summaryrefslogtreecommitdiff
path: root/hurd/documentation
diff options
context:
space:
mode:
Diffstat (limited to 'hurd/documentation')
-rw-r--r--hurd/documentation/auth.html168
-rw-r--r--hurd/documentation/hurd-paper.html760
-rw-r--r--hurd/documentation/hurd-talk.html1061
3 files changed, 1989 insertions, 0 deletions
diff --git a/hurd/documentation/auth.html b/hurd/documentation/auth.html
new file mode 100644
index 00000000..487fc1fe
--- /dev/null
+++ b/hurd/documentation/auth.html
@@ -0,0 +1,168 @@
+[[meta copyright="Copyright © 2002, 2008 Free Software Foundation, Inc."]]
+
+[[meta license="Verbatim copying and distribution of this entire article is
+permitted in any medium, provided this notice is preserved."]]
+
+[[meta title="The Authentication Server, the transcript of a talk about the
+details of the authentication mechanisms in the Hurd by Wolfgang Jährling"]]
+
+<H3><A NAME="contents">Table of Contents</A></H3>
+<UL>
+ <LI><A HREF="#intro" NAME="TOCintro">Introduction</A>
+ <LI><A HREF="#ids" NAME="TOCids">How IDs are represented and used</A>
+ <LI><A HREF="#posix" NAME="TOCposix">POSIX and beyond</A>
+ <LI><A HREF="#servers" NAME="TOCservers">Related servers</A>
+</UL>
+<HR>
+
+<H3><A HREF="#TOCintro" NAME="intro">Introduction</A></H3>
+<P>
+In this text, which mostly resembles the talk I gave at Libre Software
+Meeting 2002 in Bordeaux, I will describe what the auth server does,
+why it is so important and which cool things you can do with it, both
+on the programming and the user side. I will also describe related
+programs like the password and fakeauth servers. Note that this text
+is targeted at programmers who want to understand the auth mechanism
+in detail and are already familiar with concepts like Remote Procedure
+Calls (RPCs) as well as the way User- and Group-IDs are used in the
+POSIX world.
+
+<P>
+The auth server is a very small server, therefore it gives a useful
+example when you want to learn how a server typically looks like. One
+reason why it is so small is that the auth interface, which it
+implements, consists of only four RPCs. You can find the interface in
+hurd/hurd/auth.defs and the server itself in hurd/auth/.
+
+<H3><A HREF="#TOCids" NAME="ids">How IDs are represented and used</A></H3>
+<P>
+Each process holds (usually) one port to auth (an auth_t in C source,
+which actually is a mach_port_t, of course). The purpose of auth is
+to manage User-IDs and Group-IDs, which is the reason why users often
+will have no choice but to make use of the systems main auth server,
+which does not listen on /servers/auth; instead you inherit a port to
+auth from your parent process. Each such port is (internally in the
+auth server) associated with a set of effective User- and Group-IDs as
+well as a set of available User- and Group-IDs. So we have four sets
+of IDs in total. The available IDs can be turned into corresponding
+effective IDs at any time.
+
+<P>
+When you send an auth_getids RPC on the port you hold, you will get
+information about which IDs are associated with it, so you can figure
+out which permissions you have. But how will a server know that you
+have these permissions and therefore know which actions (e.g. writing
+into file "foo") it is supposed to do on your behalf and which not?
+The establishing of a trusted connection to a server works as follows:
+
+<P><OL>
+<LI>A user wants a server to know its IDs</LI>
+<LI>The user requests a reauthentication from the server</LI>
+<LI>In this request the user will include a port</LI>
+<LI>Both will hand this port to auth</LI>
+<LI>The user uses auth_user_authenticate</LI>
+<LI>The server uses auth_server_authenticate</LI>
+<LI>The server also passes a new port to auth</LI>
+<LI>auth matches these two requests</LI>
+<LI>The user gets the new port from auth</LI>
+<LI>The server learns about the IDs of the user</LI>
+<LI>The user uses the new port for further communication</LI>
+</OL>
+
+<P>
+We have different RPCs for users and servers because what we pass and
+what we get back differs for them: Users get a port, and servers get
+the sets of IDs, and have to specify the port which the user will get.
+
+<P>
+It is interesting to note that auth can match the requests by
+comparing two integers, because when you get the same port from two
+people, you will have the same mach_port_t (which is nothing but an
+integer).
+
+<P>
+All of this of course only works if they use the same auth server,
+which is why I said often you have no choice other than to use the
+one main auth server. But this is no serious restriction, as the auth server has
+almost no functionality one might want to replace. In fact, there is
+one replacement for the default auth implementation, but more on that
+later.
+
+<H3><A HREF="#TOCposix" NAME="posix">POSIX and beyond</A></H3>
+<P>
+Before we examine what is possible with this design, let us take a
+short look at how the POSIX semantics are implemented on top of this
+design. When a program that comes out of POSIX-land asks for its own
+effective User- or Group-ID, we will tell it about the first of the
+effective IDs. In the same sense, the POSIX real User- or Group-ID is
+the first available ID and the POSIX saved User- or Group-ID is the
+second available ID, which is why you have the same ID two times in
+the available IDs when you log into your GNU/Hurd machine (you can
+figure out which IDs you have with the program "ids", that basically
+just does an auth_getauth RPC). When you lack one of those IDs (for
+example when you have no effective Group-ID), a POSIX program asking
+for this particular information will get "-1" as the ID.
+
+<P>
+But as you can imagine, we can do more than what POSIX specifies. Fox
+example, we can modify our permissions. This is always done with the
+auth_makeauth RPC. In this RPC, you specify the IDs that should be
+associated with the new port. All of these IDs must be associated
+with either the port where the RPC is sent to or one of the additional
+ports you can specify; an exception is the superuser root, which is
+allowed to creat ports that are associated with arbitrary IDs.
+Hereby you can convert available into effective IDs.
+
+<P>
+This opens the door to a bunch of nice features. For example, we have
+the addauth program in the Hurd, which makes it possible to add an ID
+to either a single process or a group of processes if you hold the ID or know the
+appropriate password, and there is a corresponding rmauth program that
+removes an ID. So when you are working on your computer with GNU
+Emacs and want to edit a system configuration file, you switch to
+Emacs' shell-mode, do an "addauth root", enter the password, edit the
+file, and when you are done switch back to shell-mode and do "rmauth
+root". These programs have some interesting options, and there are
+various other programs, for setting the complete list of IDs (setauth)
+and so on.
+
+<H3><A HREF="#TOCservers" NAME="servers">Related servers</A></H3>
+<P>
+Finally, I want to explain two servers which are related to auth. The
+first is the password server, which listens on /servers/password. If
+you pass to it a User- or Group-ID and the correct password for it, it
+will return a port to auth to you which is associated with the ID you
+passed to it. It can create such a port because it is running as
+root. So let us assume you are an FTP server process. You will start
+as root, because you want to use port 21 (in this case, "port" does
+not refer to a mach_port_t, of course). But then, you can drop all
+your permissions so that you run without any ID. This makes it far
+less dangerous to communicate with yet unknown users over the
+network. But when someone now hands a username and password to you,
+you can ask the password server for a new auth port. The password
+server will check the data you pass to it, for example by looking into
+/etc/shadow, and if it is valid, it will ask the auth server for a new
+port. It receives this port from auth and then passes it on to you.
+So you have raised your permissions. (And for the very curious: Yes,
+we are well aware of the differences between this concept and
+capabilities; and we also do have some kinds of capabilities in
+various parts of the Hurd.)
+
+<P>
+My second example is the fakeauth server. It also implements the auth
+protocol. It is the part of the fakeroot implementation that gives a
+process the impression that it runs as root, even if it doesn't. So
+when the process asks fakeauth about its own IDs, fakeauth will tell
+the process that it runs as root. But when the process wants to make
+use of the authentication protocol described earlier in this text,
+fakeauth will forward the request to its own auth server, which will
+usually be the systems main auth server, which will then be able to
+match the auth_*_authenticate requests. So what fakeauth does is
+acting as a proxy auth server that gives someone the impression to run
+as root, while not modifying what that one is allowed to do.
+
+<P>
+At this point, I have said at least most of what can be said about the
+auth server and the protocol it implements, so I will finish by saying
+that it might be an interesting task (for you) to modify some existing
+software to take advantage of the features I described here.
diff --git a/hurd/documentation/hurd-paper.html b/hurd/documentation/hurd-paper.html
new file mode 100644
index 00000000..15d2daec
--- /dev/null
+++ b/hurd/documentation/hurd-paper.html
@@ -0,0 +1,760 @@
+[[meta copyright="Copyright © 1996, 1997, 1998, 2007, 2008 Free Software
+Foundation, Inc."]]
+
+[[meta license="Verbatim copying and distribution of this entire article is
+permitted in any medium, provided this notice is preserved."]]
+
+[[meta title="Towards a New Strategy of OS Design, an architectural overview by
+Thomas Bushnell, BSG."]]
+
+
+This article explains why FSF is developing a new operating system named the
+Hurd, which will be a foundation of the whole GNU system.
+The Hurd is built
+on top of CMU's Mach 3.0 kernel and uses Mach's virtual memory management and
+message-passing facilities.
+The GNU C Library will provide the Unix system
+call interface, and will call the Hurd for needed services it can't provide
+itself.
+The design and implementation of the Hurd is being lead by Michael
+Bushnell, with assistance from Richard Stallman, Roland McGrath,
+Jan Brittenson, and others.
