您的位置:首页 > 运维架构 > Linux

linux下C语言中的flock函数用法 .

2013-01-08 13:25 288 查看
该结构是在lock.h文件中定义。


lock.h File


功能

定义一些文件的锁的选项


Description

The flock structure in the /usr/include/sys/flock.h file, which describes a lock, contains the following fields:

l_typeDescribes the type of lock. If the value of the Command parameter to the fcntlsubroutine is F_SETLK orF_SETLKW, the l_type field
indicates the type of lock to be created. Possible values are:

F_RDLCKA read lock is requested.F_WRLCKA write lock is requested.F_UNLCKUnlock. An existing lock is to be removed.

If the value of the Command parameter to the fcntl subroutine is F_GETLK, the l_typefield describes an existing
lock. Possible values are:

F_RDLCKA conflicting read lock exists.F_WRLCKA conflicting write lock exists.F_UNLCKNo conflicting lock exists.
l_whenceDefines the starting offset. The value of this field indicates the point from which the relative offset, the l_startfield, is measured. Possible values are:

SEEK_SETThe relative offset is measured from the start of the file.SEEK_CURThe relative offset is measured from the current position.SEEK_ENDThe relative offset is measured from the end of the file.

These values are defined in the unistd.h file.
l_startDefines the relative offset in bytes, measured from the starting point in the l_whencefield.
l_lenSpecifies the number of consecutive bytes to be locked.
l_sysidContains the ID of the node that already has a lock placed on the area defined by thefcntl subroutine. This field is returned only when the value of the Command parameter isF_GETLK.
l_pidContains the ID of a process that already has a lock placed on the area defined by thefcntl subroutine. This field is returned only when the value of the Command parameter isF_GETLK.
l_vfs

Specifies the file system type of the node identified in the l_sysid field.

看一下示例吧!

[cpp] view
plaincopy

int

waldirlock(Wal *w)

{

int r;

int fd;

struct flock lk;

char path[PATH_MAX];

r = snprintf(path, PATH_MAX, "%s/lock", w->dir);

if (r > PATH_MAX) {

twarnx("path too long: %s/lock", w->dir);

return 0;

}

fd = open(path, O_WRONLY|O_CREAT, 0600);

if (fd == -1) {

twarn("open");

return 0;

}

lk.l_type = F_WRLCK;

lk.l_whence = SEEK_SET;

lk.l_start = 0;

lk.l_len = 0;

r = fcntl(fd, F_SETLK, &lk);

if (r) {

twarn("fcntl");

return 0;

}

// intentionally leak fd, since we never want to close it

// and we'll never need it again

return 1;

}

struct flock 作为fcntl函数的第三个参数,使用F_SETLK,设置了其参数。

flock() 的函数原型如下所示:

int flock(int fd, int operation);

其中,参数 fd 表示文件描述符;参数 operation 指定要进行的锁操作,该参数的取值有如下几种:

LOCK_SH:表示要创建一个共享锁,在任意时间内,一个文件的共享锁可以被多个进程拥有;

LOCK_EX:表示创建一个排他锁,在任意时间内,一个文件的排他锁只能被一个进程拥有;

LOCK_UN:表示删除该进程创建的锁;

LOCK_MAND:它主要是用于共享模式强制锁,它可以与 LOCK_READ 或者 LOCK_WRITE 联合起来使用,从而表示是否允许并发的读操作或者并发的写操作;

通常情况下,如果加锁请求不能被立即满足,那么系统调用 flock() 会阻塞当前进程。比如,进程想要请求一个排他锁,但此时,已经由其他进程获取了这个锁,那么该进程将会被阻塞。如果想要在没有获得这个排他锁的情况下不阻塞该进程,可以将 LOCK_NB 和 LOCK_SH 或者 LOCK_EX 联合使用,那么系统就不会阻塞该进程。flock() 所加的锁会对整个文件起作用。

注意:

1. 对于文件的 close() 操作会使文件锁失效;

2. 同理,进程结束后文件锁失效;

3. flock() 的 LOCK_EX 是“劝告锁”,系统内核不会强制检查锁的状态,需要在代码中进行文件操作的地方显式检查才能生效。

2012-07-31 09:50 1030人阅读 评论(0) 收藏 举报

  表头文件 #include<sys/file.h>

  定义函数 int flock(int fd,int operation);

