->next; sock_release_peer(peersock); } /* * Wake up anyone we're connected to. First, we release the * protocol, to give it a chance to flush data, etc. */ peersock = (oldstate == SS_CONNECTED) ? sock->conn : NULL; if (sock->ops) /*调用下层函数来实现我们须要的功能。ops这个域仅仅是一个函数接口集合*/ sock->ops->release(sock, peersock); if (peersock)/*假设我们还在和某个sock连接着,则将这个sock也释放了*/ sock_release_peer(peersock); --sockets_in_use; /* Bookkeeping.. ,我们应该还记得,这个字段是一个socket’计数器,也就是 系统里面正在使用的socket的数量,当我们释放了一个socket的时候,当然须要将 该值降低1~~*/ iput(SOCK_INODE(sock));//实现对该inode结构的引用计数减1 } /* * Sockets are not seekable. */ static int sock_lseek(struct inode *inode, struct file *file, off_t offset, int whence) { return(-ESPIPE); } /* * Read data from a socket. ubuf is a user mode pointer. We make sure the user * area ubuf...ubuf+size-1 is writable before asking the protocol. */ /* 我们如今应该知道的事情是:套接字也是一种文件,所以我们能够获取到套接字的文件描写叙述子(套接字描写叙述子),然后 就像文件一样操作他们。比方我们想要读写网络数据。那么我们就能够调用read来读取。当然我们能够调用write来向 套接字文件写入内容(数据)。 在BSD层。这些函数的实现仅仅是向下引用,就是将这个任务交给下一层的网络层来实现。BSD下一层是INET层。 所以这一层的这些函数的接口都指向INET层的同样接口,假设inet层依旧没有办法实现这些接口,将再次向下 引用,直到能够实现最主要的操作后。函数再向上实现。

    */ static int sock_read(struct inode *inode, struct file *file, char *ubuf, int size) { struct socket *sock; int err;  //依据inode信息获得sock结构,假设这个socket是空的,直接返回错误,没有什么错误处理 if (!(sock = socki_lookup(inode))) { printk("NET: sock_read: can't find socket for inode! "); return(-EBADF); } //发现了这个socket,可是这不知道什么标志不同意。然后也是直接返回错误 if (sock->flags & SO_ACCEPTCON) return(-EINVAL); if(size<0) return -EINVAL; if(size==0) return 0; //我们能够想象,这个函数应该是检查这个区域。详细检查什么应该在函数的实现中看到。 if ((err=verify_area(VERIFY_WRITE,ubuf,size))<0) return err; //这个虚假的调用。仅仅是指向下一层的网络接口,在sock里面读出size大小的内容。存在ubuf中 return(sock->ops->read(sock, ubuf, size, (file->f_flags & O_NONBLOCK))); } /* * Write data to a socket. We verify that the user area ubuf..ubuf+size-1 is * readable by the user process. */ static int sock_write(struct inode *inode, struct file *file, char *ubuf, int size) { struct socket *sock; int err; if (!(sock = socki_lookup(inode))) { printk("NET: sock_write: can't find socket for inode! "); return(-EBADF); } if (sock->flags & SO_ACCEPTCON) return(-EINVAL); if(size<0) return -EINVAL; if(size==0) return 0; //还是检查内存,假设出错直接返回 if ((err=verify_area(VERIFY_READ,ubuf,size))<0) return err; //继续调用下一层的函数,往sock里面写入大小为size,内容为ubuf的数据,写入的标志的指定的标志 return(sock->ops->write(sock, ubuf, size,(file->f_flags & O_NONBLOCK))); } /* * You can't read directories from a socket! */ static int sock_readdir(struct inode *inode, struct file *file, struct dirent *dirent, int count) { return(-EBADF); } /* * With an ioctl arg may well be a user mode pointer, but we don't know what to do * with it - thats up to the protocol still. */ int sock_ioctl(struct inode *inode, struct file *file, unsigned int cmd, unsigned long arg) { struct socket *sock; if (!(sock = socki_lookup(inode))) { printk("NET: sock_ioctl: can't find socket for inode! "); return(-EBADF); } //继续调用下层函数,控制套接字的模式? return(sock->ops->ioctl(sock, cmd, arg)); } static int sock_select(struct inode *inode, struct file *file, int sel_type, select_table * wait) { struct socket *sock; if (!