+
+<H2>Part 1: A More Usable Approach to OS Design</H2>
+<P>
+The fundamental purpose of an operating system (OS) is to enable a variety of
+programs to share a single computer efficiently and productively.
+This
+demands memory protection, preemptively scheduled timesharing, coordinated
+access to I/O peripherals, and other services.
+In addition, an OS can allow
+several users to share a computer.
+In this case, efficiency demands services
+that protect users from harming each other, enable them to share without
+prior arrangement, and mediate access to physical devices.
+<P>
+On today's computer systems, programmers usually implement these goals
+through a large program called the kernel.
+Since this program must be
+accessible to all user programs, it is the natural place to add functionality
+to the system.
+Since the only model for process interaction is that of
+specific, individual services provided by the kernel, no one creates other
+places to add functionality.
+As time goes by, more and more is added to the
+kernel.
+<P>
+A traditional system allows users to add components to a kernel only if they
+both understand most of it and have a privileged status within the system.
+Testing new components requires a much more painful edit-compile-debug cycle
+than testing other programs.
+It cannot be done while others are using the
+system.
+Bugs usually cause fatal system crashes, further disrupting others'
+use of the system.
+The entire kernel is usually non-pageable.
+(There are
+systems with pageable kernels, but deciding what can be paged is difficult
+and error prone.
+Usually the mechanisms are complex, making them difficult
+to use even when adding simple extensions.)
+<P>
+Because of these restrictions, functionality which properly belongs
+<STRONG>behind</STRONG>
+the wall of a traditional kernel is usually left out of systems unless it is
+absolutely mandatory.
+Many good ideas, best done with an open/read/write
+interface cannot be implemented because of the problems inherent in the
+monolithic nature of a traditional system.
+Further, even among those with
+the endurance to implement new ideas, only those who are privileged users of
+their computers can do so.
+The software copyright system darkens the mire by
+preventing unlicensed people from even reading the kernel source.
+<P>
+Some systems have tried to address these difficulties.
+Smalltalk-80 and
+the Lisp Machine both represented one method of getting around the problem.
+System code is not distinguished from user code; all of the system is
+accessible to the user and can be changed as need be.
+Both systems were
+built around languages that facilitated such easy replacement and extension,
+and were moderately successful.
+But they both were fairly poor at insulating
+users and programs from each other, failing one of the principal goals of OS
+design.
+<P>
+Most projects that use the Mach 3.0 kernel carry on the hard-to-change
+tradition of OS design.
+The internal structure is different, but the same
+heavy barrier between user and system remains.
+The single-servers, while
+fairly easy to construct, inherit all the deficiencies of the monolithic
+kernels.
+<P>
+A multi-server divides the kernel functionality up into logical blocks with
+well-defined interfaces.
+Properly done, it is easier to make changes and add
+functionality.
+So most multi-server projects do somewhat better.
+Much more
+of the system is pageable.
+You can debug the system more easily.
+You can
+test new system components without interfering with other users.
+But the
+wall between user and system remains; no user can cross it without special
+privilege.
+<P>
+The GNU&nbsp;Hurd, by contrast, is designed to make the area of
+<STRONG>system</STRONG>
+code as
+limited as possible.
+Programs are required to communicate only with a few
+essential parts of the kernel; the rest of the system is replaceable
+dynamically.
+Users can use whatever parts of the remainder of the system
+they want, and can easily add components themselves for other users to take
+advantage of.
+No mutual trust need exist in advance for users to use each
+other's services, nor does the system become vulnerable by trusting the
+services of arbitrary users.
+<P>
+This has been done by identifying those system components which users
+<STRONG>must</STRONG>
+use in order to communicate with each other.
+One of these is responsible for
+identifying users' identities and is called the
+<DFN>
+authentication server.
+</DFN>
+In
+order to establish each other's identities, programs must communicate, each
+with an authentication server they trust.
+Another component establishes
+control over system components by the superuser, provides global bookkeeping
+operations, and is called the
+<DFN>
+process server.
+</DFN>
+<P>
+Not all user programs need to communicate with the process server; it is only
+necessary for programs which require its services.
+Likewise, the
+authentication server is only necessary for programs that wish to communicate
+their identity to another.
+None of the remaining services carry any special
+status; not the network implementation, the filesystems, the program
+execution mechanism (including setuid), or any others.
+
+<H3>The Translator Mechanism</H3>
+<P>
+The Hurd uses Mach ports primarily as methods for communicating between users
+and servers.
+(A Mach port is a communication point on a Mach task where
+messages are sent and received.) Each port implements a particular set of
+protocols, representing operations that can be undertaken on the underlying
+object represented by the port.
+Some of the protocols specified by the Hurd
+are the I/O protocol, used for generic I/O operations; the file protocol,
+used for filesystem operations; the socket protocol, used for network
+operations; and the process protocol, used for manipulating processes et al.
+<P>
+Most servers are accessed by opening files.
+Normally, when you open a file,
+you create a port associated with that file that is owned by the server
+that owns the directory containing the file.
+For example, a disk-based
+filesystem will normally serve a large number of ports, each of which
+represents an open file or directory.
+When a file is opened, the server
+creates a new port, associates it with the file, and returns the port to the
+calling program.
+<P>
+However, a file can have a
+<DFN>translator</DFN>
+associated with it.
+In this case,
+rather than return its own port which refers to the contents of the file, the
+server executes a translator program associated with that file.
+This
+translator is given a port to the actual contents of the file, and is then
+asked to return a port to the original user to complete the open operation.
+<P>
+This mechanism is used for
+<CODE>mount</CODE>
+by having a translator associated with
+each mount point.
+When a program opens the mount point, the translator (in
+this case, a program which understands the disk format of the mounted
+filesystem) is executed and returns a port to the program.
+After the
+translator is started, it need not be run again unless it dies; the parent
+filesystem retains a port to the translator to use in further requests.
+<P>
+The owner of a file can associate a translator with it without special
+permission.
+This means that any program can be specified as a translator.
+Obviously the system will not work properly if the translator does not
+implement the file protocol correctly.
+However, the Hurd is constructed so
+that the worst possible consequence is an interruptible hang.
+<P>
+One way to use translators is to access hierarchically structured data using
+the file protocol.
+For example, all the complexity of the user interface to
+the
+<CODE>ftp</CODE>
+program is removed.
+Users need only know that a particular
+directory represents FTP and can use all the standard file manipulation
+commands (e.g
+<CODE>ls</CODE>
+or
+<CODE>cp</CODE>)
+to access the remote system, rather than learning
+a new set.
+Similarly, a simple translator could ease the complexity of
+<CODE>tar</CODE>
+or
+<CODE>gzip</CODE>.
+(Such transparent access would have some added cost, but it would
+be convenient.)
+
+<H3>Generic Services</H3>
+<P>
+With translators, the filesystem can act as a rendezvous for interfaces which
+are not similar to files.
+Consider a service which implements some version
+of the X protocol, using Mach messages as an underlying transport.
+For each
+X display, a file can be created with the appropriate program as its
+translator.
+X clients would open that file.
+At that point, few file
+operations would be useful (read and write, for example, would be useless),
+but new operations (
+<CODE>XCreateWindow</CODE>
+or
+<CODE>XDrawText</CODE>)
+might become meaningful.
+In this case, the filesystem protocol is used only to manipulate
+characteristics of the node used for the rendezvous.
+The node need not
+support I/O operations, though it should reply to any such messages with a
+<CODE>message_not_understood</CODE>
+return code.
+<P>
+This translator technique is used to contact most of the services in the Hurd
+that are not structured like hierarchical filesystems.
+For example, the
+password server, which hands out authorization tags in exchange for
+passwords, is contacted this way.
+Network protocol servers are also
+contacted in this fashion.
+Roland McGrath thought up this use of translators.
+
+<H3>Clever Filesystem Pictures</H3>
+<P>
+In the Hurd, translators can also be used to present a filesystem-like view
+of another part of the filesystem, with some semantics changed.
+For example,
+it would be nice to have a filesystem that cannot itself be changed, but
+nonetheless records changed versions of its files elsewhere.
+(This could be
+useful for source code management.)
+<P>
+The Hurd will have a translator which creates a directory which is a
+conceptual union of other directories, with collision resolution rules of
+various sorts.
+This can be used to present a single directory to users that
+contains all the programs they would want to execute.
+There are other useful
+variations on this theme.
+
+<H3>What The User Can Do</H3>
+<P>
+No translator gains extra privilege by virtue of being hooked into the
+filesystem.
+Translators run with the uid of the owner of the file being
+translated, and can only be set or changed by that owner.
+The I/O and
+filesystem protocols are carefully designed to allow their use by mutually
+untrusting clients and servers.
+Indeed, translators are just ordinary
+programs.
+The GNU C library has a variety of facilities to make common sorts
+of translators easier to write.
+<P>
+Some translators may need special privileges, such as the password server or
+translators which allow setuid execution.
+These translators could be run by
+anyone, but only if they are set on a root-owned node would they be able to
+provide all their services successfully.
+This is analogous to letting any
+user call the
+<CODE>reboot</CODE>
+system call, but only honoring it if that user is root.
+
+<H3>Why This Is So Different</H3>
+<P>
+What this design provides is completely novel to the Unix world.
+Until now,
+OSs have kept huge portions of their functionality in the realm of system
+code, thus preventing its modification and extension except in extreme need.
+Users cannot replace parts of the system in their programs no matter how much
+easier that would make their task, and system managers are loath to install
+random tweaks off the net into their kernels.
+<P>
+In the Hurd, users can change almost all of the things that are decided for
+them in advance by traditional systems.
+In combination with the tremendous
+control given by the Mach kernel over task address spaces and properties, the
+Hurd provides a system in which users will, for the first time, be able to
+replace parts of the system they dislike, without disrupting other users.
+<P>
+Most Mach-based OSs to date have mostly implemented a wider set of the
+<STRONG>
+same old
+</STRONG>
+Unix semantics in a new environment.
+In contrast, GNU is extending
+those semantics to allow users to improve, bypass, or replace them.
+
+
+<H2>Part 2: A Look at Some of the Hurd's Beasts</H2>
+<H3>The Authentication Server</H3>
+<P>
+One of the Hurd's more central servers is the authentication server.
+Each
+port to this server identifies a user and is associated by this server with
+an
+<DFN>id block</DFN>.
+Each id block contains sets of user and group ids.
+Either
+set may be empty.
+This server is not the same as the password server
+referred to above.
+<P>
+The authentication server exports three services.
+First, it provides simple
+boolean operations on authentication ports: given two authentication ports,
+this server will provide a third port representing the union of the two sets
+of uids and gids.
+Second, this server allows any user with a uid of zero to
+create an arbitrary authentication port.
+Finally, this server provides RPCs
+(Remote Procedure Calls between different programs and possibly different
+hosts) which allow mutually untrusting clients and servers to establish their
+identities and pass initial information on each other.
+This is crucial to
+the security of the filesystem and I/O protocols.
+<P>
+Any user could write a program which implements the authentication protocol;
+this does not violate the system's security.
+When a service needs to
+authenticate a user, it communicates with its trusted authentication server.
+If that user is using a different authentication server, the transaction will
+fail and the server can refuse to communicate further.
+Because, in effect,
+this forces all programs on the system to use the same authentication server,
+we have designed its interface to make any safe operation possible, and to
+include no extraneous operations.
+(This is why there is a separate password
+server.)
+<H3>The Process Server</H3>
+<P>
+The process server acts as an information categorization repository.
+There
+are four main services supported by this server.
+First, the process server
+keeps track of generic host-level information not handled by the Mach kernel.
+For example, the hostname, the hostid, and the system version are maintained
+by the process server.
+Second, this server maintains the Posix notions of
+sessions and process groups, to help out programs that wish to use Posix
+features.
+<P>
+Third, the process server maintains a one-to-one mapping between Mach tasks
+and Hurd processes.
+Every task is assigned a pid.
+Processes can register a
+message port with this server, which can then be given out to any program
+which requests it.
+This server makes no attempt to keep these message ports
+private, so user programs are expected to implement whatever security they
+need themselves.
+(The GNU C Library provides convenient functions for all
+this.) Processes can tell the process server their current `argv' and `envp'
+values; this server will then provide, on request, these vectors of arguments
+and environment.
+This is useful for writing
+<CODE>ps</CODE>-like
+programs and also
+makes it easier to hide or change this information.
+None of these features
+are mandatory.
+Programs are free to disregard all of this and never register
+themselves with the process server at all.
+They will, however, still have a
+pid assigned.
+<P>
+Finally, the process server implements
+<DFN>process collections</DFN>,
+which are used
+to collect a number of process message ports at the same time.
+Also,
+facilities are provided for converting between pids, process server ports,
+and Mach task ports, while ensuring the security of the ports managed.
+<P>
+It is important to stress that the process server is optional.
+Because of
+restrictions in Mach, programs must run as root in order to identify all the
+tasks in the system.
+But given that, multiple process servers could
+co-exist, each with their own clients, giving their own model of the
+universe.
+Those process server features which do not require root privileges
+to be implemented could be done as per-user servers.
+The user's hands are
+not tied.
+<H3>Transparent FTP</H3>
+<P>
+Transparent FTP is an intriguing idea whose time has come.
+The popular
+<CODE>ange-ftp</CODE>
+package available for GNU Emacs makes access to FTP files
+virtually transparent to all the Emacs file manipulation functions.
+Transparent FTP does the same thing, but in a system wide fashion.
+This
+server is not yet written; the details remain to be fleshed out, and will
+doubtless change with experience.
+<P>
+In a BSD kernel, a transparent FTP filesystem would be no harder to write
+than in the Hurd.
+But mention the idea to a BSD kernel hacker, and the
+response is that ``such a thing doesn't belong in the kernel''.
+In a sense,
+this is correct.
+It violates all the layering principles of such systems to
+place such things in the kernel.
+The unfortunate side effect, however, is
+that the design methodology (which is based on preventing users from changing
+things they don't like) is being used to prevent system designers from making
+things better.
+(Recent BSD kernels make it possible to write a user program
+that provides transparent FTP.
+An example is
+<CODE>alex</CODE>,
+but it needs to run
+with full root privileges.)
+<P>
+In the Hurd, there are no obstacles to doing transparent FTP.
+A translator
+will be provided for the node
+<CODE>/ftp</CODE>.
+The contents of
+<CODE>/ftp</CODE>
+will probably
+not be directly listable, though further subdirectories will be.
+There will
+be a variety of possible formats.
+For example, to access files on uunet, one
+could
+<CODE>
+cd /ftp/ftp.uu.net:anonymous:mib@gnu.
+</CODE>
+Or to access files on a remote
+account, one might
+<CODE>
+cd /ftp/gnu.org:mib:passwd.
+</CODE>
+Parts of this
+command could be left out and the transparent FTP program would read them
+from a user's
+<CODE>.netrc</CODE>
+file.
+In the last case, one might just
+<CODE>
+cd /ftp/gnu.org;
+</CODE>
+when the rest of the data is already in
+<CODE>.netrc</CODE>.
+<P>
+There is no need to do a
+<CODE>cd</CODE>
+first--use any file command.
+To find out about
+RFC 1097 (the Telnet Subliminal Message Option), just type
+<CODE>
+more /ftp/ftp.uu.net/inet/rfc/rfc1097.
+</CODE>
+A copy command to a local disk
+could be used if the RFC would be read frequently.
+<H3>Filesystems</H3>
+<P>
+Ordinary filesystems are also being implemented.
+The initial release of the
+Hurd will contain a filesystem upwardly compatible with the BSD 4.4 Fast File
+System.
+In addition to the ordinary semantics, it will provide means to
+record translators, offer thirty-two bit user ids and group ids, and supply a
+new id per file, called the
+<DFN>author</DFN>
+of the file, which can be set by the
+owner arbitrarily.
+In addition, because users in the Hurd can have multiple
+uids (or even none), there is an additional set of permission bits providing
+access control for
+<DFN>
+unknown user
+</DFN>
+(no uids) as distinct from
+<DFN>
+known but arbitrary user
+</DFN>
+(some uids: the existing
+<DFN>world</DFN>
+category of file
+permissions).
+<P>
+The Network File System protocol will be implemented using 4.4 BSD as a
+starting point.
+A log-structured filesystem will also be implemented using
+the same ideas as in Sprite, but probably not the same format.
+A GNU network
+file protocol may be designed in time, or NFS may be extended to remove its
+deficiencies.
+There will also be various ``little'' filesystems, such as the
+MS-DOS filesystem, to help people move files between GNU and other OSs.
+
+<H3>Terminals</H3>
+<P>
+An I/O server will provide the terminal semantics of Posix.
+The GNU C
+Library has features for keeping track of the controlling terminal and for
+arranging to have proper job control signals sent at the proper times, as
+well as features for obeying keyboard and hangup signals.
+<P>
+Programs will be able to insert a terminal driver into communications
+channels in a variety of ways.
+Servers like
+<CODE>rlogind</CODE>
+will be able to insert
+the terminal protocol onto their network communication port.
+Pseudo-terminals will not be necessary, though they will be provided for
+backward compatibility with older programs.
+No programs in GNU will depend
+on them.
+<P>
+Nothing about a terminal driver is forced upon users.
+A terminal driver
+allows a user to get at the underlying communications channel easily, to
+bypass itself on an as-needed basis or altogether, or to substitute a
+different terminal driver-like program.
+In the last case, provided the
+alternate program implements the necessary interfaces, it will be used by the
+C Library exactly as if it were the ordinary terminal driver.
+<P>
+Because of this flexibility, the original terminal driver will not provide
+complex line editing features, restricting itself to the behavior found in
+Posix and BSD.
+In time, there will be a
+<CODE>readline</CODE>-based
+terminal driver,
+which will provide complex line-editing features for those users who want
+them.
+<P>
+The terminal driver will probably not provide good support for the
+high-volume, rapid data transmission required by UUCP or SLIP.
+Those
+programs do not need any of its features.
+Instead they will be using the
+underlying Mach device ports for terminals, which support moving large
+amounts of data efficiently.
+
+<H3>Executing Programs</H3>
+<P>
+The implementation of the
+<CODE>execve</CODE>
+call is spread across three programs.
+The
+library marshals the argument and environment vectors.
+It then sends a
+message to the file server that holds the file to be executed.
+The file
+server checks execute permissions and makes whatever changes it desires in
+the exec call.
+For example, if the file is marked setuid and the fileserver
+has the ability, it will change the user identification of the new image.
+The file server also decides if programs which had access to the old task
+should continue to have access to the new task.
+If the file server is
+augmenting permissions, or executing an unreadable image, then the exec needs
+to take place in a new Mach task to maintain security.
+<P>
+After deciding the policy associated with the new image, the filesystem calls
+the exec server to load the task.
+This server, using the BFD (Binary File
+Descriptor) library, loads the image.
+BFD supports a large number of object
+file formats; almost any supported format will be executable.
+This server
+also handles scripts starting with
+<CODE>#!</CODE>,
+running them through the indicated
+program.
+<P>
+The standard exec server also looks at the environment of the new image; if
+it contains a variable
+<CODE>EXECSERVERS</CODE>
+then it uses the programs specified
+there as exec servers instead of the system default.
+(This is, of course,
+not done for execs that the file server has requested be kept secure.)
+<P>
+The new image starts running in the GNU C Library, which sends a message to
+the exec server to get the arguments, environment, umask, current directory,
+etc.
+None of this additional state is special to the file or exec servers;
+if programs wish, they can use it in a different manner than the Library.
+
+<H3>New Processes</H3>
+<P>
+The
+<CODE>fork</CODE>
+call is implemented almost entirely in the GNU C Library.
+The new
+task is created by Mach kernel calls.
+The C Library arranges to have its
+image inherited properly.
+The new task is registered with the process server
+(though this is not mandatory).
+The C Library provides vectors of functions
+to be called at fork time: one vector to be called before the fork, one after
+in the parent, and one after in the child.
+(These features should not be
+used to replace the normal fork-calling sequence; it is intended for
+libraries which need to close ports or clean up before a fork occurs.)
+The C
+library will implement both fork calls specified by the draft Posix.4a (the
+proposed standard dealing with the threads extension to the real-time
+extension).
+<P>
+Nothing forces the user to create new tasks this way.
+If a program wants to
+use almost the normal fork, but with some special characteristics, then it
+can do so.
+Hooks will be provided by the C Library, or the function can even
+be completely replaced.
+None of this is possible in a traditional Unix
+system.
+
+<H3>Asynchronous Messages</H3>
+<P>
+As mentioned above, the process server maintains a
+<DFN>
+message port
+</DFN>
+for each
+task registered with it.
+These ports are public, and are used to send
+asynchronous messages to the task.
+Signals, for example, are sent to the
+message port.
+The signal message also provides a port as an indication that
+the sender should be trusted to send the signal.
+The GNU C Library lists a
+variety of ports in a table, each of which identifies a set of signals that
+can be sent by anyone who possesses that port.
+For example, if the user
+possesses the task's kernel port, it is allowed to send any signal.
+If the
+user possesses a special
+<DFN>
+terminal id
+</DFN>
+port, it is allowed to send the
+keyboard and hangup signals.
+Users can add arbitrary new entries into the C
+library's signal permissions table.
+<P>
+When a process's process group changes, the process server will send it a
+message indicating the new process group.
+In this case, the process server
+proves its authority by providing the task's kernel port.
+<P>
+The C library also has messages to add and delete uids currently used by the
+process.
+If new uids are sent to the program, the library adds them to its
+current set, and then exchanges messages with all the I/O servers it knows
+about, proving to them its new authorization.
+Similarly, a message can
+delete uids.
+In the latter case, the caller must provide the process's task
+port.
+(You can't harm a process by giving it extra permission, but you can
+harm it by taking permission away.) The Hurd will provide user programs to
+send these messages to processes.
+For example, the
+<CODE>su</CODE>
+command will be able
+to cause all the programs in your current login session, to gain a new uid,
+rather than spawn a subshell.
+<P>
+The C library will allow programs to add asynchronous messages they wish to
+recognize, as well as prevent recognition of the standard set.
+<H3>Making It Look Like Unix</H3>
+<P>
+The C Library will implement all of the calls from BSD and Posix as well as
+some obvious extensions to them.
+This enables users to replace those calls
+they dislike or bypass them entirely, whereas in Unix the calls must be used
+``as they come'' with no alternatives possible.
+<P>
+In some environments binary compatibility will also be supported.
+This works
+by building a special version of the library which is then loaded somewhere
+in the address space of the process.
+(For example, on a VAX, it would be
+tucked in above the stack.) A feature of Mach, called system call
+redirection, is then used to trap Unix system calls and turn them into jumps
+into this special version of the library.
+(On almost all machines, the cost
+of such a redirection is very small; this is a highly optimized path in Mach.
+On a 386 it's about two dozen instructions.
+This is little worse than a
+simple procedure call.)
+<P>
+Many features of Unix, such as signal masks and vectors, are handled
+completely by the library.
+This makes such features significantly cheaper
+than in Unix.
+It is now reasonable to use
+<CODE>sigblock</CODE>
+extensively to protect
+critical sections, rather than seeking out some other, less expensive method.
+
+<H3>Network Protocols</H3>
+<P>
+The Hurd will have a library that will make it very easy to port 4.4 BSD
+protocol stacks into the Hurd.
+This will enable operation, virtually for
+free, of all the protocols supported by BSD.
+Currently, this includes the
+CCITT protocols, the TCP/IP protocols, the Xerox NS protocols, and the ISO
+protocols.
+<P>
+For optimal performance some work would be necessary to take advantage of
+Hurd features that provide for very high speed I/O.
+For most protocols this
+will require some thought, but not too much time.
+The Hurd will run the
+TCP/IP protocols as efficiently as possible.
+<P>
+As an interesting example of the flexibility of the Hurd design, consider the
+case of IP trailers, used extensively in BSD for performance.
+While the Hurd
+will be willing to send and receive trailers, it will gain fairly little
+advantage in doing so because there is no requirement that data be copied and
+avoiding copies for page-aligned data is irrelevant.
diff --git a/hurd/documentation/hurd-talk.html b/hurd/documentation/hurd-talk.html
new file mode 100644
index 00000000..d608e12a
--- /dev/null
+++ b/hurd/documentation/hurd-talk.html
@@ -0,0 +1,1061 @@
+[[meta copyright="Copyright © 2001 Marcus Brinkmann"]]
+
+[[meta license="Verbatim copying and distribution of this entire article is
+permitted in any medium, provided this notice is preserved."]]
+
+[[meta title="The Hurd, a presentation by Marcus Brinkmann"]]
+
+
+<H4><A NAME="contents">Table of Contents</A></H4>
+<UL>
+ <LI><A HREF="#int" NAME="TOCint">Introduction</A>
+ <LI><A HREF="#ove" NAME="TOCove">Overview</A>
+ <LI><A HREF="#his" NAME="TOChis">Historicals</A>
+ <LI><A HREF="#ker" NAME="TOCker">Kernel Architectures</A>
+ <LI><A HREF="#mic" NAME="TOCmic">Micro vs Monolithic</A>
+ <LI><A HREF="#sin" NAME="TOCsin">Single Server vs Multi Server</A>
+ <LI><A HREF="#mul" NAME="TOCmul">Multi Server is superior, ...</A>
+ <LI><A HREF="#the" NAME="TOCthe">The Hurd even more so.</A>
+ <LI><A HREF="#mac" NAME="TOCmac">Mach Inter Process Communication</A>
+ <LI><A HREF="#how" NAME="TOChow">How to get a port?</A>
+ <LI><A HREF="#exa" NAME="TOCexa">Example of <SAMP>hurd_file_name_lookup</SAMP></A>
+ <LI><A HREF="#pat" NAME="TOCpat">Pathname resolution example</A>
+ <LI><A HREF="#map" NAME="TOCmap">Mapping the POSIX Interface</A>
+ <LI><A HREF="#filser" NAME="TOCfilser">File System Servers</A>
+ <LI><A HREF="#act" NAME="TOCact">Active vs Passive</A>
+ <LI><A HREF="#aut" NAME="TOCaut">Authentication</A>
+ <LI><A HREF="#ope" NAME="TOCope">Operations on authentication ports</A>
+ <LI><A HREF="#est" NAME="TOCest">Establishing trusted connections</A>
+ <LI><A HREF="#pas" NAME="TOCpas">Password Server</A>
+ <LI><A HREF="#pro" NAME="TOCpro">Process Server</A>
+ <LI><A HREF="#filsys" NAME="TOCfilsys">Filesystems</A>
+ <LI><A HREF="#dev" NAME="TOCdev">Developing the Hurd</A>
+ <LI><A HREF="#sto" NAME="TOCsto">Store Abstraction</A>
+ <LI><A HREF="#deb" NAME="TOCdeb">Debian GNU/Hurd</A>
+ <LI><A HREF="#stabin" NAME="TOCstabin">Status of the Debian GNU/Hurd binary archive</A>
+ <LI><A HREF="#stainf" NAME="TOCstainf">Status of the Debian infrastructure</A>
+ <LI><A HREF="#staarc" NAME="TOCstaarc">Status of the Debian Source archive</A>
+ <LI><A HREF="#debide" NAME="TOCdebide">Debian GNU/Hurd: Good idea, bad idea?</A>
+ <LI><A HREF="#end" NAME="TOCend">End</A>
+</UL>
+<HR>
+<H3>Talk about the Hurd</H3>
+<P>
+This talk about the Hurd was written by Marcus Brinkmann for
+<UL>
+<LI>OSDEM, Brussels, 4. Feb 2001,
+<LI>Frühjahrsfachgespräche, Cologne, 2. Mar 2001 and
+<LI>Libre Software Meeting, Bordeaux, 4. Jul 2001.
+</UL>
+
+<H4><A HREF="#TOCint" NAME="int">Introduction</A></H4>
+<P>
+When we talk about free software, we usually refer to the free
+software licenses. We also need relief from software patents, so our
+freedom is not restricted by them. But there is a third type of
+freedom we need, and that's user freedom.
+
+<P>
+Expert users don't take a system as it is. They like to change the
+configuration, and they want to run the software that works best for
+them. That includes window managers as well as your favourite text
+editor. But even on a GNU/Linux system consisting only of free
+software, you can not easily use the filesystem format, network
+protocol or binary format you want without special privileges. In
+traditional unix systems, user freedom is severly restricted by the
+system administrator.
+
+<P>
+The Hurd removes these restrictions from the user. It provides an
+user extensible system framework without giving up POSIX compatibility
+and the unix security model. Throughout this talk, we will see that
+this brings further advantages beside freedom.
+
+<H4><A HREF="#TOCove" NAME="ove">Overview</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+
+<P>
+The Hurd is a POSIX compatible multi-server
+system operating on top of the GNU&nbsp;Mach microkernel.
+
+<P>
+Topics:
+<UL>
+ <LI>GNU&nbsp;Mach</LI>
+ <LI>The Hurd</LI>
+ <LI>Development</LI>
+ <LI>Debian GNU/Hurd</LI>
+</UL>
+</TD></TR></TABLE>
+
+<P>
+The Hurd is a POSIX compatible multi-server system operating on top of
+the GNU&nbsp;Mach Microkernel.
+
+<P>
+I will have to explain what GNU&nbsp;Mach is, so we start with that. Then
+I will talk about the Hurd's architecture. After that, I will give a
+short overview on the Hurd libraries. Finally, I will tell you how
+the Debian project is related to the Hurd.
+
+<H4><A HREF="#TOChis" NAME="his">Historicals</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%">
+<TR><TD VALIGN="TOP" ALIGN="LEFT">
+<UL>
+ <LI>1983: Richard Stallman founds the GNU project.</LI>
+ <LI>1988: Decision is made to use Mach 3.0 as the kernel.</LI>
+ <LI>1991: Mach 3.0 is released under compatible license.</LI>
+ <LI>1991: Thomas Bushnell, BSG, founds the Hurd project.</LI>
+ <LI>1994: The Hurd boots the first time.</LI>
+ <LI>1997: Version 0.2 of the Hurd is released.<BR><BR></LI>
+ <LI>1998: Debian hurd-i386 archive is created.</LI>
+ <LI>2001: Debian GNU/Hurd snapshot fills three CD images.</LI>
+</UL>
+</TD></TR></TABLE>
+
+<P>
+When Richard Stallman founded the GNU project in 1983, he wanted to
+write an operating system consisting only of free software. Very
+soon, a lot of the essential tools were implemented, and released
+under the GPL. However, one critical piece was missing: The kernel.
+<P>
+After considering several alternatives, it was decided not to write a
+new kernel from scratch, but to start with the Mach microkernel. This
+was in 1988, and it was not before 1991 that Mach was released under a
+license allowing the GNU project to distribute it as a part of the
+system.
+<P>
+In 1998, I started the Debian GNU/Hurd project, and in 2001 the number
+of available GNU/Hurd packages fills three CD images.
+
+<H4><A HREF="#TOCker" NAME="ker">Kernel Architectures</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Microkernel:
+<UL>
+ <LI>Enforces resource management (paging, scheduling)</LI>
+ <LI>Manages tasks</LI>
+ <LI>Implements message passing for IPC</LI>
+ <LI>Provides basic hardware support</LI>
+</UL>
+<P>
+Monolithic kernel:
+<UL>
+ <LI>No message passing necessary</LI>
+ <LI>Rich set of features (filesystems, authentication, network
+ sockets, POSIX interface, ...)</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+Microkernels were very popular in the scientific world around that
+time. They don't implement a full operating system, but only the
+infrastructure needed to enable other tasks to implement most
+features. In contrast, monolithical kernels like Linux contain
+program code of device drivers, network protocols, process management,
+authentication, file systems, POSIX compatible interfaces and much
+more.
+<P>
+So what are the basic facilities a microkernel provides? In general,
+this is resource management and message passing. Resource management,
+because the kernel task needs to run in a special privileged mode of
+the processor, to be able to manipulate the memory management unit and
+perform context switches (also to manage interrupts). Message
+passing, because without a basic communication facility the other
+tasks could not interact to provide the system services. Some
+rudimentary hardware device support is often necessary to bootstrap
+the system. So the basic jobs of a microkernel are enforcing the
+paging policy (the actual paging can be done by an external pager
+task), scheduling, message passing and probably basic hardware device
+support.
+<P>
+Mach was the obvious choice back then, as it provides a rich set of
+interfaces to get the job done. Beside a rather brain-dead device
+interface, it provides tasks and threads, a messaging system allowing
+synchronous and asynchronous operation and a complex interface for
+external pagers. It's certainly not one of the sexiest microkernels
+that exist today, but more like a big old mama. The GNU project
+maintains its own version of Mach, called GNU&nbsp;Mach, which is based on
+Mach 4.0. In addition to the features contained in Mach 4.0, the GNU
+version contains many of the Linux 2.0 block device and network card
+drivers.
+<P>
+A complete treatment of the differences between a microkernel and
+monolithical kernel design can not be provided here. But a couple of
+advantages of a microkernel design are fairly obvious.
+
+<H4><A HREF="#TOCmic" NAME="mic">Micro vs Monolithic</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Microkernel
+<UL>
+ <LI>Clear cut responsibilities
+ <LI>Flexibility in operating system design, easier debugging</LI>
+ <LI>More stability (less code to break)</LI>
+ <LI>New features are not added to the kernel</LI>
+</UL>
+<P>
+Monolithic kernel
+<UL>
+ <LI>Intolerance or creeping featuritis</LI>
+ <LI>Danger of spaghetti code</LI>
+ <LI>Small changes can have far reaching side effects</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+Because the system is split up into several components, clean
+interfaces have to be developed, and the responsibilities of each part
+of the system must be clear.
+<P>
+Once a microkernel is written, it can be used as the base for several
+different operating systems. Those can even run in parallel which
+makes debugging easier. When porting, most of the hardware dependant
+code is in the kernel.
+<P>
+Much of the code that doesn't need to run in the special kernel mode
+of the processor is not part of the kernel, so stability increases
+because there is simply less code to break.
+<P>
+New features are not added to the kernel, so there is no need to hold
+the barrier high for new operating system features.
+<P>
+Compare this to a monolithical kernel, where you either suffer from
+creeping featuritis or you are intolerant of new features (we see both
+in the Linux kernel).
+<P>
+Because in a monolithical kernel, all parts of the kernel can access
+all data structures in other parts, it is more likely that short cuts
+are used to avoid the overhead of a clean interface. This leads to a
+simple speed up of the kernel, but also makes it less comprehensible
+and more error prone. A small change in one part of the kernel can
+break remote other parts.
+
+<H4><A HREF="#TOCsin" NAME="sin">Single Server vs Multi Server</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Single Server
+<UL>
+ <LI>A single task implements the functionality of the operating system.</LI>
+</UL>
+<P>
+Multi Server
+<UL>
+ <LI>Many tasks cooperate to provide the system's functionality.</LI>
+ <LI>One server provides only a small but well-defined part of the
+ whole system.</LI>
+ <LI>The responsibilities are distributed logically among the servers.</LI>
+</UL>
+<P>
+A single-server system is comparable to a monolithic kernel system. It
+has similar
+advantages and disadvantages.
+</TD></TR></TABLE>
+<P>
+There exist a couple of operating systems based on Mach, but they all
+have the same disadvantages as a monolithical kernel, because those
+operating systems are implemented in one single process running on top
+of the kernel. This process provides all the services a monolithical
+kernel would provide. This doesn't make a whole lot of sense (the
+only advantage is that you can probably run several of such isolated
+single servers on the same machine). Those systems are also called
+single-server systems. The Hurd is the only usable multi-server
+system on top of Mach. In the Hurd, there are many server programs,
+each one responsible for a unique service provided by the operating
+system. These servers run as Mach tasks, and communicate using the
+Mach message passing facilities. One of them does only provide a
+small part of the functionality of the system, but together they build
+up a complete and functional POSIX compatible operating system.
+
+<H4><A HREF="#TOCmul" NAME="mul">Multi Server is superior, ...</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Any multi-server has advantages over single-server:
+<UL>
+ <LI>Clear cut responsibilities</LI>
+ <LI>More stability: If one server dies, all others remain</LI>
+ <LI>Easier development cycle: Testing without reboot (or replacing
+ running servers), debugging with gdb</LI>
+ <LI>Easier to make changes and add new features
+</UL>
+</TD></TR></TABLE>
+<P>
+Using several servers has many advantages, if done right. If a file
+system server for a mounted partition crashes, it doesn't take down
+the whole system. Instead the partition is "unmounted", and
+you can try to start the server again, probably debugging it this time
+with gdb. The system is less prone to errors in individual
+components, and over-all stability increases. The functionality of
+the system can be extended by writing and starting new servers
+dynamically. (Developing these new servers is easier for the reasons
+just mentioned.)
+<P>
+But even in a multi-server system the barrier between the system and
+the users remains, and special privileges are needed to cross it. We
+have not achieved user freedom yet.
+
+<H4><A HREF="#TOCthe" NAME="the">The Hurd even more so.</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+The Hurd goes beyond all this, and allows users to write and run their
+servers, too!
+<UL>
+ <LI>Users can replace system servers dynamically with their own
+ implementations.</LI>
+ <LI>Users can decide what parts of the remainder of the system they
+ want to use.</LI>
+ <LI>Users can extend the functionality of the system.</LI>
+ <LI>No mutual trust necessary to make use of other users
+ services.</LI>
+ <LI>Security of the system is not harmed by trusting users
+ services.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+To quote Thomas Bushnell, BSG, from his paper
+[[``Towards_a_New_Strategy_of_OS_design''_(1996)|hurd-paper]]:
+<BLOCKQUOTE>
+The GNU&nbsp;Hurd, by contrast, is designed to make the area of system code
+as limited as possible. Programs are required to communicate only
+with a few essential parts of the kernel; the rest of the system is
+replaceable dynamically. Users can use whatever parts of the
+remainder of the system they want, and can easily add components
+themselves for other users to take advantage of. No mutual trust need
+exist in advance for users to use each other's services, nor does the
+system become vulnerable by trusting the services of arbitrary users.
+</BLOCKQUOTE>
+
+<P>
+<EM>So the Hurd is a set of servers running on top of the Mach
+micro-kernel, providing a POSIX compatible and extensible operating
+system. What servers are there? What functionality do they provide,
+and how do they cooperate?</EM>
+
+<H4><A HREF="#TOCmac" NAME="mac">Mach Inter Process Communication</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Ports are message queues which can be used as one-way communication
+channels.
+<UL>
+ <LI>Port rights are receive, send or send-once</LI>
+ <LI>Exactly one receiver</LI>
+ <LI>Potentially many senders</LI>
+</UL>
+<P>
+MIG provides remote procedure calls on top of Mach IPC. RPCs look like
+function calls to the user.
+</TD></TR></TABLE>
+<P>
+Inter-process communication in Mach is based on the ports concept. A
+port is a message queue, used as a one-way communication channel. In
+addition to a port, you need a port right, which can be a send right,
+receive right, or send-once right. Depending on the port right, you
+are allowed to send messages to the server, receive messages from it,
+or send just one single message.
+<P>
+For every port, there exists exactly one task holding the receive
+right, but there can be no or many senders. The send-once right is
+useful for clients expecting a response message. They can give a
+send-once right to the reply port along with the message. The kernel
+guarantees that at some point, a message will be received on the reply
+port (this can be a notification that the server destroyed the
+send-once right).
+<P>
+You don't need to know much about the format a message takes to be
+able to use the Mach IPC. The Mach interface generator mig hides the
+details of composing and sending a message, as well as receiving the
+reply message. To the user, it just looks like a function call, but
+in truth the message could be sent over a network to a server running
+on a different computer. The set of remote procedure calls a server
+provides is the public interface of this server.
+
+
+<H4><A HREF="#TOChow" NAME="how">How to get a port?</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Traditional Mach:
+<UL>
+ <LI>Nameserver provides ports to all registered servers.</LI>
+ <LI>The nameserver port itself is provided by Mach.</LI>
+ <LI>Like a phone book: One list.</LI>
+</UL>
+<P>
+The Hurd:
+<UL>
+ <LI>The filesystem is used as the server namespace.</LI>
+ <LI>Root directory port is inserted into each task.</LI>
+ <LI>The C library finds other ports with hurd_file_name_lookup,
+ performing a pathname resolution.</LI>
+ <LI>Like a tree of phone books.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+So how does one get a port to a server? You need something like a
+phone book for server ports, or otherwise you can only talk to
+yourself. In the original Mach system, a special nameserver is
+dedicated to that job. A task could get a port to the nameserver from
+the Mach kernel and ask it for a port (with send right) to a server
+that registered itself with the nameserver at some earlier time.
+<P>
+In the Hurd, there is no nameserver. Instead, the filesystem is used
+as the server namespace. This works because there is always a root
+filesystem in the Hurd (remember that the Hurd is a POSIX compatible
+system); this is an assumption the people who developed Mach couldn't
+make, so they had to choose a different strategy. You can use the
+function hurd_file_name_lookup, which is part of the C library, to get
+a port to the server belonging to a filename. Then you can start to
+send messages to the server in the usual way.
+
+<H4><A HREF="#TOCexa" NAME="exa">Example of <SAMP>hurd_file_name_lookup</SAMP></A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT"><PRE>
+mach_port_t identity;
+mach_port_t pwserver;
+kern_return_t err;
+
+pwserver = hurd_file_name_lookup
+ ("/servers/password");
+
+err = password_check_user (pwserver,
+ 0 /* root */, "supass",
+ &identity);
+</PRE></TD></TR></TABLE>
+<P>
+As a concrete example, the special filename
+<SAMP>/servers/password</SAMP> can be used to request a port to the
+Hurd password server, which is responsible to check user provided
+passwords.
+<P>
+(explanation of the example)
+
+<H4><A HREF="#TOCpat" NAME="pat">Pathname resolution example</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Task: Lookup /mnt/readme.txt where /mnt has a mounted filesystem.
+<UL>
+ <LI>The C library asks the root filesystem server about
+ <SAMP>/mnt/readme.txt</SAMP>.</LI>
+ <LI>The root filesystem returns a port to the mnt filesystem server
+ (matching <SAMP>/mnt</SAMP>) and the retry name
+ <SAMP>/readme.txt</SAMP>.</LI>
+ <LI>The C library asks the mnt filesystem server about
+ <SAMP>/readme.txt</SAMP>.</LI>
+ <LI>The mnt filesystem server returns a port to itself and records
+ that this port refers to the regular file
+ <SAMP>/readme.txt</SAMP>.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The C library itself does not have a full list of all available
+servers. Instead pathname resolution is used to traverse through a
+tree of servers. In fact, filesystems themselves are implemented by
+servers (let us ignore the chicken and egg problem here). So all the
+C library can do is to ask the root filesystem server about the
+filename provided by the user (assuming that the user wants to resolve
+an absolute path), using the <SAMP>dir_lookup</SAMP> RPC. If the
+filename refers to a regular file or directory on the filesystem, the
+root filesystem server just returns a port to itself and records that
+this port corresponds to the file or directory in question. But if a
+prefix of the full path matches the path of a server the root
+filesystem knows about, it returns to the C library a port to this
+server and the remaining part of the pathname that couldn't be
+resolved. The C library than has to retry and query the other server
+about the remaining path component. Eventually, the C library will
+either know that the remaining path can't be resolved by the last
+server in the list, or get a valid port to the server in question.
+
+<H4><A HREF="#TOCmap" NAME="map">Mapping the POSIX Interface</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<TABLE BORDER="0" CELLPADDING="10">
+<TR>
+<TH>Filedescriptor</TH>
+<TH>Port to server providing the file</TH>
+</TR><TR>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>fd = open(name,...)</SAMP></TD>
+<TD VALIGN="TOP"
+ALIGN="LEFT"><SAMP>dir_lookup(..,name,..,&amp;port)</SAMP><BR>
+[pathname resolution]</TD>
+</TR><TR>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>read(fd, ...)</SAMP></TD>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>io_read(port, ...)</SAMP></TD>
+</TR><TR>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>write(fd, ...)</SAMP></TD>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>io_write(port, ...)</SAMP></TD>
+</TR><TR>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>fstat(fd, ...)</SAMP></TD>
+<TD VALIGN="TOP" ALIGN="LEFT"><SAMP>io_stat(port, ...)</SAMP></TD>
+</TR><TR>
+<TD VALIGN="TOP" ALIGN="LEFT">...</TD><TD></TD>
+</TR>
+</TABLE>
+</TD></TR></TABLE>
+<P>
+It should by now be obvious that the port returned by the server can
+be used to query the files status, content and other information from
+the server, if good remote procedure calls to do that are defined and
+implemented by it. This is exactly what happens. Whenever a file is
+opened using the C libraries <SAMP>open()</SAMP> call, the C library
+uses the above pathname resolution to get a port to a server providing
+the file. Then it wraps a file descriptor around it. So in the Hurd,
+for every open file descriptor there is a port to a server providing
+this file. Many other C library calls like <SAMP>read()</SAMP> and
+<SAMP>write()</SAMP> just call a corresponding RPC using the port
+associated with the file descriptor.
+
+<H4><A HREF="#TOCfilser" NAME="filser">File System Servers</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<UL>
+ <LI>Provide file and directory services for ports (and more).</LI>
+ <LI>These ports are returned by a directory lookup.</LI>
+ <LI>Translate filesystem accesses through their root path (hence the
+ name translator).</LI>
+ <LI>The C library maps the POSIX file and directory interface (and
+ more) to RPCs to the filesystem servers ports, but also does work on
+ its own.</LI>
+ <LI>Any user can install file system servers on inodes they own.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+So we don't have a single phone book listing all servers, but rather a
+tree of servers keeping track of each other. That's really like
+calling your friend and asking for the phone number of the blond girl
+at the party yesterday. He might refer you to a friend who hopefully
+knows more about it. Then you have to retry.
+<P>
+This mechanism has huge advantages over a single nameserver. First,
+note that standard unix permissions on directories can be used to
+restrict access to a server (this requires that the filesystems
+providing those directories behave). You just have to set the
+permissions of a parent directory accordingly and provide no other way
+to get a server port.
+<P>
+But there are much deeper implications. Most of all, a pathname never
+directly refers to a file, it refers to a port of a server. That
+means that providing a regular file with static data is just one of
+the many options the server has to service requests on the file port.
+A server can also create the data dynamically. For example, a server
+associated with <SAMP>/dev/random</SAMP> can provide new random data
+on every <SAMP>io_read()</SAMP> on the port to it. A server
+associated with <SAMP>/dev/fortune</SAMP> can provide a new fortune
+cookie on every <SAMP>open()</SAMP>.
+<P>
+While a regular filesystem server will just serve the data as stored
+in a filesystem on disk, there are servers providing purely virtual
+information, or a mixture of both. It is up to the server to behave
+and provide consistent and useful data on each remote procedure call.
+If it does not, the results may not match the expectations of the user
+and confuse him.