  函数说明 flock()会依参数operation所指定的方式对参数fd所指的文件做各种锁定或解除锁定的动作。此函数只能锁定整个文件,无法锁定文件的某一区域。

  参数 operation有下列四种情况:

  LOCK_SH 建立共享锁定。多个进程可同时对同一个文件作共享锁定。

  LOCK_EX 建立互斥锁定。一个文件同时只有一个互斥锁定。

  LOCK_UN 解除文件锁定状态。

  LOCK_NB 无法建立锁定时,此操作可不被阻断,马上返回进程。通常与LOCK_SH或LOCK_EX 做OR(|)组合。

  单一文件无法同时建立共享锁定和互斥锁定,而当使用dup()或fork()时文件描述词不会继承此种锁定。

  返回值 返回0表示成功,若有错误则返回-1,错误代码存于errno。

flock只要在打开文件后,需要对文件读写之前flock一下就可以了,用完之后再flock一下,前面加锁,后面解锁。其实确实是这么简单,但是前段时间用的时候发现点问题,问题描述如下:

一个进程去打开文件,输入一个整数,然后上一把写锁(LOCK_EX),再输入一个整数将解锁(LOCK_UN),另一个进程打开同样一个文件,直接向文件中写数据,发现锁不起作用,能正常写入(我此时用的是超级用户)。google了一大圈发现flock不提供锁检查,也就是说在用flock之前需要用户自己去检查一下是否已经上了锁,说明白点就是读写文件之前用一下flock检查一下文件有没有上锁,如果上锁了flock将会阻塞在那里(An attempt to
lock the file using one of these file descriptors may be denied by a lock that the calling process has already placed via another descriptor ),除非用了LOCK_NB。一个完整的用于测试的事例代码如下所示:

//lockfile.c

#include <stdio.h>

#include <unistd.h>

#include <sys/types.h>

#include <sys/stat.h>

#include <fcntl.h>

#include <errno.h>

int main()

{

int fd,i;

char path[]="/home/taoyong/test.txt";

extern int errno;

fd=open(path,O_WRONLY|O_CREAT);

if(fd!=-1)

{

printf("open file %s ./n",path);

printf("please input a number to lock the file./n");

scanf("%d",&i);

if(flock(fd,LOCK_EX)==0)

{

printf("the file was locked./n");

}

else

{

printf("the file was not locked./n");

}

printf("please input a number to unlock the file./n");

scanf("%d",&i);

if(flock(fd,LOCK_UN)==0)

{

printf("the file was unlocked./n");

}

else

{

printf("the file was not unlocked./n");

}

close(fd);

}

else

{

printf("cannot open file %s/n",path);

printf("errno:%d/n",errno);

printf("errMsg:%s",strerror(errno));

}

}

//testprocess.c

#include <stdio.h>

#include <unistd.h>

#include <sys/types.h>

#include <sys/stat.h>

#include <fcntl.h>

#include <errno.h>

#include <sys/file.h>

int main()

{

int fd,i;

char path[]="/home/taoyong/test.txt";

char s[]="writing.../nwriting....../n";

extern int errno;

fd=open(path,O_WRONLY|O_CREAT|O_APPEND);

if(fd!=-1)

{

printf("open file %s ./n",path);

if(flock(fd,LOCK_EX|LOCK_NB)==0)

{

printf("the file was locked by the process./n");

if(-1!=write(fd,s,sizeof(s)))

{

printf("write %s to the file %s/n",s,path);

}

else

{

printf("cannot write the file %s/n",path);

printf("errno:%d/n",errno);

printf("errMsg:%s/n",strerror(errno));

}

}

else

{

printf("the file was locked by other process.Can't write.../n");

printf("errno:%d:",errno);

}

close(fd);

}

else

{

printf("cannot open file %s/n",path);

printf("errno:%d/n",errno);

printf("errMsg:%s",strerror(errno));

}

}
内容来自用户分享和网络整理,不保证内容的准确性,如有侵权内容,可联系管理员处理 点击这里给我发消息
标签: