An Introduction to Linux

Anubhav Seth
Souri Guha
Table of Contents
1 Introduction

2 Logging In and Out

2.1 Logging In
2.2 Logging Out

3 The Graphical User Interface

4 The Linux Filesystem

4.1 The Working Directory
4.2 Listing Directory Contents
4.3 Wildcard Characters
4.4 Multiple Disks in the Filesystem

5 Viewing Files
5.1 Text Files
5.2 Other File Types

6 Online Help
6.1 Online Manual
6.2 Apropos
6.3 Info
6.4 The Gnome Help Browser

7 Manipulating and Processing Files

7.1 Copying Files
7.2 Moving and Renaming Files
7.3 Removing Files
7.4 Creating Subdirectories
7.5 Removing Subdirectories
7.6 Symbolic Links
7.7 I/O Redirection and Pipes
7.8 Printing
7.9 Searching for Files
7.10 Working with Floppy Disks

8 Security
8.1 Changing your Password
8.2 File Permissions

9 Processes
9.1 Listing Processes
9.2 Ending a Process
9.3 Running a Process in the Background

10 Creating and Editing Text Files

10.1 Text Editors
10.2 Emacs

11 Compiling Programs
11.1 Compiling From the Shell
11.2 Compiling From Within Emacs
11.3 Using Make
11.4 Debugging

12 Networking
12.1 World Wide Web
12.2 Remote Login
12.3 Remote X Windows
12.4 Novell Access

13 Customizing your Account with Shell Scripts

14 How to Learn More

Appendix: Linux Command Reference

Introduction
The original UNIX operating system was invented by Dennis Ritchie and Ken Thompson in 1969 on a DEC PDP-7 at Bell Labs.
In the years since, development has splintered into several UNIX variants, some proprietary and some open source.
All of these “UNIX-like” systems are bound by a set of standards called POSIX which ensures that code developed on one
POSIX compliant operating system can be recompiled on another. Linux (pronounced “L I-nuks”, with a short i) is a UNIX-like
operating system that was born in 1991 as the hobby of a young
student, Linus Torvalds, at the University of Helsinki in Finland. Thousands of developers have since joined the Linux project,
and Linux has grown into a mature and powerful operating system. Linux is now widely used around the world by individuals
and companies as diverse as L.L. Bean, Home Depot, Amazon.com, and Dreamworks SKG. It is supported by big computer
makers like IBM, Dell, and Compaq.
Linux is developed and distributed under the GNU General Public License (GPL). The GNU (pronounced “guh-NEW”) project
( www.gnu.org ), from which the GNU General Public License originates, was founded in 1984 with the goal of developing
high quality, free software. Most of the software bundled with Linux comes from the GNU project. GNU is a recursive acronym
for “GNU’s Not UNIX”. The name is a reaction to the original proprietary implementations of UNIX, to which the founders
objected. In a nutshell, the GPL dictates that the operating system is free of cost and that its source code is freely available for
anyone to examine or change.

2 Logging In and Out

If you do not already have a Linux account, your instructor should be arranging for one for you. After you receive an account,
you will be able to log into any of the UNIX machines with the same username and password. Each of the machines reads
password information from our file server (named sunshine ). Your personal home directory, in which you can store your
personal files, will also be accessible from every machine. Your home directory has the pathname /home/username

2.1 Logging In
If you sit down at a machine and the screen is black, wiggle the mouse a little. The screen saver will disappear and a login
screen will appear. Just type your username in the appropriate box, hit Enter, enter your password, and hit Enter again.

2.2 Logging Out

Later, when you are ready to log out, you can do so from the Panel menu in the bottom left corner of the screen. You might
want to check the button in the logout window to remember the current state of your session. Most of the programs you leave
running will be started up automatically when you next log in. However, you should always quit Netscape before logging out.
For reasons unknown, Netscape may continue to run and slow the machine down if you do not.

3 The Graphical User Interface

Basic GUI functionality in Linux is provided by a program called X Windows . The actual “look and feel” of the GUI you are using is
specified by a program called a window manager . In your case, you are probably running either the Sawfish or Enlightenment
window manager (depending upon when your account was created) in the Gnome desktop environment. Gnome essentially runs
“beneath” the window
manager to provide some useful tools and a uniform “look and feel” across different window managers. You can change the
“look and feel” by using a different window manager and/or by changing your preferences with the Gnome Control Center ,
accessible by pressing the button with the toolbox on it in the Panel at the bottom of your screen (or by typing gnomecc in a
shell window; more later).

When you are ready, click on the Panel button that looks like a monitor (or sometimes, a footprint). This will bring up a terminal
window into which you can type Linux commands. The program running in the terminal window that accepts these commands
is called a shell . The default shell is called bash (the GNU Bourne-Again SHell). When you start the shell, you only see a
prompt like:
razr@localhost>

The prompt indicates that the shell is ready to accept your commands. In this example, the prompt indicates that user razr is
logged into the computer named localhost . As you learn more about Linux, you will find that many things can be accomplished
either by “pointing and clicking” with the mouse or by typing commands in a shell. While using the mouse is often more
convenient, shell commands can give you greater control over an operation. They are also often more convenient for performing
some quick file operations. We will cover many shell commands later in this tutorial.

4 The Linux Filesystem

Linux maintains a hierarchical file system, similar to what you are used to with Windows or Mac OS. The file system is called
hierarchical because directories (or folders) can be created inside other directories to create a hierarchical tree-like structure.
Unlike Windows and Mac OS, however, this tree has only one root, called the root directory , represented by / (a forward slash).
4
4.1 The Working Directory
Relative pathnames are always specified with respect to the working directory , which is a special directory remembered by the
shell, and is interpreted to be the directory you are currently “in”. To see your current working directory in the shell, type

pwd Enter

in the terminal window. (From now on, we will omit the Enter after commands; you will need to hit this key after every
command you enter.) You should see something like
razr@localhost> pwd
/home/razr
to indicate that your are currently in your home directory. (The subdirectory razr will be, of course, replaced by your user
name.) Whenever you create or refer to a file using a relative pathname, the pathname of the working directory is simply
prepended to get the correct absolute pathname.
You can also specify pathnames that go “up” in the tree by using the special symbol “ .. ” which refers to the parent directory
of the current directory. For example, in Figure 1, if the current working directory were /bin , then the relative pathname
../home/lisa/hello.cc would refer to the file hello.cc in the subdirectory lisa .
You can use the cd command to change your working directory. For example, to change your working directory to the parent
directory, type
cd ..
in the the terminal window. To confirm the change, type pwd . You will now see:
razr@localhost> pwd
/home
To change back to your home directory, type
cd razr
(replacing razr with your username, of course). Here is a shortcut: typing cd by itself will always return you to your working
directory from anywhere.

