About this Document What is NetBSD? Changes Between The NetBSD 8 and 9 Releases Installation and Partitioning Changes Features to be removed in a later release The NetBSD Foundation Sources of NetBSD NetBSD 9.4_STABLE Release Contents NetBSD/rs6000 subdirectory structure Binary distribution sets NetBSD/rs6000 System Requirements and Supported Devices Getting the NetBSD System on to Useful Media Preparing your System for NetBSD installation Installing the NetBSD System Installing NetBSD by using a netboot setup Post installation steps Upgrading a previously-installed NetBSD System Compatibility Issues With Previous NetBSD Releases Using online NetBSD documentation Administrivia Thanks go to Legal Mumbo-Jumbo The End
This document describes the installation procedure for
NetBSD
9.4_STABLE on the
rs6000
platform.
It is available in four different formats titled
INSTALL.ext,
where
.ext
is one of
.ps, .html, .more,
or .txt:
.ps.html.moremore(1)
and
less(1)
pager utility programs.
This is the format in which the on-line
man
pages are generally presented.
.txtYou are reading the HTML version.
The NetBSD Operating System is a fully functional Open Source UNIX-like operating system derived from the University of California, Berkeley Networking Release 2 (Net/2), 4.4BSD-Lite, and 4.4BSD-Lite2 sources. NetBSD runs on many different different system architectures (ports) across a variety of distinct CPU families, and is being ported to more. The NetBSD 9.4_STABLE release contains complete binary releases for most of these system architectures, with preliminary support for the others included in source form. Please see the NetBSD website for information on them.
NetBSD is a completely integrated system. In addition to its highly portable, high performance kernel, NetBSD features a complete set of user utilities, compilers for several languages, the X Window System, firewall software and numerous other tools, all accompanied by full source code.
NetBSD is a creation of the members of the Internet community. Without the unique cooperation and coordination the net makes possible, NetBSD would not exist.
The NetBSD 9.4_STABLE release provides many significant changes, including support for many new devices, hundreds of bug fixes, new and updated kernel subsystems, and numerous userland enhancements. The result of these improvements is a stable operating system fit for production use that rivals most commercially available systems.
It is impossible to completely summarize the massive development that
went into the
NetBSD
9.4_STABLE release.
The complete list of changes can be found in the following files:
CHANGES
CHANGES-9.1
CHANGES-9.2
CHANGES-9.3
CHANGES-9.4
CHANGES-9.5
files in the top level directory of the NetBSD 9.4 release tree.
It now supports arbitrary big disks and offers GPT partitions as alternative to MBR/fdisk partitions on a lot architectures.
Unfortunately it has not been tested on all hardware supported by NetBSD. If you have problems partitioning the target disk or installing the system, please report bugs with as much details as possible. See the Administrivia section below on how to report bugs or contact other users and ask for support.
groff(1).
Man pages are now handled with
mandoc(1),
and
groff(1)
can still be found in pkgsrc as
textproc/groff.
The
NetBSD
Foundation is a tax exempt, not-for-profit 501(c)(3) corporation
that devotes itself to the traditional goals and Spirit of the
NetBSD
Project and owns the trademark of the word
``NetBSD''.
It supports the design, development, and adoption of
NetBSD
worldwide.
More information on the
NetBSD
Foundation, its composition, aims, and work can be found at:
https://www.NetBSD.org/foundation/
Refer to
mirrors
The root directory of the NetBSD 9.4_STABLE release is organized as follows:
.../NetBSD-9.4_STABLE/
CHANGESCHANGES-9.0CHANGES-9.1CHANGES-9.2CHANGES-9.3CHANGES-9.4CHANGES-9.5CHANGES.prevREADME.filesimages/source/In addition to the files and directories listed above, there is one directory per architecture, for each of the architectures for which NetBSD 9.4_STABLE has a binary distribution.
The source distribution sets can be found in subdirectories of the
source
subdirectory of the distribution tree.
They contain the complete sources to the system.
The source distribution sets are as follows:
config(1)
utility.
All the above source sets are located in the
source/sets
subdirectory of the distribution tree.
The source sets are distributed as compressed tar files.