+<P>
+A footnote from the Hurd info manual:
+<BLOCKQUOTE>
+(1) You are lost in a maze of twisty little filesystems, all
+alike....
+</BLOCKQUOTE>
+<P>
+Because a server installed in the filesystem namespace translates all
+filesystem operations that go through its root path, such a server is
+also called "active translator". You can install translators using
+the settrans command with the <SAMP>-a</SAMP> option.
+
+<H4><A HREF="#TOCact" NAME="act">Active vs Passive</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Active Translators:
+<UL>
+ <LI>"<SAMP>settrans -a /cdrom /hurd/isofs /dev/hd2</SAMP>"</LI>
+ <LI>Are running filesystem servers.</LI>
+ <LI>Are attached to the root node they translate.</LI>
+ <LI>Run as a normal process.</LI>
+ <LI>Go away with every reboot, or even time out.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+Many translator settings remain constant for a long time. It would be
+very lame to always repeat the same couple of dozens settrans calls
+manually or at boot time. So the Hurd provides a filesystem extension
+that allows to store translator settings inside the filesystem and let
+the filesystem servers do the work to start those servers on demand.
+Such translator settings are called "passive translators". A passive
+translator is really just a command line string stored in an inode of
+the filesystem. If during a pathname resolution a server encounters
+such a passive translator, and no active translator does exist already
+(for this node), it will use this string to start up a new translator
+for this inode, and then let the C library continue with the path
+resolution as described above. Passive translators are installed with
+settrans using the <SAMP>-p</SAMP> option (which is already the
+default).
+
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Passive Translators:
+<UL>
+ <LI>"<SAMP>settrans /mnt /hurd/ext2fs /dev/hd1s1</SAMP>"</LI>
+ <LI>Are stored as command strings into an inode.</LI>
+ <LI>Are used to start a new active translator if there isn't
+ one.</LI>
+ <LI>Startup is transparent to the user.</LI>
+ <LI>Startup happens the first time the server is needed.</LI>
+ <LI>Are permanent across reboots (like file data).</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+So passive translators also serve as a sort of automounting feature,
+because no manual interaction is required. The server start up is
+deferred until the service is need, and it is transparent to the user.
+<P>
+When starting up a passive translator, it will run as a normal process
+with the same user and group id as those of the underlying inode. Any
+user is allowed to install passive and active translators on inodes
+that he owns. This way the user can install new servers into the
+global namespace (for example, in his home or tmp directory) and thus
+extend the functionality of the system (recall that servers can
+implement other remote procedure calls beside those used for files and
+directories). A careful design of the trusted system servers makes
+sure that no permissions leak out.
+<P>
+In addition, users can provide their own implementations of some of
+the system servers instead the system default. For example, they can
+use their own exec server to start processes. The user specific exec
+server could for example start java programs transparently (without
+invoking the interpreter manually). This is done by setting the
+environment variable <SAMP>EXECSERVERS</SAMP>. The systems default
+exec server will evaluate this environment variable and forward the
+RPC to each of the servers listed in turn, until some server accepts
+it and takes over. The system default exec server will only do this
+if there are no security implications. (XXX There are other ways to
+start new programs than by using the system exec server. Those are
+still available.)
+<P>
+Let's take a closer look at some of the Hurd servers. It was already
+mentioned that only few system servers are mandatory for users. To
+establish your identity within the Hurd system, you have to
+communicate with the trusted systems authentication server
+<SAMP>auth</SAMP>. To put the system administrator into control over
+the system components, the process server does some global
+bookkeeping.
+<P>
+But even these servers can be ignored. However, registration with the
+authentication server is the only way to establish your identity
+towards other system servers. Likewise, only tasks registered as
+processes with the process server can make use of its services.
+
+<H4><A HREF="#TOCaut" NAME="aut">Authentication</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+A user identity is just a port to an authserver. The auth server
+stores four set of ids for it:
+<UL>
+ <LI>effective user ids</LI>
+ <LI>effective group ids</LI>
+ <LI>available user ids</LI>
+ <LI>available group ids</LI>
+</UL>
+<P>
+Basic properties:
+<UL>
+ <LI>Any of these can be empty.</LI>
+ <LI>A 0 among the user ids identifies the superuser.</LI>
+ <LI>Effective ids are used to check if the user has the
+ permission.</LI>
+ <LI>Available ids can be turned into effective ids on user
+ request.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The Hurd auth server is used to establish the identity of a user for a
+server. Such an identity (which is just a port to the auth server)
+consists of a set of effective user ids, a set of effective group ids,
+a set of available user ids and a set of available group ids. Any of
+these sets can be empty.
+
+<H4><A HREF="#TOCope" NAME="ope">Operations on authentication ports</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+The auth server provides the following operations on ports:
+<UL>
+ <LI>Merge the ids of two ports into a new one.</LI>
+ <LI>Return a new port containing a subset of the ids in a port.</LI>
+ <LI>Create a new port with arbitrary ids (superuser only).</LI>
+ <LI>Establish a trusted connection between users and servers.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+If you have two identities, you can merge them and request an identity
+consisting of the unions of the sets from the auth server. You can
+also create a new identity consisting only of subsets of an identity
+you already have. What you can't do is extending your sets, unless
+you are the superuser which is denoted by having the user id 0.
+
+<H4><A HREF="#TOCest" NAME="est">Establishing trusted connections</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<UL>
+ <LI>User provides a rendezvous port to the server (with
+ <SAMP>io_reauthenticate</SAMP>).</LI>
+ <LI>User calls <SAMP>auth_user_authenticate</SAMP> on the
+ authentication port (his identity), passing the rendezvous port.</LI>
+ <LI>Server calls <SAMP>auth_server_authenticate</SAMP> on its
+ authentication port (to a trusted auth server), passing the
+ rendezvous port and the server port.</LI>
+ <LI>If both authentication servers are the same, it can match the
+ rendezvous ports and return the server port to the user and the user
+ ids to the server.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+Finally, the auth server can establish the identity of a user for a
+server. This is done by exchanging a server port and a user identity
+if both match the same rendezvous port. The server port will be
+returned to the user, while the server is informed about the id sets
+of the user. The server can then serve or reject subsequent RPCs by
+the user on the server port, based on the identity it received from
+the auth server.
+<P>
+Anyone can write a server conforming to the auth protocol, but of
+course all system servers use a trusted system auth server to
+establish the identity of a user. If the user is not using the system
+auth server, matching the rendezvous port will fail and no server port
+will be returned to the user. Because this practically requires all
+programs to use the same auth server, the system auth server is
+minimal in every respect, and additional functionality is moved
+elsewhere, so user freedom is not unnecessarily restricted.
+
+<H4><A HREF="#TOCpas" NAME="pas">Password Server</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+The password server <SAMP>/servers/password</SAMP> runs as root and
+returns a new authentication port in exchange for a unix password.
+<P>
+The ids corresponding to the authentication port match the unix user
+and group ids.
+<P>
+Support for shadow passwords is implemented here.
+</TD></TR></TABLE>
+<P>
+The password server sits at <SAMP>/servers/password</SAMP> and runs as
+root. It can hand out ports to the auth server in exchange for a unix
+password, matching it against the password or shadow file. Several
+utilities make use of this server, so they don't need to be setuid
+root.
+
+<H4><A HREF="#TOCpro" NAME="pro">Process Server</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+The superuser must remain control over user tasks, so:
+<UL>
+ <LI>All mach tasks are associated with a PID in the system default
+ proc server.</LI>
+</UL>
+<P>
+Optionally, user tasks can store:
+<UL>
+ <LI>Their environment variables.</LI>
+ <LI>Their argument vector.</LI>
+ <LI>A port, which others can request based on the PID (like a
+ nameserver).</LI>
+</UL>
+<P>
+Also implemented in the proc server:
+<UL>
+ <LI>Sessions and process groups.</LI>
+ <LI>Global configuration not in Mach, like hostname, hostid, system
+ version.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The process server is responsible for some global bookkeeping. As
+such it has to be trusted and is not replaceable by the user.
+However, a user is not required to use any of its service. In that
+case the user will not be able to take advantage of the POSIXish
+appearance of the Hurd.
+<P>
+The Mach Tasks are not as heavy as POSIX processes. For example,
+there is no concept of process groups or sessions in Mach. The proc
+server fills in the gap. It provides a PID for all Mach tasks, and
+also stores the argument line, environment variables and other
+information about a process (if the mach tasks provide them, which is
+usually the case if you start a process with the default
+<SAMP>fork()</SAMP>/<SAMP>exec()</SAMP>). A process can also register
+a message port with the proc server, which can then be requested by
+anyone. So the proc server also functions as a nameserver using the
+process id as the name.
+<P>
+The proc server also stores some other miscellaneous information not
+provided by Mach, like the hostname, hostid and system version.
+Finally, it provides facilities to group processes and their ports
+together, as well as to convert between pids, process server ports and
+mach task ports.
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+User tasks not registering themselve with proc only have a PID assigned.
+<P>
+Users can run their own proc server in addition to the system default,
+at least for those parts of the interface that don't require superuser
+privileges.
+</TD></TR></TABLE>
+<P>
+Although the system default proc server can't be avoided (all Mach
+tasks spawned by users will get a pid assigned, so the system
+administrator can control them), users can run their own additional
+process servers if they want, implementing the features not requiring
+superuser privileges.