4.2 Listing Directory Contents

The ls command is used to list the contents of a directory. To list the contents of the working directory, type
ls
by itself. (If this is your first time using Linux, there may be no files to see quite yet.) To list the contents of another directory,
either use cd to change to that directory and then type ls , or specify the name of the directory after the ls command. For
example,
ls /bin
will list all the files and subdirectories in the /bin directory.
Linux commands can be augmented with options that follow a command and are usually prefixed by a - (a dash). For example,
type
ls -a
from your home directory. Now you will notice that more files show up. These are called dot files since they all start with a dot
(a period). These files contain configuration information for various programs (including the shell; more on this in Section 13)
and are normally hidden. The -a option (short for all ) displays these files in addition to the other files in the directory. The first
two files that you see ( . and .. ) are special characters that refer to the directory itself and the parent directory, respectively.
While we are on the topic of special characters, it is worthwhile to mention that a tilde character ( ˜ ) always refers to your home
directory. So, for example, the command
ls ˜
lists the contents of your home directory.
You may have noticed that some file names were followed by special symbols like / , * , or @ in the directory listing. The /
indicates that the file is actually a subdirectory. The * indicates that the file is an executable program. The @ indicates that the
file is actually a symbolic link (or alias) to another file somewhere else in the filesystem. (If you did not see these characters,
use the -F option with ls .)
Now try typing
ls -al
in the shell. In this case, you are combining two options — the -a option and the -l option (for long ). You will see an
expanded file listing, something similar to (but not exactly like) this:

total 27
drwx------ 70 joe joe 6144 Jul 30 13:02 ./
drwxr-xr-x 102 root root 2048 Mar 26 09:23 ../
-rw-r--r-- 1 joe joe 1067 Aug 16 1999 .Xdefaults
-rw-r--r-- 1 joe joe 1100 Mar 27 16:27 .bash_profile
-rw-r--r-- 1 joe joe 1625 Jan 12 2001 .bashrc
-rw-r--r-- 1 joe joe 4581 Feb 1 16:17 .emacs
drwxr-xr-x 12 joe joe 1536 Jul 23 10:31 .gnome/

drwxrwxr-x 2 joe joe 512 Mar 30 13:04 .gnome-desktop/

drwxr-xr-x 7 joe joe 1024 Jul 27 17:02 .netscape/
drwxrwxr-x 3 joe joe 512 Jan 3 2001 .sawfish/
-rwxr--r-- 1 joe joe 1606 Aug 18 1998 .xinitrc*
lrwxrwxrwx 1 joe root 8 Aug 13 1999 .xsession -> .xinitrc*

The first line gives the total amount of disk storage used by the files in this directory, in disk blocks. Each additional line in the
display gives information about one file or subdirectory in the directory. The first character in the first column in the file listing
tells you what type of file you are looking at. A - means that this is a normal file, a d means that this is actually a subdirectory,
and a l means that this is a symbolic link (more on links in Section 7.6). The rest of the characters in the first column give
information about file permissions. We will talk about this more in Section 8.2. The second column gives the number of hard
links to this file or directory (again, more in Section 7.6). The third columns tells who the owner of the file or directory is.
Every file in a Linux filesystem has an owner who can decide who has access to that file. In this case, joe owns every file in
his home directory except the parent directory. The fourth column indicates the group of each file or directory. Groups can be
used to give access to a file to a group of specific users. The fifth column gives the size of the file or directory in bytes. The
sixth column gives the time the file was last modified. Finally, the seventh column gives the name of the file or directory.
4.3 Wildcard Characters
Sometimes it is convenient to specify a group of files or directories as the argument to a command. If the names of these files
are similar in some way, wildcard characters may be used. There are two main wildcard characters: a * can be substituted for
any string of 0 or more characters and a ? can be substituted for exactly one character. For example, if you wanted to list all
files in the working directory that start with prog , you could type
ls prog*
in the shell. If you wanted all the files with a .cc extension, you could type
ls *.cc
in the shell. If you wanted all the files with names like prog1.cc , prog2.cc , etc., but not prog.cc , you could type
ls prog?.cc
in the shell.
Wildcard characters can be used anywhere you would normally use a file or directory name.

4.4 Multiple Disks in the Filesystem

In Windows and Mac OS, different disks maintain their own filesystems (directory trees) and each is named with a letter like
C: (Windows/DOS) or a name like Hard Drive (Macintosh). So how are multiple disks accessed in the Linux filesystem
with only one root? The answer is that each is mounted at a separate directory in the filesystem. So, for example, a separate disk
may be mounted at the directory /home . In this case, every time you access something in directory /home , you are actually
accessing a file on this separate disk. This is transparent to the user, but you can peek at which disks are mounted where on a
computer by using the df command in the shell. You will see something like this:

Filesystem 1k-blocks Used Available Use% Mounted on

/dev/hda8 497667 78511 393454 17% /
/dev/hda1 14607 2758 11095 20% /boot
/dev/hda7 5308207 1085340 3948063 22% /usr
sunshine:/export/home.linux
8357259 7024141 497393 94% /home
sunshine:/export/local.linux
8357259 7024141 497393 94% /usr/local
/dev/hdc 609750 609750 0 100% /mnt/cdrom

The first line indicates that the disk with symbolic name 5 /dev/hda8 is mounted at the root directory. The second line
indicates that another disk with symbolic name /dev/hda1 is mounted at subdirectory /boot . The third line is similar.

The disk in the fourth line is actually a subdirectory on a disk on another networked computer ( sunshine ). The /home
directory on every computer in the lab points to this same directory on sunshine . This is how you are able to access your
home directory (which is actually on sunshine ) the same way from every computer! The protocol that allows this seamless
mounting of remote disks is called the Network File System (NFS). The fifth line is similar to the fourth, and the sixth line
represents the CD-ROM drive, which is mounted at /mnt/cdrom .