(sock = socki_lookup(inode))) { printk("NET: sock_select: can't find socket for inode! "); return(0); } /* * We can't return errors to select, so it's either yes or no. */ if (sock->ops && sock->ops->select) return(sock->ops->select(sock, sel_type, wait)); return(0); } /* 该函数将实现关闭一个socket 仅仅是简单的调用sock_release and sock_release_peer 他们的调用关系应该是这种: socket_close()->sock_release()->sock_release_peer() 所以这个函数仅仅是实现简单的函数调用。 */ void sock_close(struct inode *inode, struct file *filp) { struct socket *sock; /* * It's possible the inode is NULL if we're closing an unfinished socket. */ if (!inode) /*这个inode是个空的,难道是我们正在关闭一个没有完毕的socket?应该是啊*/ return; /* 我们已经看过这个函数了。那么这个函数仅仅是实现依据inode来找到sock 仅仅是将inode结构中的sock返回来而已 */ if (!(sock = socki_lookup(inode))) { printk("NET: sock_close: can't find socket for inode! "); return; } /* 这个函数用来分配或者释放一个fasync_struct 结构,假设第三个參数为0,那么说明是释放,否则表明 是分配,这个函数调用在这里将完毕将释放当前进程相应的该结构 可是为什么不是sock而是进程呢? 由于一个套接字非常可能被多个进程所使用 */ sock_fasync(inode, filp, 0); /* 释放该socket */ sock_release(sock); } /* * Update the socket async list */ static int sock_fasync(struct inode *inode, struct file *filp, int on) { struct fasync_struct *fa, *fna=NULL, **prev; struct socket *sock; unsigned long flags; if (on) { fna=(struct fasync_struct *)kmalloc(sizeof(struct fasync_struct), GFP_KERNEL); if(fna==NULL) return -ENOMEM; } sock = socki_lookup(inode); prev=&(sock->fasync_list); save_flags(flags); cli(); for(fa=*prev; fa!=NULL; prev=&fa->fa_next,fa=*prev) if(fa->fa_file==filp) break; if(on) { if(fa!=NULL) { kfree_s(fna,sizeof(struct fasync_struct)); restore_flags(flags); return 0; } fna->fa_file=filp; fna->magic=FASYNC_MAGIC; fna->fa_next=sock->fasync_list; sock->fasync_list=fna; } else { if(fa!=NULL) { *prev=fa->fa_next; kfree_s(fa,sizeof(struct fasync_struct)); } } restore_flags(flags); return 0; } /* 实现异步唤醒 */ int sock_wake_async(struct socket *sock, int how) { if (!sock || !sock->fasync_list) return -1; switch (how) { case 0: kill_fasync(sock->fasync_list, SIGIO); break; case 1: if (!(sock->flags & SO_WAITDATA)) kill_fasync(sock->fasync_list, SIGIO); break; case 2: if (sock->flags & SO_NOSPACE) { kill_fasync(sock->fasync_list, SIGIO); sock->flags &= ~SO_NOSPACE; } break; } return 0; } /* * Wait for a connection. * 该函数仅仅用于UNIX通信域,用于处理client连接请求 * 结构中的iconn,conn结构用于unix域中的连接操作。当中iconn仅仅用于服务器端,表示等待连接但尚未 * 连接的clientsocket结构链表 */ int sock_awaitconn(struct socket *mysock, struct socket *servsock, int flags) { struct socket *last; /* * We must be listening * 检查服务器端是不是处于服务状态,也就是服务端处于监听状态,能够进行连接 */ if (!(servsock->flags & SO_ACCEPTCON)) { return(-EINVAL); } /* * Put ourselves on the server's incomplete connection queue. */ mysock->next = NULL; cli(); //将本次client连接的套接字插入到服务端的socket结构中的iconn字段指向的链表 //这种话,就表示client正在等待连接 //get the servsock list's last item,if the pointer is null , if (!(last = servsock->iconn)) servsock->iconn = mysock; else { while (last->next) last = last->next; last->next = mysock; } mysock->state = SS_CONNECTING; mysock->conn = servsock;//事实上我们须要的就是这么一句。这样我们就能够告诉服务端。我这个client正在等待连接啊,快点连接我!

