1 Introduction
The last chapter describes the basic functions to perform IO operations are carried out around the normal file IO basic question - open a file or write a file, this chapter will observe nature and other characteristics of the file system.
2, stat, fatat and functions lstat
This chapter discusses the following three functions and information is their return:
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
int stat(const char *pathname, struct stat *buf);
int fstat(int fd, struct stat *buf);
int lstat(const char *pathname, struct stat *buf);
Given a pathname, stat function returns a message related to the structure of the file with this name. fatat function to obtain information about the file is already open in the descriptor fd, lstat function is similar to stat, but the file name is a symbolic link, lstat returns information about the symbolic link. The second argument is a pointer to a structure, the following meanings for the stat field:
struct stat {
dev_t st_dev; /* ID of device containing file */
ino_t st_ino; /* inode 号 */
mode_t st_mode; /* 权限和文件类型，位图，权限位9位，类型3位，u+s 1位，g+s 1位，粘滞位(T位)1位。
位图是用一位或几位数据表示某种状态。许多要解决看似不可能的问题的面试题往往需要从位图着手。*/
nlink_t st_nlink; /* 硬链接数量 */
uid_t st_uid; /* 文件属主 ID */
gid_t st_gid; /* 文件属组 ID */
dev_t st_rdev; /* 设备号，只有设备文件才有 */
off_t st_size; /* 总大小字节数，编译时需要指定宏 -D_FILE_OFFSIZE_BITES=64，否则读取大文件可能导致溢出 */
blksize_t st_blksize; /* 文件系统块大小 */
blkcnt_t st_blocks; /* 每个 block 占用 512B，则整个文件占用的 block 数量。这个值是文件真正意义上所占用的磁盘空间 */
// 下面三个成员都是大整数，实际使用时需要先转换
time_t st_atime; /* 文件最后访问时间戳 */
time_t st_mtime; /* 文件最后修改时间戳 */
time_t st_ctime; /* 文件亚数据最后修改时间戳 */
}
Use stat function is probably the most ls -l command. Display all information about a file.
3, file type
When contacted over before learning linux file type, here a brief introduction to. In addition to regular files and directories, as well as some of the file types:
(1) ordinary files, common file types, such files contain some form of data, whether it is text data or binary data.
(2) directory file, this file contains the names of other files relating to information and a pointer pointing to these files.
(3) character special file system for a certain type of equipment.
(4) block special file, a disk device.
(5) FIFO, for inter-process communication, also called piping.
(6) sockets for inter-process communication networks.
(7) a symbolic link, this file that points to another file.
The system provides a number of macro action to determine a specific file type: only need to use the stat structure in the mode to pass in, easy to use in an if statement.
S_ISREG(m) is it a regular file? 是否为普通文件
S_ISDIR(m) directory? 是否为目录
S_ISCHR(m) character device? 是否为字符设备文件
S_ISBLK(m) block device? 是否为块设备文件
S_ISFIFO(m) FIFO (named pipe)? 是否为管道文件
S_ISLNK(m) symbolic link? (Not in POSIX.1-1996.) 是否为符号链接文件
S_ISSOCK(m) socket? (Not in POSIX.1-1996.) 是否为套接字文件4, setting the user ID and group ID setting
And a process ID associated with six or more are as follows:
Set user ID and group ID bits contain bits st_mode values, which can be tested using two S_ISUID and SISGID.
5, file access permissions
st_mode also contains access permissions for the file, the file type mentioned above which has access permissions. Each file access permissions nine bits, as in a linux file permissions into three groups, file owner permissions, group permissions file owner, other user rights, and three of each group have rights: namely, read (r) writing (w) and execute (x) can be obtained when using the file permission flags ls -l. The system also provides a macro to help us obtain the appropriate permission bits from the st_mode.
S_IRWXU 00700 mask for file owner permissions 当 st_mode & 这个宏时，计算结果将只保留 st_mode 中的文件属主权限，其它位全部被清零。注意只是将计算结果中其它位清零，并不是把 st_mode 本身的数据清零。
S_IRUSR 00400 owner has read permission
S_IWUSR 00200 owner has write permission
S_IXUSR 00100 owner has execute permission
S_IRWXG 00070 mask for group permissions
S_IRGRP 00040 group has read permission
S_IWGRP 00020 group has write permission
S_IXGRP 00010 group has execute permission
S_IRWXO 00007 mask for permissions for others (not in group)
S_IROTH 00004 others have read permission
S_IWOTH 00002 others have write permission
S_IXOTH 00001 others have execute permission6 ownership, new files and directories
User ID set up a new file for the effective user ID of the process, about the group ID, allows you to select one of the following criteria as the group ID of the new file:
(1) group ID of the new file can be an effective group ID of the process.
