The Unofficial (and no longer maintained) GNU Hurd FAQ, Version 0.13
Contributions by:
Michael I. Bushnell
Len Tower
Trent Fisher
jlr@usoft.spb.su
Remy Card
Louis-Dominique Dubeau
Original Document by: Derek Upham
==============================
Contents:
Q0. Where can I get the Unofficial GNU Hurd FAQ?
Q2. Where can I get a copy?
Q3. Why bother writing a new OS when we have Linux and 386/BSD?
Q4. What's all this about Mach 3.0 (and Mach 4.0)?
Q5. Where can I find more information?
Q6. What's a proper machine?
Q7. What sort of machines will run Hurd in the future?
Q8. What is the current development status?
Q9. What sort of system would we have if the Hurd was bootable today?
==============================
Q0. Where can I get the Unofficial GNU Hurd FAQ?
The Unofficial Hurd FAQ (what you are reading now) is occasionally
posted to the USENET newsgroup, gnu.misc.discuss. It is also
available from
http://www.enci.ucalgary.ca/~gord/hurd/hurd-faq.txt
If you don't have WWW access, you may send mail to me, Gordon
Matzigkeit with a subject line that reads:
Subject: send hurd-faq
You should receive a PGP-signed copy of the current version of this
document in a matter of minutes.
Q2. Where can I get a copy?
To put it simply, you can't. It is still under development (by
Michael Bushnell, Roland McGrath and Miles Bader). It is almost, but
not quite, at the point where you can do real work on it. Keep your
fingers crossed.
Some people have actually bootstrapped it, but the work is not easy,
and the current snapshot won't work until a new multiserver boot
mechanism is made.
If you *really* want to try it, beware that it is still pre-alpha
code, and that it will likely crash on you. See Trent Fisher's Hurd
pages (under question 5) for the latest information.
Q3. Why bother writing a new OS when we have Linux and 386/BSD?
For one thing, Linux and BSD don't scale well. Hardware designers are
shifting more and more toward multiprocessor machines for performance,
and standard Unix kernels do not provide much multiprocessor support.
The Hurd, on the other hand, runs on top of the Mach 3.0 micro-kernel
[[1]] from CMU. Mach was designed precisely for multiprocessing
machines, so its portability should carry over nicely to the Hurd.
In addition, the Hurd will be considerably more flexible and robust
than generic Unix. Wherever possible, Unix kernel features have been
moved into unprivileged space. Once there, anyone who desires can
develop custom replacements for them. Users will be able to write and
use their own file systems, their own `exec' servers, or their own
network protocols if they like, all without disturbing other users.
The Linux kernel has now been modified to allow user-level file
systems, so there is proof that people will actually use features such
as these. It will be much easier to do under the Hurd, however,
because the Hurd is almost entirely run in user space and because the
various servers are designed for this sort of modification.
Q4. What's all this about Mach 3.0 (and Mach 4.0)?
As mentioned above, Mach is a micro-kernel, written at Carnegie Mellon
University. A more descriptive term might be a greatest-common-factor
kernel, since it provides facilities common to all ``real'' operating
systems, such as memory management, interprocess communication,
processes, and a bunch of other stuff. Unfortunately, the system
calls used to access these facilities are only vaguely related to the
familiar and cherished Unix system calls. There are no "fork",
"wait", or "sleep" system-calls, no SIGHUPs, nothing like that. All
this makes it rather difficult to, say, port GNU Emacs to a Mach box.
The trick is, of course, to write an emulation library. Unix programs
can then use (what they think are) POSIX system calls and facilities
while they are really using Mach system calls and facilities.
The simplest way of going about this is to take an ordinary Unix
kernel, open it up, and rip out all the machine-specific guts; any
time the Unix kernel talks to the machine, replace the code with calls
to the Mach micro-kernel. Run this fake kernel on a Mach machine and
you end up with something that looks and acts just like Unix (even to
GNU Emacs). Note that the Unix kernel we have implemented is just one
Really Big Mach program (called a single-server).
The Hurd, on the other hand, breaks the giant Unix kernel down into
various Mach programs running as daemons. Working in concert with
facilities placed in the C library, these daemons provide all of the
POSIX system-calls and features; from the outside they look just like
a standard Unix kernel. This means that, for practical purposes,
anything that you can port to Linux will also port to the Hurd.
Of course, if a user wishes to run his own daemons, he can do that as
well....
Mach 4.0 is an enhanced version of Mach 3.0, put out by the people at
the University of Utah. They are working on another free operating
system, and part of it includes an enhanced, more flexible version of
Mach. The Hurd has moved to Mach 4.0, which is good, because it is a
lot easier to build than 3.0 was.
You can find more information on Mach by browsing the Hurd pages given
in the next answer, or by looking at the Project Mach and Flux
homepages at:
Carnegie Mellon University (for Mach versions before 4.0):
http://www.cs.cmu.edu/afs/cs.cmu.edu/project/mach/public/www/mach.html
the University of Utah (for Mach 4.0):
http://www.cs.utah.edu/projects/flux/mach4/html/
Q5. Where can I find more information?