    。。 sti(); /* * Wake up server, then await connection. server will set state to * SS_CONNECTED if we're connected. */ //唤醒服务端进程,以来服务本地client连接 wake_up_interruptible(servsock->wait); //该函数将唤醒servsock结构上的fasync_list指向的队列中的每一个元素相应的进程 sock_wake_async(servsock, 0); //由于唤醒服务端须要一定的时间,所以一般我们能够进入到这个条件语句里面执行 if (mysock->state != SS_CONNECTED) { if (flags & O_NONBLOCK) return -EINPROGRESS; //等待服务端服务本次连接 interruptible_sleep_on(mysock->wait); //本地client被唤醒后,继续检查是不是已经得到处理了。就是和是否已经完毕了与服务端的连接 if (mysock->state != SS_CONNECTED && mysock->state != SS_DISCONNECTING) { /* * if we're not connected we could have been * 1) interrupted, so we need to remove ourselves * from the server list * 这是信号中断。这种话我们须要自己将这个套接字从服务端里面删除 * 2) rejected (mysock->conn == NULL), and have * already been removed from the list * 拒绝连接,所以失败,这种话,这个client已经被删除 */ if (mysock->conn == servsock) //这是说client的连接等于服务端了吗?然后连接却失败了!! { cli(); //假设第一个等待连接的套接字就是我们自己的套接字,那么删除。为了性能真是拼了!

    if ((last = servsock->iconn) == mysock) servsock->iconn = mysock->next; else //如今我们须要找到那个套接字(也就是我们自己的套接字) { while (last->next != mysock) last = last->next; last->next = mysock->next;//然后将这个套接字删除 } sti(); } //是中断还是拒绝连接呢? //假设mysock->conn为null的话。表示服务器拒绝连接。则返回-EACCESS //否则就是被中断标志,返回-EINIT //总之,假设是服务端拒绝服务的话,那么服务端就会删除这个拒绝的连接 //否则这个连接不是空的,那么我们须要自己删除这个套接字。

    return(mysock->conn ? -EINTR : -EACCES); } } return(0); } /* * Perform the socket system call. we locate the appropriate * family, then create a fresh socket. * 继续向下层调用接口实现 */ /* 该函数将实现以下的功能: (1)、通过函数sock_alloc来分配socket结构和sock结构,这两个结构在不同的层次表示一个套接字连接 (2)、分配inode和file结构用于普通文件操作 (3)、分配一个文件描写叙述子返回给应用程序作为以后的操作句柄 */ static int sock_socket(int family, int type, int protocol) { int i, fd; struct socket *sock; //套接字 struct proto_ops *ops;//协议选项 /* Locate the correct protocol family. */ for (i = 0; i < NPROTO; ++i) { if (pops[i] == NULL) continue; if (pops[i]->family == family) break;//找到所属于的协议族 } //就是当前系统不支持这种协议族 if (i == NPROTO) { return -EINVAL; } //得到协议 ops = pops[i]; /* * Check that this is a type that we know how to manipulate and * the protocol makes sense here. The family can still reject the * protocol later. */ if ((type != SOCK_STREAM && type != SOCK_DGRAM && type != SOCK_SEQPACKET && type != SOCK_RAW && type != SOCK_PACKET) || protocol < 0) return(-EINVAL); /* * Allocate the socket and allow the family to set things up. if * the protocol is 0, the family is instructed to select an appropriate * default. */ //sock_alloc函数将返回一个新的套接字,假设系统不支持很多其它的套接字了,那么就会失败 //分配socket和sock,这两个结构在不同的层次表示一个套接字连接 if (!(sock = sock_alloc())) { printk("NET: sock_socket: no more sockets "); return(-ENOSR); /* Was: EAGAIN, but we are out of system resources! */ } //我们已经得到一个新的套接字。如今就能够初始化这个套接字了 sock->type = type; sock->ops = ops; //向下层调用函数来实现 if ((i = sock->ops->create(sock, protocol)) < 0) { sock_release(sock); return(i); } //得到套接字描写叙述子,我们能够依据这个描写叙述子来实现各种相似文件的操作 if ((fd = get_fd(SOCK_INODE(sock))) < 0) { sock_release(sock); return(-EINVAL); } //将这个套接字描写叙述子返回去 return(fd); } /* * Create a pair of connected sockets. */ static int sock_socketpair(int family, int type, int protocol, unsigned long usockvec[2]) { int fd1, fd2, i; struct socket *sock1, *sock2; int er; /* * Obtain the first socket and check if the underlying protocol * supports the socketpair call. */ if ((fd1 = sock_socket(family, type, protocol)) < 0) return(fd1); sock1 = sockfd_lookup(fd1, NULL); if (!sock1->ops->socketpair) { sys_close(fd1); return(-EINVAL); } /* * Now grab another socket and try to connect the two together. */ if ((fd2 = sock_socket(family, type, protocol)) < 0) { sys_close(fd1); return(-EINVAL); } //sockfd_lookup将会查看这个套接字是否已经被创建,假设没有被创建的话,那么函数返回失败 //套接字也是一种文件,所以创建的套接字会有一个inode存在。 