Group ID (2) of the new file can be set in its directory ID.
7, access function
#include <unistd.h>
int access(const char *pathname, int mode);Is there a test for the current process pathname file mode permissions successful return 0, else return 1, and set the errno. If pathname is a symbolic link to the test but does not expand the symbolic link file itself. mode can be selected as follows:
mode Description
R _ OK test read permission
W _ OK test write permissions
X _ OK test execution permission
if F _ OK test file exists
8, umask function
#include <sys/types.h>
#include <sys/stat.h>
mode_t umask(mode_t mask);Masking process for setting file mode, the previous value and return, the greater the value of umask, the lower authority. umask (1) command is to use the function package. Parameters mask bits constituted by the following, or may be used with bits specify multiple operation modes.
| st_mode shield | meaning |
| S_IRUSR S_IWUSR S_IXUSR |
Reading owner Write owner Owner execution |
| S_IRGRP S_IWGRP S_IXGRP |
It is a group reading It is a group write It is a set of execution |
| S_IROTH S_IWOTH S_IXOTH |
Other Reading Other Write Other executive |
9, chmod function and fchmode
#include <sys/stat.h>
int chmod(const char *path, mode_t mode);
int fchmod(int fd, mode_t mode);Chmod function to change the existing family of functions for file access permissions, chmod operation on the specified file, fchmod operate on open file. In order to change a file's permission bits, effective user ID of the process must be equal to the file's owner, or the process must have superuser permissions. Constant parameter mode as bitwise OR operation. chmod chmod command is to use the function package.
10-bit stick
In earlier versions of unix, there is a place called stick position, if an executable file this bit is set, then the first execution of the program and the end, a body of text stored in the exchange program area. This makes it faster next time into memory area when executing the program. However, now use unix page cache technique, frequently used data can reside in memory, so the effect of the stuck bit weaker and weaker.
11, chown, fchown, lchown function
chown to change the file's user ID and group ID.
#include <unistd.h>
int chown(const char *path, uid_t owner, gid_t group);
int fchown(int fd, uid_t owner, gid_t group);
int lchown(const char *path, uid_t owner, gid_t group);chown family of functions used to modify the file owner, modifying successful return 0, else return -1 and set errno. Modify the file owner can only be superuser to do. Three functions similar operations, except for the object is not the same. Prevent users from changing the file owner so please limit disk space to it.
12, file length
st_size stat structure member comprising a length of documents in bytes, this field is meaningful only ordinary files, directories, files and symbolic links. For ordinary file, the file length may be 0, when reading this document, the resulting file indication. For the directory file, the file size is typically a number, for the symbolic link, file size is the actual number of bytes in a file name.
13, truncated file
Sometimes the need to cut off some of the data in the file to shorten the end of the file, you need to truncate function and ftruncate.
#include <unistd.h>
#include <sys/types.h>
int truncate(const char *path, off_t length);
int ftruncate(int fd, off_t length);truncate函数族文件用于截断path所指定的文件到length个字节，如果想要清空一个文件可以在open时指定flags为O_TRUNC。如果length参数小于文件之前的长度那么length之后的数据被丢弃，如果大于，是系统而定，一般是在文件末尾用\0填充，使文件达到length的长度。
