The Hurd Installation Guide

  This document aims to provide an easy and relatively painless set of
  instructions on how to get the Hurd up and running with a minimum
  amount of effort.

    1. Overview                         Where we are going
    2. Real Estate or Finding A Home    Finding a Home
    3. The Boot Loader                  Getting Grub
    4. Cross Install                    Cross Installing the Hurd
    5. Booting the Hurd
    6. Native Install                   Finishing the Installation
    7. Configuration                    Making the System Usable
    8. Final Words                      The FAQ
    9. Works Cited                      Referenced Materials
    _________________________________________________________________


                                 1. Overview

  The Debian GNU Hurd distribution, unlike distributions of other
  operating systems, does not have a nice installation program. One day
  it will and maybe you will help design and implement it; however,
  until that day, installing the GNU Hurd requires another operating
  system, specifically, another Unix-like system(1). Users have
  indicated successful installations using different flavors of
  GNU/Linux as well as the BSDs. The minimum requirements of the
  bootstrap operating system are the ability: to create an ext2 file
  system; to extract a tar archive on to it; and to install GNU Grub.

  The GNU Hurd is similar in nature to any Unix-like system: after
  logging in, the user is presented with a shell and the familiar Unix
  VFS, virtual filesystem. Although GNU tries to be POSIX compliant, it
  is not Unix. The Hurd builds upon many of the Unix concepts and
  extends them to either add new functionality or to fix what has been
  perceived as flaws in the original design. The most noticeable
  difference is translators, user space programs which interact with the
  VFS. These filesystems do not live living in the kernel nor do they
  need to be run as root; they only need access to the backing store and
  the mount point. Another difference is that processes, rather having a
  single user identity fixed at creation time, have identity tokens
  which are disjoint from the process, i.e. they may be added with the
  appropriate permission from an authority or destroyed.

  Being familiar with the Unix environment is an imperative for feeling
  at ease in GNU. Having experience with the Debian tools will also
  prove invaluable to the configuration and maintenance of a GNU/Hurd
  box.

  This guide endeavors to make installing the Hurd as painless a process
  as possible. If there are errors, they are most certainly the
  author's. Please report them, along with any other suggestions or
  criticisms, to him; all are gladly accepted.
    _________________________________________________________________


                      2. Real Estate or Finding A Home

  If you do not have an available partition or an extra hard drive, this
  can be the longest step. In this case, you will need to repartition
  the hard drive. One solution is to use GNU's partition editor, Parted.
  It features not only basic partition editing but also partition
  resizing and moving functionality. It can be found at
  http://www.gnu.org/software/parted. The manual is quite complete and
  includes several tutorials.

  The Hurd can only support partition sizes of up to approximately two
  gigabytes; anything larger than this will not work. This limitation is
  due to a design decision that was made several years ago in which the
  filesystem server maps the entire filesystem into virtual memory. As
  the amount of virtual memory available on an ia32 is only four
  gigabytes of which Mach allocates three gigabytes to the application
  and, of that, a significant portion is reserved for the code, the
  stack and the heap, the final, maximum contiguous virtual memory area
  that remains is generally about two gigabytes. This limitation is
  scheduled to be removed.

  Having said that, a single two gigabyte filesystem is more than enough
  for a working system. Many, however, prefer at least two filesystems:
  a root filesystem and a second for `/home'. This latter scheme is
  highly advised for developers: compiling the Hurd can take up quite a
  bit of space.

  The Hurd supports several extensions to the ext2fs filesystem format.
  Foremost among these are passive translators and a fourth set of
  permission bits for unknown users (users without an identity--not the
  other user). To use these extensions, the owner of the partition must
  be set to hurd. mke2fs, unless specifically overridden on the command
  line, will set the owner to whatever operating system it is running
  on. As the Hurd will diligently respect this setting, care must be
  taken to set this appropriately or the Hurd will fail in subtle ways.
  Be aware that even if a file system is owned by a particular operating
  system, others may still use it; they just may not be able to use
  certain extensions.