sock2 = sockfd_lookup(fd2, NULL); //下层函数来做。 if ((i = sock1->ops->socketpair(sock1, sock2)) < 0) { sys_close(fd1); sys_close(fd2); return(i); } //让两个新的套接字描写叙述字连接在一起,这两个套接字描写叙述字全双工的 sock1->conn = sock2; sock2->conn = sock1; sock1->state = SS_CONNECTED; sock2->state = SS_CONNECTED; er=verify_area(VERIFY_WRITE, usockvec, 2 * sizeof(int)); if(er) { sys_close(fd1); sys_close(fd2); return er; } put_fs_long(fd1, &usockvec[0]); put_fs_long(fd2, &usockvec[1]); return(0); } /* * Bind a name to a socket. Nothing much to do here since it's * the protocol's responsibility to handle the local address. * * We move the socket address to kernel space before we call * the protocol layer (having also checked the address is ok). */ //我们在创建一个套接字描写叙述子之后。这个套接字资源还仅仅能在进程内使用。其它进程无法识别这个新的 //套接字资源,所以须要给这个新的套接字绑定一个名字,这样其它进程就能够依据这个套接字名字来訪问 //这个套接字了。也就是能够和这个套接字通信了 static int sock_bind(int fd, struct sockaddr *umyaddr, int addrlen) { struct socket *sock; int i; char address[MAX_SOCK_ADDR]; int err; //assert if (fd < 0 || fd >= NR_OPEN || current->files->fd[fd] == NULL) return(-EBADF); //查看这个套接字是否已经存在,假设不存在的话,返回失败 //第二个參数将返回套接字结构的file结构,假设我们须要知道套接字的inode信息的话,这个參数不应该为空 //可是这里仅仅是为了測试是不是存在,所以不须要保存file结构 if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); //在使用这个名字之前 。将套接字的地址把move到内核空间 if((err=move_addr_to_kernel(umyaddr,addrlen,address))<0) return err; //下层函数继续做苦力 if ((i = sock->ops->bind(sock, (struct sockaddr *)address, addrlen)) < 0) { return(i); } return(0); } /* * Perform a listen. Basically, we allow the protocol to do anything * necessary for a listen, and if that works, we mark the socket as * ready for listening. */ //ok。对于tcp连接,该函数将创建监听队列,这个队列里面保存全部请求连接该套接字的套接字 //当我们调用accept函数的时候。accept函数将创建一个新的socket,然后取出这个监听队列中 //的一个套接字。用这个新的套接字来与之通信。须要注意的是,由于我们仅仅须要让这个新的套接字 //和特定的套接字通信,所以不须要绑定名字 //第二个參数是监听队列的长度 static int sock_listen(int fd, int backlog) { struct socket *sock; //assert if (fd < 0 || fd >= NR_OPEN || current->files->fd[fd] == NULL) return(-EBADF); //assert whether the socket already exist. if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); //assert the socket's state if (sock->state != SS_UNCONNECTED) { return(-EINVAL); } //继续调用下层函数来做事情 if (sock->ops && sock->ops->listen) sock->ops->listen(sock, backlog); sock->flags |= SO_ACCEPTCON; return(0); } /* * For accept, we attempt to create a new socket, set up the link * with the client, wake up the client, then return the new * connected fd. We collect the address of the connector in kernel * space and move it to user at the very end. This is buggy because * we open the socket then return an error. */ //这个函数视图新建一个新的套接字来为一个client的连接请求服务,这个套接字将作为结果返回给用户以便 //用户能够知道为他服务的套接字是哪一个。 //上面说了,函数会将连接的套接字的地址从内核空间move到用户空间。这样用户就能够控制这个套接字了。