+
+<H4><A HREF="#TOCfilsys" NAME="filsys">Filesystems</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Store based filesystems
+<UL>
+ <LI><SAMP>ext2fs</SAMP></LI>
+ <LI><SAMP>ufs</SAMP></LI>
+ <LI><SAMP>isofs</SAMP> (iso9660, RockRidge, GNU extensions)</LI>
+ <LI><SAMP>fatfs</SAMP> (under development)</LI>
+</UL>
+<P>
+Network file systems
+<UL>
+ <LI><SAMP>nfs</SAMP></LI>
+ <LI><SAMP>ftpfs</SAMP></LI>
+</UL>
+<P>
+Miscellaneous
+<UL>
+ <LI><SAMP>hostmux</SAMP></LI>
+ <LI><SAMP>usermux</SAMP></LI>
+ <LI><SAMP>tmpfs</SAMP> (under development)</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+We already talked about translators and the file system service they
+provide. Currently, we have translators for the ext2, ufs and iso9660
+filesystems. We also have an nfs client and an ftp filesystem.
+Especially the latter is intriguing, as it provides transparent access
+to ftp servers in the filesystem. Programs can start to move away
+from implementing a plethora of network protocols, as the files are
+directly available in the filesystem through the standard POSIX file
+interface.
+
+
+<H4><A HREF="#TOCdev" NAME="dev">Developing the Hurd</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Over a dozen libraries support the development of new servers.
+<P>
+For special server types highly specialized
+libraries require only the implementation of a
+number of callback functions.
+<UL>
+ <LI>Use <SAMP>libdiskfs</SAMP> for store based filesystems.</LI>
+ <LI>Use <SAMP>libnetfs</SAMP> for network filesystems, also for
+ virtual filesystems.</LI>
+ <LI>Use <SAMP>libtrivfs</SAMP> for simple filesystems providing only
+ a single file or directory.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The Hurd server protocols are complex enough to allow for the
+implementation of a POSIX compatible system with GNU extensions.
+However, a lot of code can be shared by all or at least similar
+servers. For example, all storage based filesystems need to be able to
+read and write to a store medium splitted in blocks. The Hurd comes
+with several libraries which make it easy to implement new servers.
+Also, there are already a lot of examples of different server types in
+the Hurd. This makes writing a new server easier.
+<P>
+<SAMP>libdiskfs</SAMP> is a library that supports writing store based
+filesystems like ext2fs or ufs. It is not very useful for filesystems
+which are purely virtual, like <SAMP>/proc</SAMP> or files in
+<SAMP>/dev</SAMP>.
+<P>
+<SAMP>libnetfs</SAMP> is intended for filesystems which provide a rich
+directory hierarchy, but don't use a backing store (for example ftpfs,
+nfs).
+<P>
+<SAMP>libtrivfs</SAMP> is intended for filesystems which just provide
+a single inode or directory. Most servers which are not intended to
+provide a filesystem but other services (like
+<SAMP>/servers/password</SAMP>) use it to provide a dummy file, so
+that file operations on the servers node will not return errors. But
+it can also be used to provide meaningful data in a single file, like
+a device store or a character device.
+
+<H4><A HREF="#TOCsto" NAME="sto">Store Abstraction</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Another very useful library is libstore, which is used by all store
+based filesystems. It provides a store media abstraction. A store
+consists of a store class and a name (which itself can sometimes
+contain stores).
+<P>
+Primitive store classes:
+<UL>
+ <LI>device store like device:hd2, device:hd0s1, device:fd0</LI>
+ <LI>file store like file:/tmp/disk_image</LI>
+ <LI>task store like task:PID</LI>
+ <LI>zero store like zero:4m (like /dev/zero, of size 4 MB)</LI>
+</UL>
+</TD></TR></TABLE>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Composed store classes:
+<UL>
+ <LI>copy store like copy:zero:4m</LI>
+ <LI>gunzip/bunzip2 store like gunzip:device:fd0</LI>
+ <LI>concat store like concat:device:hd0s2:device:hd1s5</LI>
+ <LI>ileave store (RAID-0(2))</LI>
+ <LI>remap store like remap:10+20,50+:file:/tmp/blocks</LI>
+ <LI>...</LI>
+</UL>
+<P>
+Wanted: A similar abstraction for streams (based on channels), which
+can be used by network and character device servers.
+</TD></TR></TABLE>
+<P>
+<SAMP>libstore</SAMP> provides a store abstraction, which is used by
+all store based filesystems. The store is determined by a type and a
+name, but some store types modify another store rather than providing
+a new store, and thus stores can be stacked. For example, the device
+store type expects a Mach device, but the remap store expects a list
+of blocks to pick from another store, like remap:1+:device:hd2, which
+would pick all blocks from hd2 but the first one, which skipped.
+Because this functionality is provided in a library, all libstore
+using filesystems support many different store kinds, and adding a new
+store type is enough to make all store based filesystems support it.
+
+<H4><A HREF="#TOCdeb" NAME="deb">Debian GNU/Hurd</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Goal:
+<UL>
+ <LI>Provide a binary distribution of the Hurd that is easy to
+ install.</LI>
+</UL>
+<P>
+Constraints:
+<UL>
+ <LI>Use the same source packages as Debian GNU/Linux.</LI>
+ <LI>Use the same infrastructure:
+ <UL>
+ <LI>Policy</LI>
+ <LI>Archive</LI>
+ <LI>Bug tracking system</LI>
+ <LI>Release process</LI>
+ </UL></LI>
+</UL>
+<P>
+Side Goal:
+<UL>
+ <LI>Prepare Debian for the future:
+ <UL>
+ <LI>More flexibility in the base system</LI>
+ <LI>Identify dependencies on the Linux kernel</LI>
+ </UL></LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The Debian distribution of the GNU&nbsp;Hurd that I started in 1998 is
+supposed to become a complete binary distribution of the Hurd that is
+easy to install.
+
+<H4><A HREF="#TOCstabin" NAME="stabin">Status of the Debian GNU/Hurd binary archive</A></H4>
+<P>
+See
+<A HREF="http://buildd.debian.org/stats/graph.png">http://buildd.debian.org/stats/graph.png</A>
+for the most current version of the statistic.
+
+<H4><A HREF="#TOCstainf" NAME="stainf">Status of the Debian infrastructure</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Plus:
+<UL>
+ <LI>Source packages can identify build and host OS using
+ dpkg-architecture.</LI>
+</UL>
+<P>
+Minus:
+<UL>
+ <LI>The binary architecture field is insufficient.</LI>
+ <LI>The BTS has no architecture tag.</LI>
+ <LI>The policy/FHS need (small) Hurd specific extensions.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+While good compatibiity can be achieved at the source level,
+the binary packages can not always express their relationship
+to the available architectures sufficiently.
+<P>
+For example, the Linux version of makedev is binary-all, where
+a binary-all-linux relationship would be more appropriate.
+<P>
+More work has to be done here to fix the tools.
+
+<H4><A HREF="#TOCstaarc" NAME="staarc">Status of the Debian Source archive</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<UL>
+ <LI>Most packages just work.</LI>
+ <LI>Maintainers are usually responsive and cooperative.</LI>
+ <LI>Turtle, the autobuilder, crunches through the whole list right
+ now.</LI>
+</UL>
+<P>
+Common pitfalls are POSIX incompatibilities:
+<UL>
+ <LI>Upstream:
+ <UL>
+ <LI>Unconditional use of <SAMP>PATH_MAX</SAMP>
+ (<SAMP>MAXPATHLEN</SAMP>), <SAMP>MAXHOSTNAMELEN</SAMP>.</LI>
+ <LI>Unguarded use of Linux kernel features.</LI>
+ <LI>Use of legacy interfaces (<SAMP>sys_errlist</SAMP>,
+ <SAMP>termio</SAMP>).</LI>
+ </UL></LI>
+ <LI>Debian:
+ <UL>
+ <LI>Unguarded activation of extensions available with Linux.</LI>
+ <LI>Low quality patches.</LI>
+ <LI>Assuming GNU/Linux in package scripts.</LI>
+ </UL></LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+Most packages are POSIX compatible and can be compiled without
+changes on the Hurd. The maintainers of the Debian source packages
+are usually very kind, responsiver and helpful.
+<P>
+The Turtle autobuilder software (<A
+HREF="http://turtle.sourceforge.net" >http://turtle.sourceforge.net</A>)
+builds the Debian packages on the Hurd automatically.
+
+<H4><A HREF="#TOCdebide" NAME="debide">Debian GNU/Hurd: Good idea, bad idea?</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Upstream benefits:
+<UL>
+ <LI>Software packages become more portable.</LI>
+</UL>
+<P>
+Debian benefits:
+<UL>
+ <LI>Debian becomes more portable.</LI>
+ <LI>Maintainers learn about portability and other systems.</LI>
+ <LI>Debian gets a lot of public recognition.</LI>
+</UL>
+<P>
+GNU/Hurd benefits:
+<UL>
+ <LI>Large software base.</LI>
+ <LI>Great infrastructure.</LI>
+ <LI>Nice community to partner with.</LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+The sheet lists the advantages of all groups involved.
+
+<H4><A HREF="#TOCend" NAME="end">End</A></H4>
+<TABLE BORDER="1" CELLPADDING="5" WIDTH="100%"><TR><TD VALIGN="TOP" ALIGN="LEFT">
+<P>
+Join us at
+<UL>
+ <LI><A HREF="http://hurd.gnu.org/" >http://hurd.gnu.org/</A></LI>
+ <LI><A HREF="http://www.debian.org/ports/hurd"
+ >http://www.debian.org/ports/hurd</A></LI>
+ <LI><A HREF="http://www.hurdfr.org"
+ >http://www.hurdfr.org</A></LI>
+</UL>
+</TD></TR></TABLE>
+<P>
+List of contacts.