  To create a filesystem, use mke2fs and pass it `-o hurd' to designate
  the Hurd as the owner of the new file system. For instance, assuming
  the parition is `/dev/hda2':


# mke2fs -o hurd /dev/hda2

                             3. The Boot Loader

  Unlike GNU/Linux and the BSDs, the Hurd does not have its own boot
  loader; any boot loader that supports the multiboot standard can be
  used to load the Hurd. At the moment, there is only one project which
  that satisfies these requirements: Grub, the GRand Unified Boot
  loader.

  A word about Grub. Unlike traditional boot loaders on the x86, such as
  LILO, Grub is very powerful. It has a command line interface, bootp,
  dummy terminal support and a plethora of other features. In addition,
  it can boot most any operating system. If you have ever booted an
  alpha or sparc, you will understand what Grub can do. Therefore, do
  not be scared: Grub is better. You will like it. You will not go back.

  To find Grub, visit http://www.gnu.org/software/grub/. Here, there is
  a source tarball and a floppy image. If you choose to download the
  tarball, it is a normal configure, make and make install. Included is
  a wonderfully complete manual on how Grub works. Read it. If, on the
  other hand, you choose to download the floppy image, it is sufficient
  to dump it to a floppy disk to get a working Grub, for example:


# dd if=grub-boot-image of=/dev/fd0

  You can always install Grub onto your hard drive at a later date.
    _________________________________________________________________

                              4. Cross Install

  The next step is to download the base system at:
  ftp://alpha.gnu.org/gnu/hurd/contrib/marcus/gnu-latest.tar.gz.

  The tarball is setup to extract everything into the current directory.
  After the filesystem is mounted, the archive can be extracted.
  Assuming that the filesystem is on `/dev/hda2', the mount point is
  `/gnu' and archive is in current user's home directory, the following
  is required:


# mount -t ext2 /dev/hda2 /gnu
# cd /gnu
# tar --same-owner -xvzpf ~/gnu-latest.tar.gz
    _________________________________________________________________

                             5. Booting the Hurd

  All is now in readiness to boot the Hurd for the first time. After
  verifying that the Grub boot disk is in the drive, reboot. If all goes
  well, either a Grub menu or command line will be displayed. If
  presented with a menu, press c to go to the command line.

  First, GNU Mach needs to be loaded. This requires knowing the
  filesystem and the path to GNU Mach. Grub uses a partition
  nomenclature that is a bit different from both Linux and the Hurd:
  both IDE and SCSI drives are named `(hdN,M)'. N is the drive number
  (zero based) as enumerated by the BIOS. That is, Grub makes no
  distinction between IDE and SCSI disks. M identifies the partition on
  the drive. It is also zero based index. If this sounds confusing,
  relax: Grub is also helpful.

  To determine on which filesystem a particular file resides, Grub
  provides the find command. When this command is issued along with a
  filename, Grub searches on each filesystem for the specified file and
  prints where it was found. For example, to search for the kernel,
  `/boot/gnumach.gz':


 grub> find /boot/gnumach.gz
  (hd0,0)

  Here, Grub is indicates that `/boot/gnumach.gz' is on `(hd0,0)'.

  Before loading the kernel, at least one option, the root partition,
  must be specified on the command line. This will be used by the Hurd
  itself (i.e. not Grub). As such, it must be in terms that the Hurd can
  understand.

  GNU Mach enumerates disks starting at zero. IDE drives are prefixed
  with hd, while SCSI disks are prefixed with sd. Like Linux, drives are
  number by their position on the controller. For instance, the primary
  master is hd0 and the secondary slave is hd3. Partitions use the BSD
  slice naming convention and append sM to the drive name to indicate a
  given partition. Note that M is a one, not zero, based index. The
  slice number is simple to calculate: just increment what was used for
  Grub by one.

  Since the Hurd has not yet been configured, it must be started in
  single user mode. Adding a `-s' to the kernel command line is all that
  is required.

  Assuming that the first drive (i.e. `(hd0)') is the master on the
  secondary controller, we would have:


 grub> kernel (hd0,0)/boot/gnumach.gz root=device:hd2s1 -s
  [Multiboot-elf, ...]

  Next, the root filesystem server and the exec server must be loaded.
  This is done using Grub's boot module capability. The ${var} are
  filled in by GNU Mach. The arguments are used by the Hurd to indicate
  what type of information is being provided. Since the ext2fs command
  line is very long, it can be broken up by escaping the newline
  character in the normal Unix fashion. Be sure that there is not space
  after the slash at the end of each line. Also be sure to differentiate
  { and } from ( and ).


 grub> module (hd0,0)/hurd/ext2fs.static \
  --multiboot-command-line=${kernel-command-line} \
  --host-priv-port=${host-port} \
  --device-master-port=${device-port} \
  --exec-server-task=${exec-task} -T typed ${root} \
  $(task-create) $(task-resume)
   [Multiboot-module  0x1c4000, 0x2cfe6a bytes]
 grub> module (hd0,0)/lib/ld.so.1 /hurd/exec $(exec-task=task-create)
   [Multiboot-module  0x494000, 0x27afe bytes]

  Once the GNU Hurd is running, process can be automated by adding the
  appropriate commands to Grub's `/boot/grub/menu.lst' configuration
  file.

  The GNU Hurd can be now booted:


grub> boot

  Sit back and watch the messages. This is actually more important than
  most people believe: there is a bug in GNU Mach whereby hitting a key
  during the boot process causes the kernel to panic.

  If the Hurd fails to boot, it could be due to shared IRQs: GNU Mach
  does not play well with these. You can verify your situation by
  looking at, for instance, the `/proc/interrupts' file under GNU/Linux.
  Also, as GNU Mach does not support loadable kernel modules, many of
  the drivers are compiled into the default kernel. If there are old
  peripherals, this can be a problem: a device may incorrectly respond
  to a probe intended for a completely unrelated device and thereby
  cause a crash. Building a new kernel with only the required device
  drivers will usually solve this problem. GNU Mach is easily cross
  compiled. If you are running Debian, try installing the `gcc-i386-gnu'
  package.

  If this does not help, explore the resources listed at the end of this
  document. Finally, ask on the appropriate mailing list.
    _________________________________________________________________

                              6. Native Install

  Once you are presented with a shell prompt, and any time that the the
  Hurd is in single user mode, it is necessary to set the terminal type:


# export TERM=mach

  Be warned that CONTROL-C and family will not work in single user mode.

  We can now run the native-install script. This will configure the
  packages and set up several important translators:


# ./native-install

  Before the script terminates, it will indicate that it needs to be run
  a second time. Follow its instructions and reboot using the reboot
  command. Again, go into single user mode and run ./native-install.
    _________________________________________________________________


                              7. Configuration

7.1 The Network

  To configure the network, the pfinet translator must be configured.
  This is done using the settrans command to attach a translator to a
  given filesystem node. When programs access the node by, for example
  sending an RPC, the operating system will transparently start the
  server to handle the request.


# settrans -fgap /servers/socket/2 /hurd/pfinet -i eth0 \
  -a a.b.c.d -g e.f.g.h -m i.j.k.l

  Here, settrans is passed several options. The first two, `fg', force
  any existing translator to go away. The next two, `ap', make both
  active and passive translators. By making the translator active, we
  will immediately see any error messages on `stderr'. The latter, saves
  the translator and arguments in the node so it can be transparently
  restarted later (i.e. making the setting persistent across reboots).
  The options are followed by the node to which the translator is to be
  attached, then the program (i.e. translator) to run and any arguments
  to give it. The `-i' option is the interface pfinet will listen on,
  `-a' is the ip address, `-g' is the gateway and `-m' is the network
  mask.

  Be sure to add name servers to your `/etc/resolv.conf' file:


 nameserver 192.168.1.1

  To test the configuration, ping -c2 gateway. The `-c' is important to
  limit the number of pings; recall, CONTROL-C does not work in single
  user mode.

  DHCP does not yet work on the Hurd. This is due to limitations of
  pfinet: it is based on the Linux' TCP/IP code and unable to listen on
  `0.0.0.0'.

  Help on settrans can be obtained by passing it the `--help' option.
  Help on a specific translator can be gotten by invoking it from the
  command line with the same argument, e.g.:


# /hurd/pfinet --help

  As there can be a lot of output, consider piping this through a pager
  such as less.
    _________________________________________________________________


7.2 Other File Systems

  Next, edit `/etc/fstab' to add any additional filesystems as well as
  swap space. It is very important that swap space be used; the Hurd
  will be an order of magnitude more stable. Note that the Hurd can
  transparently share a swap partition with Linux but will happily page
  to any device including a raw partition such as your home partition.
  By default, nano is the only editor installed by the the base
  distribution.

  Here is an example `/etc/fstab' file:


# <file system> <mount point>   <type>  <options>  <dump>  <pass>
/dev/hd2s1      /               ext2    rw         0       1
/dev/hd2s2      /home           ext2    rw         0       2
/dev/hd2s3      none            swap    sw         0       0

  Remember to create any devices using the MAKEDEV command:


# cd /dev
# ./MAKEDEV hd2s1 hd2s2 hd2s3

  To mount an nfs filesystem, /hurd/nfs translator is used. When run as
  non-root, the translator will connect to the server using a port above
  1023. By default, GNU/Linux will reject this. To tell GNU/Linux to
  accept connections originating from a non-reserved port, add the
  `insecure' option to the export line. Here is an example
  `/etc/exports' file assuming the client's ip address is `192.168.1.2':


 /home  192.168.1.2(rw,insecure)

  To mount this from a Hurd box and assuming that nfs server's ip
  address is `192.168.1.1':


# settrans -cgap /mount/point /hurd/nfs 192.168.1.1:/home
    _________________________________________________________________


7.3 Rebooting

  Finally, reboot into multiuser mode, i.e. in the same way single user
  mode was brought up minus the `-s' option when loading the kernel. For
  details, refer to See section 5. Booting the Hurd.

  Happy Hacking!
    _________________________________________________________________


                               8. Final Words

8.1 Documentation

  To understand the Hurd, start with the home page on Debian's site:
  http://www.debian.org/ports/hurd/ and GNU's site: http://hurd.gnu.org.

  Also consider reading the source code and writing more documentation.
    _________________________________________________________________


8.2 The Grub Menu

  Having to always load the kernel by hand can be very tedious. Edit the
  `/boot/grub/menu.lst' and tailor it appropriately; booting will become
  much quicker and easier.
    _________________________________________________________________


8.3 Adding Devices

  By default, only a few devices are created in the `/dev' directory.
  Use the MAKEDEV script to create any needed device nodes.
    _________________________________________________________________


8.4 Installing More Packages

  There are several ways to add packages. Downloading and using dpkg -i
  works but is very inconvenient. The easiest method is to use apt-get.
  Edit `/etc/apt/sources.list' and add the following two entries:


deb ftp://alpha.gnu.org/gnu/hurd/debian unstable main
deb ftp://ftp.debian.org/debian unstable main

  ftp://alpha.gnu.org contains packages that have hacks or patches that
  have not yet been integrated upstream. There are no mirror sites. To
  use a local Debian mirror, visit
  http://www.debian.org/distrib/ftplist.

  If GNU Mach does not recognize your network card or you use a modem,
  the only way to upgrade will be to download the packages and then
  transfer them to the GNU Hurd system. The easiest way to do this is to
  use apt off-line. Refer to `/usr/share/doc/apt/offline' for detailed
  instructions.
    _________________________________________________________________

8.5 XFree86

  XFree86 has been ported and all video cards, which it supports that do
  not require a kernel module should work.

  First, set up the keyboard translator:


# cd /dev
# ./MAKEDEV kbd

  And then the mouse translator. For a serial port mouse, run the
  following replace `com0' with the appropriate communication port:


# settrans /dev/mouse /hurd/mouse --device=com0 --protocol=microsoft

  Make sure that `/dev/com0' actually exists. If it does not, create it
  using MAKEDEV in the usual fashion.

  PS/2 so not require a device node. It is simple a matter of:


# settrans /dev/mouse /hurd/mouse --protocol=ps/2

  Other mice can be used; run `/hurd/mouse' with the `--help' option for
  details.

  You will need several X packages. x-window-system-core, rxvt and twm
  or fvwm are a good start.

  Debconf can be used to configure XFree86, however, it is not Hurd
  aware and the configuration file will need to be tweaked. Change the
  pointer section to read:


Section "Pointer"
 Protocol "osmouse"
 Device "/dev/mouse"
EndSection

  `Emulate3Buttons' may be optionally added. Nothing else will work.

  The GNU Hurd does not use ld.so.conf. Since `/X11R6/lib' is not in the
  default library search path, it is necessary to add the following to
  either `/etc/profile' or each user's `.profile':


export LD_LIBRARY_PATH=/X11R6/lib:$LD_LIBRARY_PATH

  Finally, run startx.

  There are several caveats to be aware of. xterm does not work
  correctly as it is SETGID (and thus ignores LD_LIBRARY_PATH and fails
  to load the appropriate shared libraries); try rxvt. update-menu has
  not yet been ported. As such, there are no fine Debian menus. Although
  a pthreads implementation exists not all pthread packages have been
  ported: do not expect Gnome and KDE to work.
    _________________________________________________________________


8.6 Virtual Terminals

  The Hurd does not have virtual terminals although it is in
  development. Use the screen package in the interim.
    _________________________________________________________________


8.7 Mailing Lists

   1. [email protected] Discussion on the Hurd as it relates to
      Debian.
        1. Archive: http://lists.debian.org/#debian-hurd
   2. [email protected] Development of the Hurd web pages at
      http://hurd.gnu.org.
        1. Archive: http://mail.gnu.org/pipermail/web-hurd/
   3. [email protected] Help on the Hurd in general.
        1. Archive: http://mail.gnu.org/pipermail/help-hurd/
   4. [email protected] Bug reports and general development. Send patches
      here.
        1. Archive: http://mail.gnu.org/pipermail/bug-hurd/
    _________________________________________________________________


8.8 Other Resources

  The Hurd Wiki, http://hurd.gnufans.org/bin/view/Hurd/WebHome, answers
  commonly addressed problems that new users face.
    _________________________________________________________________


                               9. Works Cited

  "The Easy Guide to Installing Hurd on a Linux Box" Copyright (C) 1999
  Matthew Vernon [email protected].
  http://www.pick.ucam.org/~mcv21/hurd.html
    _________________________________________________________________


                                  Footnotes

 (1)

  Philip Charles has created a set of CDs (available at
  http://www.debian.org/ports/hurd/hurd-cd) which contains a live Debian
  GNU/Linux system thereby arguably eliding this requirement, however,
  we maintain that this is functionally equivalent.
    _________________________________________________________________


                              Table of Contents

  1. Overview
      2. Real Estate or Finding A Home
      3. The Boot Loader
      4. Cross Install
      5. Booting the Hurd
      6. Native Install
      7. Configuration

  7.1 The Network
      7.2 Other File Systems
      7.3 Rebooting

  8. Final Words
  8.1 Documentation
      8.2 The Grub Menu
      8.3 Adding Devices
      8.4 Installing More Packages
      8.5 XFree86
      8.6 Virtual Terminals
      8.7 Mailing Lists
      8.8 Other Resources

  9. Works Cited
    _________________________________________________________________

                           Short Table of Contents

    1. Overview
    2. Real Estate or Finding A Home
    3. The Boot Loader
    4. Cross Install
    5. Booting the Hurd
    6. Native Install
    7. Configuration
    8. Final Words
    9. Works Cited
    _________________________________________________________________

                             About this document

  This document was generated by Neal H. Walfield on January, 21 2003
  using texi2html

    _________________________________________________________________

  This document was generated by Neal H. Walfield on January, 21 2003
  using texi2html