5 Viewing Files
5.1 Text Files
The easiest way to view the contents of a file is by using the cat command. The name of the command sounds odd at first, but
it will make more sense when we revisit it in Section 7.7. For example, to view the file named /etc/passwd , simply type
cat /etc/passwd
in the shell. (This file contains some encrypted passwords for this machine.) The cat command is really only good for looking
at small files, since the entire file is displayed without stopping. To view longer files, it is better to use the more command. For
example, try to view the file again by typing
more /etc/passwd
in the shell. Notice that, this time, the command stops after one page is displayed. To view the next line, hit Enter . To view
the next page, hit the space bar. To move backward one page, type b . To exit, type q . If you want to search for something
while in more , type / followed by what you want to find and then hit Enter .

5.2 Other File Types

Files other than text files can viewed by using an appropriate program, as indicated in the table below. Follow the command
with the name of the file to view it or omit the file name and select the file from the menu once the program is running.

File Type Program Name Command

Postscript gv (based on GhostView) gv
Portable Document Format (PDF) Adobe Acrobat Reader acroread
Image files XV xv
Image files The Gimp gimp
LaTeX ouput (DVI) xdvi xdvi
Hypertext (HTML) Netscape netscape
MP3 and MPEG gtv gtv

6 Online Help
The commands we have covered so far are just a tiny portion of those available. So how do you find out about these commands
and remember all those you have learned? You don’t; you use the online manuals instead!

6.1 Online Manual

The online manual can be invoked with the man command. For instance, to find more information about the ls command, type

man ls
from a shell. The manual page is displayed using more , so hit the space bar to see each page. If you want to find out more about
man itself, type man man . There are manual pages on most Linux shell commands and on most standard C/C++ functions
and Linux system calls.

6.2 Apropos
If you’re looking for how to do something, but don’t know the command yet, try the apropos command. For example, if you
want to display a new clock window, try typing apropos clock . A list of matching manual sections will be listed. There is
an stunning amount of software installed on these machines. Exploring is the only way to found out about it.

6.3 The Gnome Help Browser

You can also access the manual pages and info pages through the Gnome Help Browser. If you are sitting at the console of a
machine, this option is preferable to the command line alternatives since navigation is all via point-and-click. The Gnome Help
Browser can be accessed by pressing the Panel button with a question mark on it (or with gnome-help-browser in the
shell).
If the online help functions do not answer your question, you may want to consult one of the Linux books kept in the depart-
ment office (Olin 201) or a book in the bibliography at the end of this document. If you are really stuck, ask Tony Silveira
( [email protected] ) or Professor Havill ( [email protected] ).

7 Manipulating and Processing Files

7.1 Copying Files
To make a copy of an existing file, use the cp command. For example, to make a copy of a file myfile.txt called
myfile2.txt in the working directory, you would type:
cp myfile.txt myfile2.txt
If you want to create a copy of myfile.txt in a cs100 subdirectory, type:
cp myfile.txt cs100
Notice that, if the destination file name is missing, the copy will have the same name as the original.

7.2 Moving and Renaming Files

The mv command is similar to the cp command except that the original file is not preserved. Instead, the file is “moved” to the
destination location. For example, if you had typed
mv myfile.txt cs100
above instead of cp myfile.txt cs100 , the file would have been moved to the subdirectory rather than copied.
The mv command is also used to rename files. For example, typing
mv myfile2.txt myfile3.txt
will rename myfile2.txt to myfile3.txt .

7.3 Removing Files

To remove (delete) a file or group of files, use the rm command. For example,
rm myfile3.txt
deletes the file myfile3.txt . The command
rm *
will remove all the files in the current directory. (Be careful using rm with wildcard characters!)

7.4 Creating Subdirectories

In order to organize all of your files, you will want to create subdirectories in your home directory. To create a new subdirectory,
use the mkdir command. For example, to create a new subdirectory named cs100 in your home directory, change to your
home directory and type
mkdir cs100
in the shell. If you now type ls , you will see your new subdirectory.

7.5 Removing Subdirectories

To delete the subdirectory cs100 , type:
rmdir cs100
To remove (delete) a subdirectory, the subdirectory must be empty. To delete files in a directory, you will need to use the rm
command (Section 7.3).
The rm -R command can be used to remove a non-empty directory. The -R option is short for “recursive”. This command
deletes all the files in the directory and then the directory itself.

7.6 Symbolic Links

A symbolic (or soft) link is an indirect reference to a file, similar to an “alias” in Mac OS or a “shortcut” in Windows. Sup-
pose you use a particular file often, but it is deeply nested in a subdirectory 4 levels down from your home directory. Rather
than typing the full pathname of this file (or changing to that subdirectory) every time you want to access it, you can cre-
ate a symbolic link to it in your home directory (or somewhere else handy). Then whenever you want to access that file,
you can just access the symbolic link instead of the file itself. For example, suppose this commonly accessed file is called
˜/personal/misc/letters/contacts . To create a symbolic link to this file called buddies in your home directory,
you would type:
ln -s ˜/my/misc/stuff/friends ˜/buddies
Typing ls -l in your home directory would now yield an entry like the following for your symbolic link:

srw-r--r-- 1 joe joe 40 Aug 4 11:01 buddies -> /home/joe/my/misc/stuff/friends

Referring to this symbolic link (with cat , for instance) will always refer to the real file /home/joe/my/misc/stuff/friends .
You can also use symbolic links with directories.
Linux actually supports two different kinds of links: hard and symbolic (or soft ). A hard link is a direct link to a lower level
operating system maintained entry for a file on disk, and is less exible than a symbolic link. Using the ln command without
the -s option creates a hard link instead of a symbolic link.
7.9 Searching for Files
7.9.1 Searching by File Name

If you need to access a specific file in the filesystem and you do not know where it is located, you can use the find command
to find it. For example, the command
find . -name paper.txt -print
will search for the file paper.txt starting in the current working directory ( . ). The command
find /bin -name ma* -print
will search for all files that begin with the characters ma , starting in the directory /bin . In each case, the absolute pathname of
each matching file is printed. Although the syntax of the find command looks strange, it gets the job done. There are many
other ways to search for files with find . Consult the online manual (Section 6) for more information.

7.9.2 Searching by Content

Sometimes it is useful to be able to search for a file based on the file’s content instead of its name. The grep command can be
used to do just this. For example, the command
grep example ./*
will search all files in the current directory for the string example . You can also use wildcards (and, in general, regular
expressions) with grep . For example, the command
grep ex*le ./*
will search all files in the current directory for strings starting with ex and ending with le .

8 Security
8.1 Changing your Password
It is always a good idea to change your password periodically. To do so, use the yppasswd command. The yp at the beginning
of the command is short for “yellow pages”, the old name of the Network Information Service (NIS), the system that we use to
maintain passwords on our network. When you issue the yppasswd command, you will prompted for both your old password
and your new password.
When you choose a new password, it is important to choose something that is easy for you to remember, yet hard for others
(including automatic “cracker” software) to guess. Use a combination of lower and upper case letters, punctuation, and digits
in your password.

8.2 File Permissions

Recall that when you issued the ls -al command earlier, you were presented with an expanded file listing with a line like the
following for each file:

-rwxr--r-- 1 joe joe 1606 Aug 18 1998 .xinitrc*

Recall that the first character in the first column ( - ) indicates that the file is a normal file. The next 9 characters indicate the
access permissions for the file. The 9 characters are divided into 3 groups of 3. Each group indicates whether read (r), write
(w), and execute (x) permission is granted to a particular group of users. The first group, corresponding to the first 3 characters,
is the “user” (or file owner), the second group is the file “group”, and the third group is “other”, which consists of all users on
the system. So the line above indicates that the owner has read, write, and execute permission for the file and everyone else has
only read permission.
When the file is actually a directory, read and write permissions refer to the ability to read the contents of the directory and
create new entries in the directory. Executable permission means that one can search in the directory (or list its contents with
ls ), but not read from or write to it.
The chmod command is used to change the access permissions of a file or a directory. There are two ways to specify permis-
sions for a file with chmod . The first is to use a symbolic mode representation of the permissions. In this mode, you use some
combination of u for user, g for group, o for other, or a for all to specify which of these group(s) of users’ permissions to
modify. Then use + to add a permission or - to take away a permission. Lastly, specify the permission(s) to add/subtract: r for
read, w for write, or x for execute. For example, suppose the current working directory contained the following files:

drwxr-xr-x 1 joe joe 512 Jul 23 16:56 doc

-rwxr-xr-x 1 joe joe 23034 Aug 4 11:03 prog1
-rw-r--r-- 1 joe joe 1501 Aug 4 11:01 prog1.cc

Then the command

chmod g+w *
gives write permission to the group for all files in the current directory. The permissions of the files would then look like:

drwxrwxr-x 1 joe joe 512 Jul 23 16:56 doc

-rwxrwxr-x 1 joe joe 23034 Aug 4 11:03 prog1
-rw-rw-r-- 1 joe joe 1501 Aug 4 11:01 prog1.cc

If you then issued the

chmod a-rw prog1*
command, this would prevent anyone from reading or writing prog1 and prog1.cc . The resulting directory listing would
look like:

drwxrwxr-x 1 joe joe 512 Jul 23 16:56 doc

---x--x--x 1 joe joe 23034 Aug 4 11:03 prog1
---------- 1 joe joe 1501 Aug 4 11:01 prog1.cc

It is important to restrict access to your home directory. The command

chmod go-rwx ˜
takes all permissions for your home directory away from everyone but the user. This is how the permissions of your home
directory should be set. If they are not, type this command now.
The second way to use the chmod command is to use octal mode . Think of the rwx permissions for a group ( u , g , or o ) as a
representation of a 3 bit binary number (0 – 7 in decimal). If a permission is set, that bit is a 1; otherwise it is a 0. Use the 3
octal (or decimal) digits for the 3 groups as the permission. For example, r-x is equivalent to 4+1=5 and -w- is equivalent to
2. Taken as a group, rw-r--r-- is equivalent to 644 . If you wanted to grant these permissions to the file prog1.cc , you
would type:
 chmod 644 prog1.cc
To give yourself all permissions to your home directory, but everyone else none, you would type:
chmod 700 ˜

9 Processes
Linux, like most modern operating system, is multitasking , which means that it can execute many programs simultaneously. A
program that is currently executing is called a process . A Linux system consists of several dozen active processes at any time.
Some of these processes are system processes that perform important “behind the scenes” tasks and some are user processes
corresponding to programs like Netscape or Emacs.

9.1 Listing Processes

You can view the processes that are currently running on the system with the ps command. Type
ps
now and you will see something like the following:

PID TTY TIME CMD

23985 pts/0 00:00:01 bash
24001 pts/0 00:00:00 ps

This tells you that you are running 2 processes and the commands ( CMD ) used to start them were bash (your shell) and ps
(the process you just started when you entered ps ). The first column in the list, headed by PID , gives the process ID for each
process, an integer used to uniquely identify each process. The next column, TTY , gives the terminal name to which the process
is attached. The third column, TIME , gives the amount of time the process has been running for (or the amount of time it has
been running since it was last sleeping or stopped).
Using ps by itself only gives information about processes that you own, which, unless you are a system administrator, is usually
all the information you need. If you want information about all the processes running on the system, you can issue the
ps aux
command. This command combines 3 options (the normal dash before options is not necessary) which, together, give informa-
tion about all processes that have users have attached to them. Here is a (significantly trimmed) example of what this output
might look like:

USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.0 1324 76 ? S Jul23 0:12 init [5]
root 2 0.0 0.0 0 0 ? SW Jul23 0:21 [kflushd]
root 3 0.0 0.0 0 0 ? SW Jul23 0:03 [kupdate]
root 4 0.0 0.0 0 0 ? SW Jul23 0:00 [kpiod]
root 5 0.0 0.0 0 0 ? SW Jul23 0:14 [kswapd]
root 6 0.0 0.0 0 0 ? SW< Jul23 0:00 [mdrecoveryd]
root 415 0.0 0.1 1548 160 ? S Jul23 0:02 syslogd -m 0
rpc 440 0.0 0.1 1472 252 ? S Jul23 0:00 portmap
root 456 0.0 0.0 0 0 ? SW Jul23 0:00 [lockd]
root 500 0.0 0.2 5744 272 ? S Jul23 0:00 [ypbind]
nobody 566 0.0 0.0 7616 24 ? S Jul23 0:00 [identd]
.
.
.
root 1006 14.7 24.7 62796 31612 ? R Jul23 53:50 /etc/X11/X
joe 7892 0.0 1.2 7580 1564 ? S 14:15 0:00 gnome-session
joe 7957 0.0 1.3 4700 1700 ? S 14:15 1:39 sawmill

joe 7966 0.0 2.1 9336 2784 ? S 14:15 3:51 panel

joe 7969 0.0 0.8 8404 1092 ? S 14:15 0:00 gmc
joe 7972 0.0 0.7 3560 968 ? S 14:15 0:11 xscreensaver
joe 7992 0.0 1.3 7676 1676 ? S 14:26 0:21 gnome-terminal
joe 7994 0.0 0.6 2440 796 pts/0 S 14:26 0:01 bash
joe 12504 0.0 0.1 2004 156 pts/0 S 15:08 0:00 netscape
joe 24059 0.0 0.7 2728 900 pts/0 R 15:42 0:00 ps aux

This listing gives a lot of information about the 100 or so processes that are probably currently executing on the system. For
example, you can gather that the computer was rebooted last on July 23. You can see that X windows ( /etc/X11/X ) is using
almost 25% of the computer’s memory at the moment and 15% of the CPU time. You can also view the current status ( STAT )
of each of the processes in the system. An R in this column means that the processing is currently “runnable” which, as you
will learn when you take Operating Systems, means that the process is either currently running or in the ready queue, waiting
to gain access to the processor. An S indicates that the process is “sleeping” (currently not waiting to use the processor). A
W means that, in addition to sleeping, the process has been “swapped out” (not currently resident in memory but on the disk
instead). Lastly, the < marks a process that is given higher priority than normal while waiting for the CPU.

9.2 Ending a Process

Sometimes you will come across a process that locks up and just won’t quit, no matter what you do. In this case, the only
alternative is to force the process to quit. If the process is one that you started from the shell in the foreground, you can end
it by typing Ctrl - C . If the process is running in the background, you will need to issue the kill command followed by a
process ID. For example, to kill the netscape process above, you would type:
kill 12504
If this does not work, use the -9 (or -s SIGKILL ) option to forcibly kill the process.
The kill command (or Ctrl - C ) also comes in handy when you must kill a program that you wrote with an infinite loop!
9.3 Running a Process in the Background
When you run a program in Linux, it can either be run in the foreground or in the background . When you run a program in the
foreground (which is the default), the resulting process must finish before you get another prompt in the shell. This is fine when
you run short processes like ls , for example. However, if you want to run Netscape (for example) from the shell, you probably
do not want to have to quit Netscape in order to get a new prompt to, say, copy a file. The alternative is to run Netscape in
the background, which means that the new process will run concurrently with the shell prompt, and a prompt will be returned
to you immediately. To run a program in the background, simply follow the command with an ampersand ( & ) character. For
example, to run Netscape in the background, type:
netscape &
Generally speaking, any time you run a program that generates its own window on the desktop, it is best to run it in the
background.
If you start a program in the foreground by mistake and want to turn it into a background process, hit Ctrl - Z , which suspends
the process. Then type
bg
which restarts the currently suspended process in the background.

10 Creating and Editing Text Files

To write programs, you will need to use a text editor . A text editor is a program that edits text files and is different from a word
processor in that the resulting files do not contain formatting information of any kind.

10.1 Text Editors

In Linux, there is a wide range of text editors available. Below we list a few of these. See the man pages or the application’s
help function for more information on each one.

• Text-oriented Editors (work in a terminal window):

vi and vim (“vi improved”), pico , and emacs (more information in the next section)
• Basic GUI Editors :
gnp (Gnotepad+), gedit , and kedit
• Code-oriented GUI Editors (with syntax highlighting, indentation, etc.):
glimmer , nedit , gvim , kwrite , and xemacs (more information in the next section)

10.2 Emacs
Although you may use whichever text editor you prefer, we recommend that you learn Emacs. Emacs is an extremely powerful
and exible program, capable of doing much more than just editing text. There exist fervent Emacs aficionados that prefer to
do all of their work through Emacs; they rarely ever type anything into a terminal window! However, Emacs can also be used
in a very intuitive manner, just like any other text editor you may have used in the past.
There are two versions of Emacs available. One version, simply called emacs , works either in text mode in a terminal window
or with a GUI in its own window. Which mode is invoked depends upon whether you are logged in remotely or locally. The
other version, xemacs , only has a GUI, but with more extensive menus and a nicer scroll bar.

In Figure 2, you will find a table of some of the more commonly used Emacs key combinations. If you wish to learn more
advanced features, you are encouraged to browse the detailed, easy-to-use help facility available within the application. There
is also an Emacs book in the department office that we recommend you consult, as appropriate.

11 Compiling Programs
The standard UNIX compilers are the GNU compilers, distributed as part of the GNU project. The C++ compiler is called g++
and the C compiler is called gcc . You can learn more about g++ later by looking at the man pages. (There’s a lot there!) For
now, we’ll just look at the basics.

11.1 Compiling From the Shell

To compile a C++ program named sort.cc in the shell, simply type
g++ -o sort sort.cc
in the terminal window. (The .cc extension is the one most commonly used for C++ programs in Linux. You can also use
.cpp extension, if you prefer.) The -o option gives the name of the executable to create; if omitted, the executable will be
called a.out .
In general you’ll probably want to compile programs with more options specified. Here is a better way to compile a program:
g++ -g -Wall -o sort sort.cc
The -g option produces debugging information in the executable, for use with a debugger such as DDD (more in Section 11.4).
The -Wall option instructs the compile to display “all warnings”. Be sure to take warnings seriously, as they may point to a
looming problem.
To compile multiple source files into one executable, just include them all on the command line. For example, to compile
sort.cc and swap.cc into a single executable called sort , type:
g++ -g Wall -o sort sort.cc swap.cc
Once you have a working executable program, go to a terminal window to execute it. Make sure you are in the correct directory
and then type the name of the executable program.
11.3 Using Make
When you are creating large software projects, it is usually a good idea to break the code into multiple source files. This way,
errors are usually easier to find and edit, and when you recompile the program you only need to recompile the file you changed
and those upon which it depends. Of course, this can become confusing and complicated when the system becomes large.
Linux has a powerful tool called make that allows you to efficiently manage a complex project and compile it efficiently. The
make utility reads a file called a makefile that describes the files involved in the project and the dependencies between them.
Each line in a makefile is of the form:
targets: dependencies
<Tab> commands
Here, targets is a list of target files separated by spaces and dependencies is a list of files on which the targets
depend. In other words, if any of the files in a dependency list is modified, make should recompile and/or relink the target. The
commands are the commands used to recreate the targets. (Note that the <Tab> preceding commands really must be a tab
character and not several spaces.)
For example, consider the following simple makefile:

# A simple makefile

sort: sort.o list.o

g++ sort.o list.o -o sort

sort.o: sort.cc sort.h

g++ -g -c sort.cc

list.o: list.cc list.h

g++ -g -c list.cc

This file states that the executable sort is dependent on the object files sort.o and list.o . If either of these object files
change, the command g++ sort.o list.o -o sort should be used to recreate (link) sort from the object files. The
next two groups show how to create the object files. The first states that sort.o should be recreated (compiled using g++ ) if
sort.cc or sort.h changes. (The -c option to g++ tells the compiler to only compile, creating object files, and not to link
them together.) The third group does a similar thing for list.o . So, you can see that if, for example, list.h is modified,
list.o and then sort will need to be recreated but sort.o will not.
To use this makefile, it should be placed in the same directory as the source files and named Makefile . Then, execute
make sort
to compile and link everything.
Here is a more complicated example of a makefile:

# Makefile for compiling a network sender project

CC = g++
LD = g++

INCDIR = -I../
CFLAGS = -g -c $(INCDIR)
LDFLAGS =

all: sender

sender: sender.o packet.o ack.o

$(LD) $(LDFLAGS) sender.o packet.o ack.o -o sender

sender.o: sender.cc packet.cc ack.cc proj.h packet.h ack.h

$(CC) $(CFLAGS) sender.cc

packet.o: packet.cc ack.cc packet.h ack.h

$(CC) $(CFLAGS) packet.cc

ack.o: ack.cc ack.h

$(CC) $(CFLAGS) ack.cc

clean:
rm *.o

In this makefile, you see the use of macros like CC (the name of the compiler) and LD (the name of the linker). Macros can
sometimes simplify a makefile and make it easier to modify. For example, if we wanted to change the compiler we could
just change the definition of the CC macro instead of changing each and every command list. The INCDIR , CFLAGS , and
LDFLAGS macros contain an option giving additional directories in which to look for header files, compiler options, and linker
options, respectively.
The all target can be followed by the default target(s) you want to create if no target is specified on the command line. In this
case, if make is given on the command line by itself, the target sender will be created. The clean target at the end can be
used to remove unneeded object files from the directory.
12 Networking
Linux was designed from the start to be a net-centric operating system. For the most part, network resources can be used
in Linux seamlessly. Many times, you do not even realize you are using the network. We saw two examples of this earlier:
remotely stored passwords with NIS and accessing your home directory via NFS.
Linux is complete with all the standard network tools including Netscape Navigator, FTP, Telnet, etc.

12.1 World Wide Web

To use Netscape, just type netscape on the command line.
There is also a text-oriented browser called Lynx ( lynx ) that you might want to use on the rare occasions that you only have a
text interface to a computer.

12.2 Remote Login

You can remotely log in to any Linux computer on our network by using telnet . For example, if you currently using tigger
and you want to log in to pooh , open another terminal window and type:
telnet pooh
You will be presented with a login prompt and then you can use a shell on pooh as if you were actually seated in front of that
machine. This is especially useful if you want to do some text-oriented work from somewhere else on campus. To end your
telnet session, type exit .
There is one problem with telnet however. When you type your password in telnet , it is sent over the network as cleartext.
Theoretically, someone can “sniff” out your password by intercepting the data stream on the local network. Therefore, a better
way to remotely log in to a machine is by using ssh (secure shell). Ssh sets up a secure encrypted connection between two
machines. You can use ssh basically the same way as telnet . For example, use
ssh tigger
to securely log in to tigger . You may see a message like the following:

The authenticity of host ’tigger (140.141.132.78)’ can’t be established.

DSA key fingerprint is 70:fa:15:6f:9f:da:2a:a8:b3:5d:61:8b:56:3a:c0:a9.
Are you sure you want to continue connecting (yes/no)?

Simply answer yes and you will be prompted for your password. To end your ssh session, type exit .

12.3 Remote X Windows

Beyond text-based remote login, X Windows actually allows you to see GUI applications that you are running on another
computer on the display in front of you. This is very useful if you are working on a machine (say, in your room) that does not
have a particular application installed, but it is installed on another computer to which you have access (say, in an Olin lab). It
may also be convenient to use this capability even if you have the application locally but you need to access a file in your home
directory which is only available from machines in the labs.
X Windows is an event-driven system, meaning that X Windows intercepts events like mouse clicks and key presses, and puts
them in an event queue associated with the corresponding application. The X Windows system that is running on the machine
in front of you (and intercepting events and displaying applications) is called an X Windows server . Application programs are
called clients . X Windows clients may be either local or remote . A local client is an application that is running on the same
machine as the server that is hosting it. A remote client is an application that is running on a different machine on the network,
but being displayed on the screen managed by the server. When an event is detected on the local machine that corresponds to a
remote client’s window, that event is sent to the remote client over the network. The remote client then reacts to the event (by
redrawing its window, for example) and this reaction is re ected on the local screen.
To start up a remote client application, you will first need to give permission to applications on the remote machine to use this
server. To do this, issue the command

xhost +remote-host
in a local shell, where remote-host is the name of the remote host. Then log in to the remote host on which you want to
start the application. Once you are logged in, you have to specify that applications you start should be displayed on your local
machine instead of on the remote one. To do this, you need to set the DISPLAY environment variable:
DISPLAY=local-host:0
The value local-host is the name of the host running the X server (the one at which you are sitting). The :0 is the display
number. Now you are ready to execute a client application. For example, executing netscape now on the remote host will
display the Netscape application on your screen.

13 Customizing your Account with Shell Scripts

Besides a command interpreter, Linux shells also include a programming language of their own. Programs written in a shell
programming language are usually called shell scripts . Shell scripts are executed by typing their name on the command line, just
like binary executables. (A shell script must have executable permission to be executed.) When executed, the shell interprets
the commands in the shell script as if they were being typed on the command line. We will specifically discuss the bash
shell here. The shell scripts with which you are most likely to come into contact are called .bash profile , .bashrc , and
.bash logout , all located in your home directory. The first script is executed automatically by the shell when you log in, the
second is executed every time you start a new bash shell, and the third is executed when you log out. An in-depth discussion
of bash shell programming is well beyond the scope of this tutorial, but we will very brie y discuss some of the basics here
— just enough to get you familiar with these initialization files.
Here is an example of a simple .bash profile file:
 # User specific environment variables
PATH="$PATH:.:˜/bin:/bin:/usr/bin:/usr/local/bin:/sbin:/usr/sbin:/usr/X11R6/bin"
CDPATH="$CDPATH:.:˜/"
PS1="\u@\h> "
PRINTER="pr219"
USER=$(id -un)
HOSTNAME=$(/bin/hostname)

export PATH CDPATH PS1 USER MAIL PRINTER HOSTNAME

umask 077

# Get the aliases and functions from .bashrc

if [ -f ˜/.bashrc ]; then
. ˜/.bashrc
fi

The first line in the file above is a comment, which is ignored by the shell. The next seven lines assign values to a variety of
environment variables . There are two types of variables in bash : environment variables and user-defined variables . The values
of environment variables are used by the shell and programs executed from the shell to customize their execution. User-defined
variables are used as temporary locations for data during a shell script. You can enter the command
printenv
to see the values of all currently set environment variables. The value of an environment variable can be set by using an
assignment statement like those above. For example, the statement
PS1="\u@\h> "
sets your shell prompt to the form user@host> . The special character \u stands for your user name and the special character
\h stands for the machine’s host name. Similarly, in the statement
PS1="\w$ "
the \w is a special character that stands for the current working directory. So this command changes your prompt to the name
of the working directory followed by a $ and a space.
The PATH variable contains a list of directories (separated by colons) in which the shell will look for programs whose names
you type on the command line. If you need to execute a common program by typing its absolute pathname on the command
line, add the directory to this list and you can simply type the program name from now on. Notice the mention of $PATH at
the beginning of the list. When a variable name is prefixed with $ , the value of the variable is returned. So, in this case, we are
appending the list of directories to the current value of PATH . The CDPATH variable contains a list of directories in which the
shell looks for relative pathnames specified with the cd command. The PRINTER variable contains the name of your default
printer. The USER variable contains your user name. The syntax on the right hand side of the assignment is for command
substitution . This means that the command inside the $(command) construct is executed and the result is assigned to the
variable. In this case, the id -un command returns your user name. The HOSTNAME variable contains the host name of the
machine. The /bin/hostname command gets this information and the result is stored in the variable.
After the variable assignments is the export command. This command causes the variables that follow to be copied by the
shell to every program executed from the shell. For example, if we did not export PRINTER here, the default printer name
would not be known to the lpr command later.
The umask command specifies the default access permissions for new files that you create. The bit mask argument of umask
specifies which permissions should be turned off when a new file is created. The default permissions are obtained by subtracting
the argument of umask from 666 for files and 777 for directories. In this case, the default permissions are set to 600
( rw------- ) for files and 700 ( rwx------ ) for directories.
The last part of the file is an example of a conditional construct. This particular statement tests to see if the file ˜/.bashrc
exists and if so, executes it. See one of the references in the bibliography for more information about shell programming if you
are curious about how this works.
In your .bashrc file, you will find a number of alias commands. Aliases are shorthand for commonly used commands.
For example, the command

alias rm="rm -i"

makes rm (the remove file command) an alias for rm -i , which means that every time you delete a file, you are asked whether
you are sure you want to delete it. (The -i option is short for “interactive”. Without this option, rm would delete a file without
asking first.)
If you want to print the value of a variable (or anything else) on the screen, you can use the simple echo command. For
example,
echo Hello World!
prints Hello World! in the shell. To print the value of a variable, prefix the variable name with $ . For example,
echo $PS1
prints the value of the PS1 variable in the shell.
You are encouraged to make changes to these configuration files as you please. After you make a change, execute
source ˜/.bash profile
(or substitute another shell script) to execute the script and institute the changes.

14 How to Learn More

We encourage you to play around with these systems so that you begin to feel more comfortable with them. Linux provides
a very powerful environment in which to work (and play!). There is tons more to learn, if you are so inclined. The menu in
the Panel at the bottom of the screen is a good starting point. Also refer to the command reference in Appendix B and the
bibliography below for information on more applications and commands.
Appendix: Linux Command Reference
The following is a (brief) summary of many of the commands available in Linux. Look at the man pages to learn more about
them. In what follows, text parameters enclosed in square brackets ( [ ] ) are optional.

Special Symbols
. the current directory
.. the parent directory
˜ your home directory
/ separator between directories in a pathname
* wildcard in a filename that matches any sequence of characters
? wildcard in a filename that matches any single character
& run a process in the background
Example: emacs & runs emacs in the background, freeing up the terminal.

Account Management
yppasswd Change your password on the NIS server.
ypchfn Change your full name on the NIS server.
ypchsh Change your default shell on the NIS server.