14、文件系统
不详细看了，之前看过。大致都是用node存文件信息，用block存文件实际内容。
15、link、unlink、remove和rename函数
#include <unistd.h>
#include <stdio.h>
int link(const char *oldpath, const char *newpath);
int unlink(const char *pathname);
int remove(const char *pathname);
int rename(const char *oldpath, const char *newpath);linux中ln命令是用来创建文件连接的，ln产生硬链接。ln -s产生符号链接。两者的区别，硬链接的inode没有改变，inode号是文件的唯一标识，所以硬链接没有产生新文件。符号链接产生了新的inode号，产生了新的文件，但是不分配硬盘块。符号链接可以跨分区，可以为目录文件建立符号链接。link和ulink函数用于创建和删除符号链接。ulink删除目录项，并且将pathname所引用的文件连接计数减1，如果该文件还有其他连接，仍可以通过其他链接获取文件的数据。只有当文件的连接计数达到0时，该文件的内容才可以被删除，另外一个最值删除文件内容的条件-只要有进程打开了该文件，其内容也不能删除。ulink这种特性经常被程序用来确保即使是在程序崩溃时，它创建的临时文件也不会遗留下来。用open或creat创建一个文件，然后立即调用unlink，因为文件是打开的，所以不会删除内容，只有进程关闭改文件会终止时，该文件内容才会被删除。
remove相当于rm命令，是使用unlink，rmdir函数封装的，它在删除文件的时候其实没有将文件的数据块从磁盘上移除，而是在被删除的文件没有任何进程引用时才将数据块释放。rename函数用于重命名文件或目录。对于文件remove的功能与unlink相同，对于目录remove的功能与rmdir相同。
文件或目录用rename函数更名。
16、符号链接
符号链接是对一个文件的间接指针，与上一节所述的硬链接有所不同，硬链接直接指向文件的i节点，引进符号链接是为了避免一些硬链接的限制：a）硬链接通常要求连接和文件处于同一文件系统。b）只有超级用户才能创建到目录的硬链接。而符号链接以及它指向什么没有文件系统限制，任何用户都能创建指向目录的符号链接。
当使用以名字引用一个文件的函数时，应该了解该函数是否处理符号链接功能，也就是是否跟随符号链接连接到它所连接的文件。
17、symlink和readlink函数
#include <unistd.h>
int symlink(const char *target, const char *linkpath);
ssize_t readlink(const char *pathname, char *buf, size_t bufsiz);
symlink函数创建一个指向target的新目录项linkpath，创建此符号连接时，并不要求target存在，并且两个也不需要在同一个文件系统中。readlink函数组合了open read close 所有操作，如果此函数成功，则返回读入buf的字节数。在buf中返回的符号链接的内容不以null字符终止。
18、文件的时间
每个文件有三个时间字段，如下表：
19、utime函数
#include <sys/types.h>
#include <utime.h>
int utime(const char *filename, const struct utimbuf *times);
struct timeval {
long tv_sec; /* seconds */
long tv_usec; /* microseconds */
};
一个文件的存取和修改时间可以用utime函数更改，所用结构如上图。
20、mkdir和rmdir函数
//mkdir - create a directory
#include <sys/stat.h>
#include <sys/types.h>
int mkdir(const char *pathname, mode_t mode);
//rmdir - delete a directory
#include <unistd.h>
int rmdir(const char *pathname);与mkdir和rmdir命令一样，用来创建和删除目录，rmdir只能删除空目录，如果要删除费空目录，需要自行递归实现。对目录而言至少需要有执行权限。
21、读目录
对目录的写和执行权限决定了能否在该目录中创建新文件以及删除文件，并不表示能写目录本身。所谓的读目录其实是读出目录中文件列表，对目录的访问分两种方式：glob和xxxdir函数族，这里主要介绍xxxdir函数族。
//opendir, fdopendir - open a directory
#include <sys/types.h>
#include <dirent.h>
DIR *opendir(const char *name);
DIR *fdopendir(int fd);
//readdir - read a directory
#include <dirent.h>
struct dirent *readdir(DIR *dirp);
//closedir - close a directory
#include <sys/types.h>
#include <dirent.h>
int closedir(DIR *dirp);
//telldir - return current location in directory stream
#include <dirent.h>
long telldir(DIR *dirp);
//seekdir - set the position of the next readdir() call in the directory stream.
#include <dirent.h>
void seekdir(DIR *dirp, long loc);opendir函数用于打开一个目录，返回一个指向目录的结构体指针。
readdir函数用于读取目录项，每次调用返回一个目录项，循环调用就可以读出一个目录中所有目录项，返回NULL标书目录中目录项读取完毕。
closedir用于回收资源和open成对使用。而telldir和seekdir用来定位目录项位置指针。
22、chdir、fchdir和getcwd函数
// chdir, fchdir - change working directory
#include <unistd.h>
int chdir(const char *path);
int fchdir(int fd);
char *getcwd(char *buf, size_t size);chdir和fchdir用于改变进程的工作目录，参数path和fd表示要修改的目标目录。cd命令就是用这个函数封装的。getcwd函数用于获取当前工作路径的绝对路径，pwd命令就是用该函数封装的。
23、特殊块设备
st_dev和st_rdev字段很容易引起混淆。每个文件系统都是由主次设备号为人所知，设备号所用的数据类型是基本系统数据类型dev_t。通常可以使用定义的宏，major和minor来获取主次设备号。系统中每个文件名的st_dev是文件系统的设备号，只有字符特殊文件和块特殊文件才有st_rdev值，这个值包含实际设备的设备号。
24、sync和fsync函数
#include <unistd.h>
void sync(void);
int fsync(int fd);
sync将所有修改过的块的缓存排入写队列，然后返回，并不等待实际IO操作结束。系统精灵进程一般隔30秒调用一次sync函数，保证定期刷新内核的块缓存，命令sync也调用sync函数。fsync只引用单个文件，有文件描述符指定。等待IO结束返回。可用于数据库这样的应用，确保修改过的块立即写到磁盘上。
25、文件存取位小结
已经说明了所有文件的存取许可位，某些位有多种用途。如下表：
26、小结
围绕stat函数，介绍stat结构体中的每一个成员。对unix文件的各个属性有所了解。