    // static int sock_accept(int fd, struct sockaddr *upeer_sockaddr, int *upeer_addrlen) { struct file *file; struct socket *sock, *newsock; int i; char address[MAX_SOCK_ADDR]; int len; //asert if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); //get the fd's file structure. if (!(sock = sockfd_lookup(fd, &file))) return(-ENOTSOCK); //asert if (sock->state != SS_UNCONNECTED) { return(-EINVAL); } //assert if (!(sock->flags & SO_ACCEPTCON)) { return(-EINVAL); } //new socket if (!(newsock = sock_alloc())) { printk("NET: sock_accept: no more sockets "); return(-ENOSR); /* Was: EAGAIN, but we are out of system resources! */ } //this is the server socket. newsock->type = sock->type; newsock->ops = sock->ops; //copy the sock to newsock if ((i = sock->ops->dup(newsock, sock)) < 0) { sock_release(newsock); return(i); } //call down.. i = newsock->ops->accept(sock, newsock, file->f_flags); if ( i < 0) { sock_release(newsock); return(i); } if ((fd = get_fd(SOCK_INODE(newsock))) < 0) { sock_release(newsock); return(-EINVAL); } if (upeer_sockaddr) { newsock->ops->getname(newsock, (struct sockaddr *)address, &len, 1); move_addr_to_user(address,len, upeer_sockaddr, upeer_addrlen); } return(fd); } /* * Attempt to connect to a socket with the server address. The address * is in user space so we verify it is OK and move it to kernel space. */ //这是一个共client使用的函数,他用来连接一个服务端,服务端会在listen里面监听到这个链接请求 //然后将这个新的连接请求放在连接队列里面,然后服务器将用accept来为队列里面的套接字服务 static int sock_connect(int fd, struct sockaddr *uservaddr, int addrlen) { struct socket *sock; struct file *file; int i; char address[MAX_SOCK_ADDR]; int err; //asert if (fd < 0 || fd >= NR_OPEN || (file=current->files->fd[fd]) == NULL) return(-EBADF); //get the file structure if (!(sock = sockfd_lookup(fd, &file))) return(-ENOTSOCK); //move addr from user space to kernel so that the kernel can get the control in this socket if((err=move_addr_to_kernel(uservaddr,addrlen,address))<0) return err; //do something according to the socket's state switch(sock->state) { case SS_UNCONNECTED: /* This is ok... continue with connect */ break; case SS_CONNECTED: /* Socket is already connected */ if(sock->type == SOCK_DGRAM) /* Hack for now - move this all into the protocol */ break; return -EISCONN; case SS_CONNECTING: /* Not yet connected... we will check this. */ /* * FIXME: for all protocols what happens if you start * an async connect fork and both children connect. Clean * this up in the protocols! */ break; default: return(-EINVAL); } //next level i = sock->ops->connect(sock, (struct sockaddr *)address, addrlen, file->f_flags); if (i < 0) { return(i); } return(0); } /* * Get the local address ('name') of a socket object. Move the obtained * name to user space. */ //函数将得到自己的套接字的名字 // static int sock_getsockname(int fd, struct sockaddr *usockaddr, int *usockaddr_len) { struct socket *sock; char address[MAX_SOCK_ADDR]; int len; int err; if (fd < 0 || fd >= NR_OPEN || current->files->fd[fd] == NULL) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); err=sock->ops->getname(sock, (struct sockaddr *)address, &len, 0); if(err) return err; if((err=move_addr_to_user(address,len, usockaddr, usockaddr_len))<0) return err; return 0; } /* * Get the remote address ('name') of a socket object. Move the obtained * name to user space. */ //这个函数将会得到对方的套接字地址 static int sock_getpeername(int fd, struct sockaddr *usockaddr, int *usockaddr_len) { struct socket *sock; char address[MAX_SOCK_ADDR]; int len; int err; if (fd < 0 || fd >= NR_OPEN || current->files->fd[fd] == NULL) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); err=sock->ops->getname(sock, (struct