The kernel maintains a single hierarchical directory structure to organize all files in the system. (This contrasts with operating systems such as Microsoft Windows, where each disk device has its own directory hierarchy.) At the base of this hierarchy is the root directory, named / (slash). All files and directories are children or further removed descendants of the root directory. Figure 2-1 shows an example of this hierarchical file structure.
Within the file system, each file is marked with a type, indicating what kind of file it is. One of these file types denotes ordinary data files, which are usually called regular or plain files to distinguish them from other file types. These other file types include devices, pipes, sockets, directories, and symbolic links.
The term file is commonly used to denote a file of any type, not just a regular file.
A directory is a special file whose contents take the form of a table of filenames coupled with references to the corresponding files. This filename-plus-reference association is called a link, and files may have multiple links, and thus multiple names, in the same or in different directories.
Directories may contain links both to files and to other directories. The links between directories establish the directory hierarchy shown in Figure 2-1.
Every directory contains at least two entries: . (dot), which is a link to the directory itself, and .. (dot-dot), which is a link to its parent directory, the directory above it in the hierarchy. Every directory, except the root directory, has a parent. For the root directory, the dot-dot entry is a link to the root directory itself (thus, /.. equates to /).
Like a normal link, a symbolic link provides an alternative name for a file. But whereas a normal link is a filename-plus-pointer entry in a directory list, a symbolic link is a specially marked file containing the name of another file. (In other words, a symbolic link has a filename-plus-pointer entry in a directory, and the file referred to by the pointer contains a string that names another file.) This latter file is often called the target of the symbolic link, and it is common to say that the symbolic link “points” or “refers” to the target file. When a pathname is specified in a system call, in most circumstances, the kernel automatically dereferences (or synonymously, follows) each symbolic link in the pathname, replacing it with the filename to which it points. This process may happen recursively if the target of a symbolic link is itself a symbolic link. (The kernel imposes limits on the number of dereferences to handle the possibility of circular chains of symbolic links.) If a symbolic link refers to a file that doesn’t exist, it is said to be a dangling link.
Often hard link and soft link are used as alternative terms for normal and symbolic links. The reasons for having two different types of links are explained in Chapter 18.
On most Linux file systems, filenames can be up to 255 characters long. Filenames may contain any characters except slashes (/) and null characters (\0). However, it is advisable to employ only letters and digits, and the . (period), _ (underscore), and - (hyphen) characters. This 65-character set, [-._a-zA-Z0-9], is referred to in SUSv3 as the portable filename character set.
We should avoid the use of characters in filenames that are not in the portable filename character set because those characters may have special meanings within the shell, within regular expressions, or in other contexts. If a filename containing characters with special meanings appears in such contexts, then these characters must be escaped; that is, specially marked—typically with a preceding backslash (\)—to indicate that they should not be interpreted with those special meanings. In contexts where no escape mechanism is available, the filename is not usable.
We should also avoid filenames beginning with a hyphen (-), since such filenames may be mistaken for options when specified in a shell command.
A pathname is a string consisting of an optional initial slash (/) followed by a series of filenames separated by slashes. All but the last of these component filenames identifies a directory (or a symbolic link that resolves to a directory). The last component of a pathname may identify any type of file, including a directory. The series of component filenames preceding the final slash is sometimes referred to as the directory part of a pathname, while the name following the final slash is sometimes referred to as the file or base part of the pathname.
A pathname is read from left to right; each filename resides in the directory specified by the preceding part of the pathname. The string .. can be used anywhere in a pathname to refer to the parent of the location so far specified in the pathname.
A pathname describes the location of a file within the single directory hierarchy, and is either absolute or relative:
An absolute pathname begins with a slash (
/) and specifies the location of a file with respect to the root directory. Examples of absolute pathnames for files in Figure 2-1 are/home/mtk/.bashrc,/usr/include, and/(the pathname of the root directory).A relative pathname specifies the location of a file relative to a process’s current working directory (see below), and is distinguished from an absolute pathname by the absence of an initial slash. In Figure 2-1, from the directory
usr, the filetypes.hcould be referenced using the relative pathnameinclude/sys/types.h, while from the directoryavr, the file.bashrccould be accessed using the relative pathname../mtk/.bashrc.
Each process has a current working directory (sometimes just referred to as the process’s working directory or current directory). This is the process’s “current location” within the single directory hierarchy, and it is from this directory that relative pathnames are interpreted for the process.
A process inherits its current working directory from its parent process. A login shell has its initial current working directory set to the location named in the home directory field of the user’s password file entry. The shell’s current working directory can be changed with the cd command.
Each file has an associated user ID and group ID that define the owner of the file and the group to which it belongs. The ownership of a file is used to determine the access rights available to users of the file.
For the purpose of accessing a file, the system divides users into three categories: the owner of the file (sometimes termed the user of the file), users who are members of the group matching the file’s group ID (group), and the rest of the world (other). Three permission bits may be set for each of these categories of user (making a total of nine permission bits): read permission allows the contents of the file to be read; write permission allows modification of the contents of the file; and execute permission allows execution of the file, which is either a program or a script to be processed by some interpreter (usually, but not always, one of the shells).
These permissions may also be set on directories, although their meanings are slightly different: read permission allows the contents of (i.e., the filenames in) the directory to be listed; write permission allows the contents of the directory to be changed (i.e., filenames can be added, removed, and changed); and execute (sometimes called search) permission allows access to files within the directory (subject to the permissions on the files themselves).