Except for the
pkgsrc
set, which is traditionally unpacked into
/usr/pkgsrc,
all sets may be unpacked into
/usr/src
with the command:
# cd / ; tar -zxpf set_name.tgz
In each of the source distribution set directories, there are files which contain the checksums of the files in the directory:
MD5SHA512The SHA512 digest is safer, but MD5 checksums are provided so that a wider range of operating systems can check the integrity of the release files.
rs6000
subdirectory of the distribution:
.../NetBSD-9.4_STABLE/rs6000/.
It contains the following files and directories:
INSTALL.htmlINSTALL.psINSTALL.txtINSTALL.more.more
file contains underlined text using the
more(1)
conventions for indicating italic and bold display.
binary/kernel/netbsd-GENERIC.gzsets/installation/misc/rs6000/binary/sets
subdirectory
of the
NetBSD
9.4_STABLE
distribution tree, and are as follows:
/usr/include)
and the various system libraries (except the shared
libraries, which are included as part of the
base
set).
This set also includes the manual pages for
all of the utilities it contains, as well as the
system call and library manual pages.
/etc
and in several other places.
This set
must
be installed if you are installing the system from scratch, but should
not
be used if you are upgrading.
GENERIC
kernel, named
/netbsd.
You
must
install this distribution set.
/usr/share.
/rescue.
groff(1),
all related programs, and their manual pages.
NetBSD maintains its own set of sources for the X Window System in order to assure tight integration and compatibility. These sources are based on X.Org. Binary sets for the X Window System are distributed with NetBSD. The sets are:
The rs6000 binary distribution sets are distributed as gzipped tar files
named with the extension
.tgz,
e.g.
base.tgz.
The instructions given for extracting the source sets work equally
well for the binary sets, but it is worth noting that if you use that
method, the filenames stored in the sets are relative and therefore
the files are extracted
below the current directory.
Therefore, if you want to extract the binaries into your system, i.e.
replace the system binaries with them, you have to run the
tar -xzpf
command from the root directory (
/
) of your system.
Since
NetBSD/rs6000
is an experimental port, the hardware support is in a constant state of flux.
Check the
port-prep@NetBSD.org
mailing list and the
NetBSD/rs6000
port homepage for an up-to-date list of known supported hardware.
http://www.NetBSD.org/ports/rs6000/
NetBSD/rs6000
9.4_STABLE works diskless via netboot.
Installation is supported from several media types, including:
The steps necessary to prepare the distribution sets for installation depend upon which installation medium you choose. The steps for the various media are outlined below.
binary/sets
and
rs6000/binary/sets.
(You only need to know this if you are mixing installer and installation
media from different versions - the installer will know the proper
default location for the sets it comes with).
Proceed to the instructions on installation.
Once you have this information, you can proceed to the next step in the installation or upgrade process. If you're installing NetBSD from scratch, go to the section on preparing your hard disk, below. If you're upgrading an existing installation, go directly to the section on upgrading.
/etc/exports
file on the NFS server and resetting its mount daemon (mountd).
(Both of these actions will probably require superuser
privileges on the server.)
You need to know the numeric IP address of the NFS server, and, if you don't have DHCP available on your network and the server is not on a network directly connected to the machine on which you're installing or upgrading NetBSD, you need to know the numeric IP address of the router closest to the NetBSD machine. Finally, you need to know the numeric IP address of the NetBSD machine itself.
Once the NFS server is set up properly and you have the information mentioned above, you can proceed to the next step in the installation or upgrade process. If you're installing NetBSD from scratch, go to the section on preparing your hard disk, below. If you're upgrading an existing installation, go directly to the section on upgrading.
If you're making the tape on a UNIX-like system, the easiest way to do so is probably something like:
# tar -cf tape_device dist_sets
where
tape_device
is the name of the tape device that
represents the tape drive you're using.
This might be
/dev/rst0,
or something similar, but it will vary from system to system.
In the above example,
dist_sets
is a list of filenames corresponding to the distribution sets that you
wish to place on the tape.
For instance, to put the
kern-GENERIC, base, and etc
distributions on tape (the absolute minimum required for installation),
you would do the following:
# cd .../NetBSD-9.4_STABLE
# cd rs6000/binary
# tar -cf tape_device kern-GENERIC.tgz base.tgz etc.tgz
Once you have the files on the tape, you can proceed to the next step in the installation or upgrade process. If you're installing NetBSD from scratch, go to the section on preparing your hard disk, below. If you're upgrading an existing installation, go directly to the section on upgrading.
To netboot a rs6000, you must configure one or more servers to provide
information and files to your rs6000 (the
`client').
If you are using
NetBSD
(any architecture) on your netboot server(s), the information
provided here should be sufficient to configure everything.
Additionally, you may wish to look at the
diskless(8)
manual page and the manual pages for each daemon you'll be configuring.
If the server(s) are another operating system, you should consult the
NetBSD Diskless HOW-TO, which will walk you through the steps necessary to
configure the netboot services on a variety of platforms:
https://www.NetBSD.org/docs/network/netboot/
You may either netboot the installer so you can install onto a locally attached disk, or you may run your system entirely over the network.
Briefly, the netboot
process involves discovery, bootstrap, kernel and file system stages.
In the first stage, the client discovers information
about where to find the bootstrap program.
Next, it downloads and executes the bootstrap program.
The bootstrap program goes through another discovery phase to determine
where the kernel is located.
The bootstrap program tries to mount the NFS share containing the kernel.
Once the kernel is loaded, it starts executing.
For RAM disk kernels, it mounts the RAM disk file system and begins
executing the installer from the RAM disk.
For normal (non-RAM disk) kernels, the
kernel tries to mount the NFS share that had the kernel and starts
executing
the installation tools or
init(8).
TFTP is used in the bootstrap phase to download
NFS is used in both the kernel and file system stages to download the
kernel, and to access files on the file server.
We will use
`CC:CC:CC:CC:CC:CC'
as the MAC address (ethernet hardware address) of your netboot client
machine.
You should have determined this address in an earlier stage.
In this example, we will use
`192.168.1.10'
as the IP address of your client and
`client.test.net'
as its name.
We will assume you're providing all of your netboot services
on one machine called
`server.test.net'
with the client's files exported from the directory
/export/client/root.
You should, of course, replace all of these with the names, addresses,
and paths appropriate to your environment.
You should set up each netboot stage in order (i.e., discovery, bootstrap, kernel, and then file system) so that you can test them as you proceed.
tftpd(8)
The default configuration of the TFTP server is to run in a
chroot(8)
environment in the
/tftpboot
directory.
Thus, the first order of business is to create this directory:
# mkdir -p /tftpboot
Next, edit
/etc/inetd.conf
and uncomment the line with the TFTP daemon:
tftp dgram udp wait root /usr/libexec/tftpd tftpd -l -s /tftpboot
Now, restart
inetd(8).
If the server is running
NetBSD,
you can achieve this with:
# /etc/rc.d/inetd restart
Now, you need to copy the
bootloader
for your rs6000 machine to
/tftpboot.
Just to be sure, let's make everything readable.
# chmod -R a+rX /tftpboot
Sometimes, the
arp(8)
table gets messed up, and the TFTP server can't communicate with the
client.
In this case, it will write a log message (via
syslogd(8))
to
/var/log/messages
saying:
`tftpd: write: Host is down'.
If this is the case, you may need to force the server to map your client's
ethernet address to its IP address:
# arp -s client CC:CC:CC:CC:CC:CC
nfsd(8),
mountd(8),
and
rpcbind(8)
Now your system should be able to load the bootstrap program and start looking for the kernel. Let's set up the NFS server. Create the directory you are exporting for the netboot client:
# mkdir -p /export/client/root
Put the following line in
/etc/exports
to enable NFS sharing:
/export/client/root -maproot=root client.test.net
If your server is currently running an NFS server, you only need to
restart
mountd(8).
Otherwise, you need to start
rpcbind(8)
and
nfsd(8).
If the server is running
NetBSD,
you can achieve this with:
# /etc/rc.d/rpcbind start
# /etc/rc.d/nfsd start
# /etc/rc.d/mountd restart
Now, if you place a kernel named
netbsd
in
/export/client/root
your client should boot the kernel.
You can skip this step if you do not plan to run your client diskless after installation. Otherwise, you need to extract and set up the client's installation of NetBSD. The Diskless HOW-TO describes how to provide better security and save space on the NFS server over the procedure listed here. See for details.
# cd /export/client/root
# tar -xpzf /path/to/files/base.tgz
# tar -xpzf /path/to/files/etc.tgz
Continue with the other non-essential distribution sets if desired.
# mkdir /export/client/root/swap
# dd if=/dev/zero of=/export/client/swap bs=4k count=4k
# echo '/export/client/swap -maproot=root:wheel client.test.net' >> /etc/exports
# /etc/rc.d/mountd restart
# cd /export/client/root/dev
# ./MAKEDEV all
This procedure only works on NetBSD hosts.
fstab(5)
Create a file in
/export/client/root/etc/fstab
with the following lines:
server:/export/client/swap none swap sw,nfsmntpt=/swap
server:/export/client/root / nfs rw 0 0
rc.conf(5)
Edit
/export/client/root/etc/rc.conf
rc_configured=YES
hostname="client"
defaultroute="192.168.1.1"
nfs_client=YES
auto_ifconfig=NO
net_interfaces=""
Make sure rc does not reconfigure the network device since it will lose its connection to the NFS server with your root file system.
hosts(5)
file.
Edit
/export/client/root/etc/hosts
::1 localhost
127.0.0.1 localhost
192.168.1.10 client.test.net client
192.168.1.5 server.test.net server
If you want
these services to start up every time you boot
your server, make sure the following lines are present in your
/etc/rc.conf:
nfs_server=YES # enable server daemons
mountd=YES
rpcbind=YES rpcbind_flags="-l" # -l logs libwrap
Also, you'll need to make sure the
tftpd
line in
/etc/inetd.conf
remains uncommented.
Once you've got the operating system running, there are a few things you need to do in order to bring the system into a properly configured state. The most important steps are described below.
postinstall(8).
/etc/rc.conf
If you or the installation software haven't done any configuration of
/etc/rc.conf
(sysinst
normally will),
the system will drop you into single user mode on first reboot with the
message
/etc/rc.conf is not configured. Multiuser boot aborted.
and with the root file system
(/)
mounted read-only.
When the system asks you to choose a shell, simply press
RETURN
to get to a
/bin/sh
prompt.
If you are asked for a terminal type, respond with
vt220
(or whatever is appropriate for your terminal type)
and press
RETURN.
You may need to type one of the following commands to get your delete key
to work properly, depending on your keyboard:
# stty erase '^h'
# stty erase '^?'
At this point, you need to configure at least
one file in the
/etc
directory.
You will need to mount your root file system read/write with:
# /sbin/mount -u -w /
Change to the
/etc
directory and take a look at the
/etc/rc.conf
file.
Modify it to your tastes, making sure that you set
rc_configured=YES
so that your changes will be enabled and a multi-user boot can
proceed.
Default values for the various programs can be found in
/etc/defaults/rc.conf,
where some in-line documentation may be found.
More complete documentation can be found in
rc.conf(5).
When you have finished editing
/etc/rc.conf,
type
exit
at the prompt to
leave the single-user shell and continue with the multi-user boot.
Other values that may need to be set in
/etc/rc.conf
for a networked environment are
hostname
and possibly
defaultroute.
You may also need to add an
ifconfig_int
for your
<int>
network interface,
along the lines of
ifconfig_le0="inet 192.0.2.123 netmask 255.255.255.0"
or, if you have
myname.my.dom
in
/etc/hosts:
ifconfig_le0="inet myname.my.dom netmask 255.255.255.0"
To enable proper hostname resolution, you will also want to add an
/etc/resolv.conf
file or (if you are feeling a little more adventurous) run
named(8).
See
resolv.conf(5)
or
named(8)
for more information.
Instead of manually configuring networking,
DHCP can be used by setting
dhcpcd=YES
in
/etc/rc.conf.
After reboot, you can log in as
root
at the login prompt.
If you didn't set a password in
sysinst,
there
is no initial password.
You should create an account for yourself (see below) and protect it and the
``root''
account with good passwords.
By default, root login from the network is disabled (even via
ssh(1)).
One way to become root over the network is to log in as a different
user that belongs to group
``wheel''
(see
group(5))
and use
su(1)
to become root.
Use the
useradd(8)
command to add accounts to your system.
Do not
edit
/etc/passwd
directly! See
vipw(8)
and
pwd_mkdb(8)
if you want to edit the password database.
If you installed the X Window System, you may want to read the chapter about X in the NetBSD Guide:
If you wish to install any of the software freely available for UNIX-like systems you are strongly advised to first check the NetBSD package system, pkgsrc. pkgsrc automatically handles any changes necessary to make the software run on NetBSD. This includes the retrieval and installation of any other packages the software may depend upon.
rs6000/9.4_STABLE/All
subdir.
If you installed
pkgin(1)
in the
sysinst
post-installation configuration menu, you can use it to automatically install
binary packages over the network.
Assuming that
/usr/pkg/etc/pkgin/repositories.conf
is correctly configured, you can install them with the following commands:
# pkgin install tcsh # pkgin install bash # pkgin install perl # pkgin install apache # pkgin install kde # pkgin install firefox ...
/pub/pkgsrc
directory.
The above commands will install the Tenex-csh and Bourne Again shells, the Perl programming language, Apache web server, KDE desktop environment and the Firefox web browser as well as all the packages they depend on.
pkgsrc(7)
framework for compiling packages can be obtained by
retrieving the file
/usr/pkgsrc
(though other locations work fine) with the commands:
# cd /usr
# tar -zxpf pkgsrc.tar.gz
After extracting, see the
doc/pkgsrc.txt
file in the extraction directory (e.g.,
/usr/pkgsrc/doc/pkgsrc.txt)
for more information.
/etc/mail/aliases
to forward root mail to the right place.
Don't forget to run
newaliases(1)
afterwards.
/etc/rc.local
to run any local daemons you use.
/etc
files are documented in section 5 of the manual; so just invoking
# man 5 filename
is likely to give you more information on these files.
The easiest way to upgrade to NetBSD 9.4_STABLE is with binaries, and that is the method documented here.
To do the upgrade, you must have one form of boot media available. You must also have at least the base and kern binary distribution sets available. Finally, you must have sufficient disk space available to install the new binaries. Since files already installed on the system are overwritten in place, you only need additional free space for files which weren't previously installed or to account for growth of the sets between releases.
Since upgrading involves replacing the kernel, boot blocks, and most of the system binaries, it has the potential to cause data loss. You are strongly advised to back up any important data on the NetBSD partition or on another operating system's partition on your disk before beginning the upgrade process.
The upgrade procedure is similar to an installation, but without the hard disk partitioning.
Fetching the binary sets is done in the same manner as the installation procedure; refer to the installation part of the document for help. File systems are checked before unpacking the sets.
After a new kernel has been copied to your hard disk, your
machine is a complete
NetBSD
9.4_STABLE
system.
However, that doesn't mean that you're finished with the upgrade process.
You will probably want to update the set of device nodes you have in
/dev.
If you've changed the contents of
/dev
by hand, you will need to be careful about this, but if
not, you can just cd into
/dev,
and run the command:
# sh MAKEDEV all
sysinst
will attempt to merge the settings stored in your
/etc
directory with the new version of
NetBSD
using the
postinstall(8)
utility.
However,
postinstall(8)
is only able to deal with changes that are easily automated.
It is recommended that you use the
etcupdate(8)
tool to merge any remaining configuration changes.
Users upgrading from previous versions of NetBSD may wish to bear the following problems and compatibility issues in mind when upgrading to NetBSD 9.4_STABLE.
Note that sysinst will automatically invoke
postinstall fix
A number of things have been removed from the NetBSD 9.4_STABLE release. See the ``Components removed from NetBSD'' section near the beginning of this document for a list.
Documentation is available if you installed the manual
distribution set.
Traditionally, the
``man pages''
(documentation) are denoted by
`name(section)'.
Some examples of this are
intro(1),
man(1),
apropos(1),
passwd(1),
and
passwd(5).
The section numbers group the topics into several categories, but three are of primary interest: user commands are in section 1, file formats are in section 5, and administrative information is in section 8.