sockaddr *)address, &len, 1); if(err) return err; if((err=move_addr_to_user(address,len, usockaddr, usockaddr_len))<0) return err; return 0; } /* * Send a datagram down a socket. The datagram as with write() is * in user space. We check it can be read. */ //向一个套接字写数据 static int sock_send(int fd, void * buff, int len, unsigned flags) { struct socket *sock; struct file *file; int err; //check if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); //check if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); //nothing to do if(len<0) return -EINVAL; //验证这个地址,检查是否可读 err=verify_area(VERIFY_READ, buff, len); if(err) return err; //下层调用 return(sock->ops->send(sock, buff, len, (file->f_flags & O_NONBLOCK), flags)); } /* * Send a datagram to a given address. We move the address into kernel * space and check the user space data area is readable before invoking * the protocol. */ //这个函数是对于udp来说的,向一个套接字发送数据。可是没有那么麻烦 static int sock_sendto(int fd, void * buff, int len, unsigned flags, struct sockaddr *addr, int addr_len) { struct socket *sock; struct file *file; char address[MAX_SOCK_ADDR]; int err; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); if(len<0) return -EINVAL; err=verify_area(VERIFY_READ,buff,len); if(err) return err; if((err=move_addr_to_kernel(addr,addr_len,address))<0) return err; return(sock->ops->sendto(sock, buff, len, (file->f_flags & O_NONBLOCK), flags, (struct sockaddr *)address, addr_len)); } /* * Receive a datagram from a socket. This isn't really right. The BSD manual * pages explicitly state that recv is recvfrom with a NULL to argument. The * Linux stack gets the right results for the wrong reason and this need to * be tidied in the inet layer and removed from here. * We check the buffer is writable and valid. */ //接收数据,对于tcp static int sock_recv(int fd, void * buff, int len, unsigned flags) { struct socket *sock; struct file *file; int err; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); if(len<0) return -EINVAL; if(len==0) return 0; //验证这个地址是不是可写的 err=verify_area(VERIFY_WRITE, buff, len); if(err) return err; //下层函数 return(sock->ops->recv(sock, buff, len,(file->f_flags & O_NONBLOCK), flags)); } /* * Receive a frame from the socket and optionally record the address of the * sender. We verify the buffers are writable and if needed move the * sender address from kernel to user space. */ //对于udp来说的 static int sock_recvfrom(int fd, void * buff, int len, unsigned flags, struct sockaddr *addr, int *addr_len) { struct socket *sock; struct file *file; char address[MAX_SOCK_ADDR]; int err; int alen; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); if(len<0) return -EINVAL; if(len==0) return 0; err=verify_area(VERIFY_WRITE,buff,len); if(err) return err; len=sock->ops->recvfrom(sock, buff, len, (file->f_flags & O_NONBLOCK), flags, (struct sockaddr *)address, &alen); if(len<0) return len; if(addr!