\documentclass[11pt]{article} \setlength{\oddsidemargin}{0in} \setlength{\evensidemargin}{0in} \setlength{\topmargin}{0.0in} \setlength{\headheight}{0in} \setlength{\headsep}{0in} \setlength{\textwidth}{6.5in} \setlength{\textheight}{9.0in} \setlength{\parindent}{0in} \setlength{\parskip}{0.12in} \usepackage{times} \begin{document} \newcommand{\bfs}[1]{{\bf #1}} \title{File Systems in Unix} \author{Norman Matloff \\ Department of Computer Science \\ University of California at Davis} \date{October 19, 1998} \maketitle \tableofcontents \section{Introduction} In Unix, the files are organized into a tree structure with a root named by the character '/'. The first few levels of the tree look like this: \begin{verbatim} / | --------------------- / | | | \ etc bin usr tmp dev | | ------ -------- / \ / \ ls .. csh ucb ... lib \end{verbatim} Your own files form a subtree of this tree. For example, in many systems the user files are subdirectories of a directory named `home' within `usr'; if we had users Jack and Jill, for example, Jack's home directory would be /usr/home/jack, and all his files would be within that subtree, and the analogous statement would hold for Jill. Suppose Jill's directory looks like this: \begin{verbatim} jill | ---------------------- / | | | \ hill water pail story misc | | ----- -------- / \ | fresh salt rocks \end{verbatim} File names can be given either in relative terms, or with full path names. Look at the file `salt' above. If we are in the directory `water', we can refer to this file as simply \begin{verbatim} salt \end{verbatim} If we are in the directory above, i.e. the one named `jill', then we must write \begin{verbatim} water/salt \end{verbatim} If we are in the directory `misc', we can write either \begin{verbatim} ../salt \end{verbatim} or \begin{verbatim} ~/water/salt \end{verbatim} If we are not in any of Jill's directories, we can write \begin{verbatim} ~jill/water/salt \end{verbatim} In any case, the full pathname will work: \begin{verbatim} /usr/home/jill/water/salt \end{verbatim} \section{File Types} There are four types of files in the Unix file system. \subsection{Ordinary Files} An ordinary file may contain text, a program, or other data. It can be either an ASCII file, with each of its bytes being in the numerical range 0 to 127, i.e. in the 7-bit range, or a binary file, whose bytes can be of all possible values 0 to 255, in the 8-bit range. \subsection{Directory Files} Suppose that in the directory x I have a, b and c, and that b is a directory, containing files u and v. Then b can be viewed not only as a directory, containing further files, but also as a file itself. The \underline{file} b consists of information about the \underline{directory} b; i.e. the file b has information stating that the directory b has files u and v, how large they are, when they were last modified, etc.\footnote{The \bfs{ls} command gets its information about the directory b by reading the file b.} \subsection{Device Files} In Unix, physical devices (printers, terminals etc.) are represented as ``files.'' This seems odd at first, but it really makes sense: This way, the same \bfs{read()} and \bfs{write()} functions used to read and write real files can also be used to read from and write to these devices. \subsection{Link Files} Suppose we have a file X, and type \begin{verbatim} ln X Y \end{verbatim} If we then run \bfs{ls}, it will appear that a new file, Y, has been created, as a copy of X, as if we had typed \begin{verbatim} cp X Y \end{verbatim} However, the difference is the \bfs{cp} does create a new file, while \bfs{ln} merely gives an alternate name to an old file. If we make Y using \bfs{ln}, then Y is merely a new name for the same physical file X. \section{Obtaining Information About the Files in a Given Directory} The `a' (``all'') and `l' (``long'') options of the \bfs{ls} command will give us a lot of information about files in a specified directory (if we don't specify a directory, then the current directory is assumed). Here is a sample output from typing \begin{verbatim} ls -al \end{verbatim} \begin{verbatim} drwxr-xr-x 6 ecs4005 1024 Apr 22 13:30 ./ drwxr-xr-x 74 root 1536 Mar 24 12:51 ../ -rw------- 1 ecs4005 188 Apr 13 15:53 .login -rw------- 1 ecs4005 6 Mar 24 11:29 .logout -rw------- 1 ecs4005 253 Apr 10 12:50 .xinitrc -rw-r--r-- 1 ecs4005 516 Apr 10 13:00 .twmrc -rw-r--r-- 1 ecs4005 1600 Apr 22 10:59 test2.out \end{verbatim} The output is separated into six columns: \begin{verbatim} 1st column - access permissions (see below) 2nd column - number of file entries (in the case of directory files) 3rd column - owner 4th column - size in bytes 5th column - date and time of last modification 6th column - name \end{verbatim} \section{File Access Control} In Unix, all files are protected under some access control mechanism, so that the owner of a file can deny access of his files to other users. The first column of the long directory list shows the access characteristics of a file, in te form of 10 flags, e.g. drwxr-xr-x. The meanings of the flags are shown below: \begin{verbatim} Position 1 file type: d (directory) - (ordinary file) l (symbolic link) Position 2-4 permissions for the owner: r (read) w (write) x (execute) Position 5-7 permissions for other users in the same group Position 8-10 permissions for all other users \end{verbatim} Note that a hyphen (`-') denotes lack of the given permission type. For example, r-x would mean that read and execute permission are granted, but not write permission. In order to remove a file, you must have write permission for it. In order to view the contents of a directory, i.e. see what files are there, you need read permission for that directory. In order to actually access a file (read from it, write to it, or execute it) in the directory, you need execute permission for the directory. \section{Some File Commands} \subsection{chmod} You can use this command to change the access permissions of any file for which you are the owner. The notation used is: \begin{verbatim} u user (i.e. owner) g group o others + add permission - remove permission r read w write x execute \end{verbatim} As an example, the command \begin{verbatim} chmod ugo+rw .login \end{verbatim} would add read and write permission for all users to the .login file of the person issuing this command. In some cases it is useful for a user to deny himself/herself permission to write to a file, e.g. to make sure he/she doesn't remove the file by mistake. \subsection{du and df} The du command displays the sizes in kilobytes of all files in the specified directory, and the total of all those sizes; if no directory is specified, the current directory is assumed. The df command displays the amount of unused space left in your disk systems. \subsection{diff} This command displays line-by-line differences between two ASCII files. If for example, you have two versions of a C source file but don't remember how the new version differs from the old one, you could type \begin{verbatim} diff oldprog.c newprog.c \end{verbatim} \section{Wild Cards} These will save you a lot of typing! There are two wild-card characters in Unix, `*' and `?'. The wildcard `*' will matches with any string of characters. For example, \begin{verbatim} rm *.c \end{verbatim} would delete all files in the current directory whose names end with `.c'. The wildcard `?' will match with any single character. For example, \begin{verbatim} rm x?b.c \end{verbatim} would delete all files whose names consisted of five characters, the first of which was `x' and the last three of which were `b.c'. Example: rm prog?.c will delete all the files (in the current directory) The files x3b.c and xrb.c would be deleted, while the file xuvb.c would not. In addition, \begin{verbatim} [0-9] matches character from `0' through `9' [a-z] matches character from `a' through `z' \end{verbatim} For instance, \begin{verbatim} rm test[1-3].c \end{verbatim} would remove test1.c, test2.c and test3.c but not test4.c. \end{document}