Applications
abiword a word processor
acroread Adobe Acrobat Reader
emacs , xemacs a very well endowed text editor.
exmh an electronic mail client
gedit a simple text editor
gimp the GNU image manipulation program (similar to PhotoShop)
glimmer a code-oriented text editor
gnp the gnotepad+ text editor
gnumeric a spreadsheet program
gv file display postscript file file
kedit , kwrite KDE text editors
kword , kspread ,
kpresenter ,
killustrator the KOffice office application suite
latex a typesetting program used by mathematicians and scientists (see references below)
nedit a text editor
netscape the web browser (and e-mail client, etc.)
soffice StarOffice is an office application suite, complete with a word processor, spreadsheet, HTML editor,
etc. It is Microsoft Office compatible.
To use StarOffice, you must first run usr/local/office52/program/setup from your
home directory. This will install about 2 MB of stuff in a directory called office52 . After
installation, you can run StarOffice by executing soffice in the office52 subdirectory.

Disk and File System Manipulation

cd [dir] Change the current working directory to dir .
cp [-f][-i][-R] file1 file2 Copy file1 to file or directory file2 .
Options: -f force copy without asking
-i interactive — ask before copying
-R recursively copy directory and its contents
Examples: cp sort.cc sort/sort.cc copies sort.cc from the current directory to the subdirectory sort .
cp /home/173/example.cc . copies example.cc from /home/173/ to the current directory.
df Display disk space available on each mounted volume.
du [directories] Display the space used by files in the current directory or directories if
specified.
fdformat /dev/fd0 Format a oppy disk. /dev/fd0 is the name of the oppy disk drive.
find dir -name file -print Find and print locations of file on the disk by descending the file system
from directory dir .
Examples: find . -name foo.bar -print prints location of foo.bar , starting from current directory.
find /home -name ’*.txt’ -print prints location of files ending in .txt , starting from /home .
ln -s source [dest] Make a symbolic link (alias) for the file source in the current directory or
name the link dest .
ls [-a][-F][-l][-R] [files] List contents of current directory or files , which may be files or directories.
Options: -a show all files, even those prefaced with a .
-F follow directory names with a /
-l show file sizes, dates, and permissions
-R list subdirectories recursively
mkdir dirs Create directory(ies) with name(s) dirs .
mtools Tools to read DOS disks ( mcopy , mdir , etc.)
mv [-f][-i] file1 file2 Move file1 to file or directory file2
Options: -f force move without asking
-i interactive — ask before moving
Examples: mv sort/sort.cc . moves sort.cc from subdirectory sort to the current directory.
mv sort.cc sorting.cc renames sort.cc to sorting.cc .
pwd Print the name of the current working directory.
rm [-R][-f] files Delete (remove) files .
Options: -R remove files and directories recursively
-f force delete – do not ask before deleting
rmdir [dir] Delete (remove) directory dir . The directory must be empty.
File Manipulation
cat [files] Print files to standard output. Reads from standard input by default.
Examples: cat file.txt displays file.txt .
cat > file.txt creates file.txt from standard input.
cat f1.txt f2.txt > f3.txt writes concatenation of f1.txt and f2.txt to f3.txt .
diff file1 file2 Display the differences between file1 and file2 . The output prefaces lines
from file1 with < and lines from file2 with > .
expand [-t num] files Convert the tabs in files to spaces. Specify the number of spaces with the -t
option.
grep [-i][-n] regexp files Find instances of regular expression regexp in files .
Options: -i ignore case
-n display line numbers with occurrences
gzip [-r][-v] [files] Compress files (or standard input, by default) using Lempel-Ziv encoding.
Options: -r descend recursively into directories
-v verbose output
gunzip [-r][-v] [files] Decompress files (or standard input by default) compressed with gzip .
head [-n num] [files] Display first num (10 by default) lines of files .
more [files] Display files one page at a time. During the display, type spacebar to see next
page, return to show next line, = to show line number, /pattern to search for
pattern , or q to quit.
tail [-n num] [files] Display last num (10 by default) lines of files .
tar cf tarfile files Create an archive called tarfile containing files .
tar xf tarfile Extract files from an archive called tarfile .

File Permissions
chmod mode files Set access mode (permissions) for files . There are two ways to specify permissions for a file.
1. First, use some combination of u for user, g for group, o for other, or a for all to specify
which of these group(s) of users’ permissions to modify. Then use + to add a permission or -
to take away a permission. Lastly, specify the permission(s) to add/subtract: r for read, w for
write, or x for execute.
Example: chmod go-rw * takes away read and write permission from group and other for all
files in the current directory.
2. Think of the rwx for a group ( u , g , or o ) as a representation of a 3 bit binary number (0 – 7
in decimal). If a permission is set, that bit is a 1; otherwise it is a 0. Use the 3 decimal digits for
the 3 groups as the permission.
Example: chmod 744 * gives all permissions to user, and only read permission to group and
other.
Help
apropos [-w wildcard] keyword Search manual pages for keyword .
info [topic] Display help on command .
man [section] [topic] Display manual page on command in section .

Miscellaneous
cal [[month] year] Display a calendar for a month or year. By default, display the current month.
clear Clear the terminal screen.
date Display current date and time.
xclock a graphical analog clock

Networking
finger name@host Lookup a user by name on machine host .
ftp host Initiate file transfer protocol with remote machine host .
host [host] Convert host name to IP address, or vice versa, using default name server.
hostname [-d] Display host name of this machine. The -d option displays domain instead.
ncpmount -S server -U user dir Mount NetWare server server on directory dir as user user .
ncpumount dir Unmount the NetWare server mounted on dir .
ssh [host] Remotely log in to the machine host securely.
telnet [host] Remotely log in to the machine host .
rusers Query users logged onto systems on local network.

Printing
lpq [-Pprinter] Display contents of default printer queue or queue of printer .
lpr [-Pprinter] [files] Print files to default printer or printer .
lprm [-Pprinter] [job nums] Remove jobs with numbers nums from a printer queue.

Process Management
kill [-9] pid End process with process id pid . The -9 option kills the process unconditionally.
nice command Execute command with lower than normal priority (to be “nice”).
ps [aux] Display information about currently executing processes. The aux options show a lot of
information about all processes in the system.
top Display the top processes with respect to CPU utilization. Hit q to quit.
Programming
ddd Data Display Debugger (you must compile with the -g option to use the debugger on your program)
gcc the GNU C compiler
g++ the GNU C++ compiler
gdb the GNU command line debugger
xxgdb X front end to GDB

System Resource Information

free Show free memory and swap space.
uptime Show amount of time passed since machine was last booted.
who Show who is logged into the local host.
whoami Show your login name.