=NULL && (err=move_addr_to_user(address,alen, addr, addr_len))<0) return err; return len; } /* * Set a socket option. Because we don't know the option lengths we have * to pass the user mode parameter for the protocols to sort out. */ static int sock_setsockopt(int fd, int level, int optname, char *optval, int optlen) { struct socket *sock; struct file *file; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); //下层函数 return(sock->ops->setsockopt(sock, level, optname, optval, optlen)); } /* * Get a socket option. Because we don't know the option lengths we have * to pass a user mode parameter for the protocols to sort out. */ static int sock_getsockopt(int fd, int level, int optname, char *optval, int *optlen) { struct socket *sock; struct file *file; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); if (!sock->ops || !sock->ops->getsockopt) return(0); return(sock->ops->getsockopt(sock, level, optname, optval, optlen)); } /* * Shutdown a socket. */ //关闭一个套接字,关闭这个套接字的话,下次连接须要又一次连接 //how指定怎样关闭,能够关闭该套接字不再可写,或者不再可读,或者不再可读写 static int sock_shutdown(int fd, int how) { struct socket *sock; struct file *file; if (fd < 0 || fd >= NR_OPEN || ((file = current->files->fd[fd]) == NULL)) return(-EBADF); if (!(sock = sockfd_lookup(fd, NULL))) return(-ENOTSOCK); //仅仅能向下看了 return(sock->ops->shutdown(sock, how)); } /* * Perform a file control on a socket file descriptor. */ int sock_fcntl(struct file *filp, unsigned int cmd, unsigned long arg) { struct socket *sock; sock = socki_lookup (filp->f_inode); if (sock != NULL && sock->ops != NULL && sock->ops->fcntl != NULL) return(sock->ops->fcntl(sock, cmd, arg)); return(-EINVAL); } /* * System call vectors. Since I (RIB) want to rewrite sockets as streams, * we have this level of indirection. Not a lot of overhead, since more of * the work is done via read/write/select directly. * * I'm now expanding this up to a higher level to separate the assorted * kernel/user space manipulations and global assumptions from the protocol * layers proper - AC. */ asmlinkage int sys_socketcall(int call, unsigned long *args) { int er; switch(call) { case SYS_SOCKET: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_socket(get_fs_long(args+0), get_fs_long(args+1), get_fs_long(args+2))); case SYS_BIND: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_bind(get_fs_long(args+0), (struct sockaddr *)get_fs_long(args+1), get_fs_long(args+2))); case SYS_CONNECT: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_connect(get_fs_long(args+0), (struct sockaddr *)get_fs_long(args+1), get_fs_long(args+2))); case SYS_LISTEN: er=verify_area(VERIFY_READ, args, 2 * sizeof(long)); if(er) return er; return(sock_listen(get_fs_long(args+0), get_fs_long(args+1))); case SYS_ACCEPT: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_accept(get_fs_long(args+0), (struct sockaddr *)get_fs_long(args+1), (int *)get_fs_long(args+2))); case SYS_GETSOCKNAME: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_getsockname(get_fs_long(args+0), (struct sockaddr *)get_fs_long(args+1), (int *)get_fs_long(args+2))); case SYS_GETPEERNAME: er=verify_area(VERIFY_READ, args, 3 * sizeof(long)); if(er) return er; return(sock_getpeername(get_fs_long(args+0), (struct sockaddr *)get_fs_long(args+1), (int *)get_fs_long(args+2))); case SYS_SOCKETPAIR: er=verify_area(VERIFY_READ, args, 4 * sizeof(long)); if(er) return er; return(sock_socketpair(get_fs_long(args+0), get_fs_long(args+1), get_fs_long(args+2), (unsigned long *)get_fs_long(args+3))); case SYS_SEND: er=verify_area(VERIFY_READ, args, 4 * sizeof(unsigned long)); if(er) return er; return(sock_send(get_fs_long(args+0), (void *)get_fs_long(args+1), get_fs_long(args+2), get_fs_long(args+3))); case SYS_SENDTO: er=verify_area(VERIFY_READ, args, 6 * sizeof(unsigned long)); if(er) return er; return(sock_sendto(get_fs_long(args+0), (void *)get_fs_long(args+1), get_fs_long(args+2), get_fs_long(args+3), (struct sockaddr *)get_fs_long(args+4), get_fs_long(args+5))); case SYS_RECV: er=verify_area(VERIFY_READ, args, 4 * sizeof(unsigned long)); if(er) return er; return(sock_recv(get_fs_long(args+0), (void *)get_fs_long(args+1), get_fs_long(args+2), get_fs_long(args+3))); case SYS_RECVFROM: er=verify_area(VERIFY_READ, args, 6 * sizeof(unsigned long)); if(er) return er; return(sock_recvfrom(get_fs_long(args+0), (void *)get_fs_long(args+1), get_fs_long(args+2), get_fs_long(args+3), (struct sockaddr *)get_fs_long(args+4), (int *)get_fs_long(args+5))); case SYS_SHUTDOWN: er=verify_area(VERIFY_READ, args, 2* sizeof(unsigned long)); if(er) return er; return(sock_shutdown(get_fs_long(args+0), get_fs_long(args+1))); case SYS_SETSOCKOPT: er=verify_area(VERIFY_READ, args, 5*sizeof(unsigned long)); if(er) return er; return(sock_setsockopt(get_fs_long(args+0), get_fs_long(args+1), get_fs_long(args+2), (char *)get_fs_long(args+3), get_fs_long(args+4))); case SYS_GETSOCKOPT: er=verify_area(VERIFY_READ, args, 5*sizeof(unsigned long)); if(er) return er; return(sock_getsockopt(get_fs_long(args+0), get_fs_long(args+1), get_fs_long(args+2), (char *)get_fs_long(args+3), (int *)get_fs_long(args+4))); default: return(-EINVAL); } } /* * This function is called by a protocol handler that wants to * advertise its address family, and have it linked into the * SOCKET module. */ int sock_register(int family, struct proto_ops *ops) { int i; cli(); for(i = 0; i < NPROTO; i++) { if (pops[i] != NULL) continue; pops[i] = ops; pops[i]->family = family; sti(); return(i); } sti(); return(-ENOMEM); } /* * This function is called by a protocol handler that wants to * remove its address family, and have it unlinked from the * SOCKET module. */ int sock_unregister(int family) { int i; cli(); for(i = 0; i < NPROTO; i++) { if (pops[i] == NULL) continue; if(pops[i]->family == family) { pops[i]=NULL; sti(); return(i); } } sti(); return(-ENOENT); } void proto_init(void) { extern struct net_proto protocols[]; /* Network protocols */ struct net_proto *pro; /* Kick all configured protocols. */ pro = protocols; while (pro->name != NULL) { (*pro->init_func)(pro); pro++; } /* We're all done... */ } void sock_init(void) { int i; printk("Swansea University Computer Society NET3.019 "); /* * Initialize all address (protocol) families. */ for (i = 0; i < NPROTO; ++i) pops[i] = NULL; /* * Initialize the protocols module. */ proto_init(); #ifdef CONFIG_NET /* * Initialize the DEV module. */ dev_init(); /* * And the bottom half handler */ bh_base[NET_BH].routine= net_bh; enable_bh(NET_BH); #endif } int socket_get_info(char *buffer, char **start, off_t offset, int length) { int len = sprintf(buffer, "sockets: used %d ", sockets_in_use); if (offset >= len) { *start = buffer; return 0; } *start = buffer + offset; len -= offset; if (len > length) len = length; return len; }

  • 相关阅读:
    js闭包
    mysql 创建索引
    模拟title实现提示功能
    数组的深拷贝和浅拷贝
    JavaScript总结之DOM基本操作(三)
    JavaScript总结之数组操作(二)
    JavaScript总结之字符串操作(一)
    Visual Studio Code中JavaScript开发环境的配置
    PhpStorm+PhpStudy开发环境的配置
    如何在个人网站上添加logo图标
  • 原文地址:https://www.cnblogs.com/zhchoutai/p/8452975.html
Copyright © 2011-2022 走看看