The man command is used to view the documentation on a topic, and is started by entering man [section] topic. The brackets [] around the section should not be entered, but rather indicate that the section is optional. If you don't ask for a particular section, the topic with the lowest numbered section name will be displayed. For instance, after logging in, enter
# man passwd
to read the documentation for
passwd(1).
To view the documentation for
passwd(5),
enter
# man 5 passwd
instead.
If you are unsure of what man page you are looking for, enter
# apropos subject-word
where subject-word is your topic of interest; a list of possibly related man pages will be displayed.
If you've got something to say, do so!
We'd like your input.
There are various mailing lists available via the mailing list
server at
majordomo@NetBSD.org.
See
https://www.NetBSD.org/mailinglists/
for details.
There are various mailing lists set up to deal with comments and questions about this release. Please send comments to: netbsd-comments@NetBSD.org.
To report bugs, use the
send-pr(1)
command shipped with
NetBSD,
and fill in as much information about the problem as you can.
Good bug reports include lots of details.
Bugs also can be submitted and queried with the web interface at
https://www.NetBSD.org/support/send-pr.html
There are also port-specific mailing lists, to discuss aspects of
each port of
NetBSD.
Use majordomo to find their addresses, or visit
https://www.NetBSD.org/mailinglists/
If you're interested in doing a serious amount of work on a specific port, you probably should contact the `owner' of that port (listed below).
If you'd like to help with NetBSD, and have an idea as to how you could be useful, send us mail or subscribe to: netbsd-users@NetBSD.org.
As a favor, please avoid mailing huge documents or files to these mailing lists. Instead, put the material you would have sent up for FTP or WWW somewhere, then mail the appropriate list about it. If you'd rather not do that, mail the list saying you'll send the data to those who want it.
Keith Bostic Ralph Campbell Mike Karels Marshall Kirk McKusick
for their work on BSD systems, support, and encouragement.
All product names mentioned herein are trademarks or registered trademarks of their respective owners.
The following notices are required to satisfy the license terms of the software that we have mentioned in this document:
NetBSD is a registered trademark of The NetBSD Foundation, Inc.
In the following statement, the phrase ``this text'' refers to portions
of the system documentation.
Portions of this text are reprinted and reproduced in electronic form in
NetBSD, from IEEE Std 1003.1, 2004 Edition, Standard for
Information Technology -- Portable Operating System Interface (POSIX),
The Open Group Base Specifications Issue 6, Copyright (C) 2001-2004 by the
Institute of Electrical and Electronics Engineers, Inc and The Open Group.
In the event of any discrepancy between these versions and the original
IEEE and The Open Group Standard, the original IEEE and The Open Group
Standard is the referee document.
The original Standard can be obtained online at
http://www.opengroup.org/unix/online.html.
This notice shall appear on any product containing this material.
In the following statement, "This software" refers to the parallel port driver:
Some files have the following copyright:
Permission to use, copy, modify and distribute this software and its
documentation is hereby granted, provided that both the copyright
notice and this permission notice appear in all copies of the
software, derivative works or modified versions, and any portions
thereof, and that both notices appear in supporting documentation.
CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS
CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
Carnegie Mellon requests users of this software to return to
any improvements or extensions that they make and grant Carnegie the
rights to redistribute these changes.
Some files have the following copyright:
Author: Chris G. Demetriou
Permission to use, copy, modify and distribute this software and
its documentation is hereby granted, provided that both the copyright
notice and this permission notice appear in all copies of the
software, derivative works or modified versions, and any portions
thereof, and that both notices appear in supporting documentation.
Carnegie Mellon requests users of this software to return to
any improvements or extensions that they make and grant Carnegie the
rights to redistribute these changes.
Some files have the following copyright:
Permission to use, copy, modify, and distribute this
software and its documentation for any purpose and without
fee is hereby granted, provided that the above copyright
notice appear in all copies. Stanford University
makes no representations about the suitability of this
software for any purpose. It is provided "as is"
without express or implied warranty.
Copyright (c) 1991,1990,1989 Carnegie Mellon University
All Rights Reserved.
Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU
School of Computer Science
Carnegie Mellon University
Pittsburgh PA 15213-3890
All rights reserved.
CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND
FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU
School of Computer Science
Carnegie Mellon University
Pittsburgh PA 15213-3890