The June 1995 GNU's Bulletin contains the following official
information:
The GNU Hurd now runs programs native. We have implemented both
shared libraries using ELF, & the popular `ext2' file system used
by Linux. It can run GCC, `make', Emacs, & most other GNU
utilities. Progress is being made so rapidly that by the time you
read this it probably does much more. It is right on the verge of
being self-hosting (able to run on its own well enough to compile
its own source code, & be used for its own development). We have
much better device supportm [sic] & some new utilities, including a
fancy `ps' & `settrans'. For a complete system we still have much
more work to do, but we will make an alpha release as soon as the
network software is finished & shared libraries have been well
tested. We have a mailing list to announce progress; to be added
to it, ask `hurd-announce-request@gnu.org'.
The Portland State University CS department (via Trent Fisher)
maintains a WWW server with various Hurd documents, including Michael
Bushnell's Hurd paper, all the collected GNU's Bulletins, and various
announcements posted to "gnu.misc.discuss". The top-level GNU page is
http://www.cs.pdx.edu/~trent/gnu/gnu.html
and the Hurd page is
http://www.cs.pdx.edu/~trent/gnu/hurd/hurd.html
People in Europe might want to try the GNU WWW server for DESY
Germany, first:
http://info.desy.de/gnu/www
This site lacks culled, Hurd-specific information at the moment, but
it does have the last two GNU's Bulletins plus lots of general
information.
There is a snapshot of the Hurd development tree on
"alpha.gnu.ai.mit.edu" in the "/gnu" directory. It is updated as
significant changes are made, and not guaranteed to run.
You can subscribe to the Hurd announcement list by sending a request
to "hurd-announce-request@gnu.org". This is a moderated list
for distributing Hurd info to ``all and sundry'', and anyone can join.
In addition, there is a private (invitation-only) list for developers
to coordinate their efforts. It's not even worth thinking about
unless you (a) have a lot of free time on your hands, (b) know Unix
internals and Mach very well, and (c) have a proper machine.
Q6. What's a proper machine?
A ``proper machine'', at the moment, means an x86 box running Mach 3.0
(or 4.0), with FreeBSD 2.x, NetBSD 1.x, or Linux.
A single-server OS is no longer required for development because by
the time the Hurd bootstrap mechanism is finished, the Hurd will
probably be self-hosting.
Linux, FreeBSD, or NetBSD will only be required to splat the Hurd
binaries onto a partition of some sort, and to provide a way of
transferring files to the Hurd until the networking code is ready.
Q7. What sort of machines will run Hurd in the future?
The first thing a prospective Hurd machine needs is a Mach 3.0 port.
According to the most recent "comp.os.mach" FAQ (which hasn't been
updated since February 1994), the following chips have redistributable
Mach micro-kernels and device drivers:
Intel 80x86 (ISA and PS/2 buses)
Motorola 68000 (Sun 3)
Motorola 88000 (Omron Luna)
DEC Vax
DEC Pmax (DECstation 3100)
DEC Alpha
MIPS R4000 (DECstation 5000 et al.)
IBM RS/6000
Apple Macintosh
IBM is planning to run WorkplaceOS (the OS/2 successor) over Mach 3.0
on the PowerPC chip (closely related to the RS/6000), so the PowerPC
will likely be added to this list soon. The University of Utah has
ported Mach 4.0 to the HP700, but it is not yet stable.
Sun Sparc machines have a redistributable Mach microkernel, but the
device drivers require a SunOS 4.1.1 source license.
In addition, any prospective Hurd machine needs a port of the GNU C
library. Version 1.07.4 of the library can handle the following
chips:
Intel 80x86 (BSD, Dynix, Hurd, SCO, SysV)
Motorola 68000 (HP BSD, NEWS, Sun 4)
MIPS R4000 (Ultrix)
Sun Sparc (Solaris 2, Sun 4)
DEC Alpha (OSF/1, mostly finished)
So if the next Hurd snapshot is self-hosting, we will be able to run
it (in theory) on Intel 80x86s, Motorola 68000s, MIPS R4000s and DEC
Alphas.
People who can port the Mach micro-kernel to new architectures are
encouraged to do so. People who can port the GNU C library to new
chips (a much larger group) are also encouraged to do so. You can
help out here without knowing anything about Mach or having any
special machine. Note that once the GNU C library exists for a new
chip, for _any_ OS, making a Hurd port later is simple (and making
ports to other chips becomes easier as well---the effects are
cumulative).
By current indications, the other hardware requirements (RAM, disk
space, and the like) will be about the same as those of BSD 4.4.
Q8. What is the current development status?
Please see Trent Fisher's Hurd pages for details.
Q9. What sort of a system would we have if the Hurd was bootable
today?
Quite likely, if you already use an end-user system like Linux,
FreeBSD, or NetBSD, you'll be disappointed with the Hurd. It will
take some time before the OS hackers really get to work on
applications and major enhancements.
But, rest assured, Hurd development should proceed very rapidly.
Of course, if you think you can help, or you just enjoy neat stuff,
then you'll probably like the Hurd. When you actually understand a
fraction of what's going on behind the scenes, it's very impressive.
All I'm saying is that I'm not expecting all the Windows '95 users in
the world to switch to the Hurd right away. Wait a little while,
maybe 5-6 years (ample time for GNUStep and Guile to be in use), and
GNU users everywhere will be very happy that the FSF proceeded with
the Hurd. :)
==============================
Footnotes:
[[1]] Yes, I know that ``micro-kernel'' is about as apt a description
as ``Reduced Instruction Set Chip'', but we're stuck with it.
Updated:
23 Feb 1998 grat-w