path: root/doc/mach.texi

\input texinfo                  @c -*- Texinfo -*-
@setfilename mach.info
@settitle The GNU Mach Reference Manual
@setchapternewpage odd

@comment Tell install-info what to do.
@dircategory Kernel
@direntry
* GNUMach: (mach).              Using and programming the GNU Mach microkernel.
@end direntry

@c Should have a glossary.
@c Unify some of our indices.
@syncodeindex pg cp
@syncodeindex vr fn
@syncodeindex tp fn

@c Get the Mach version we are documenting.
@include version.texi
@set EDITION 0.4
@c @set ISBN X-XXXXXX-XX-X

@ifinfo
This file documents the GNU Mach microkernel.

This is edition @value{EDITION}, last updated on @value{UPDATED}, of @cite{The
GNU Mach Reference Manual}, for version @value{VERSION}.

Copyright @copyright{} 2001, 2002, 2006, 2007 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``Free Software Needs Free Documentation'' and
``GNU Lesser General Public License'', the Front-Cover texts being (a)
(see below), and with the Back-Cover Texts being (b) (see below).  A
copy of the license is included in the section entitled ``GNU Free
Documentation License''.

(a) The FSF's Front-Cover Text is:

     A GNU Manual

(b) The FSF's Back-Cover Text is:

     You have freedom to copy and modify this GNU Manual, like GNU
     software.  Copies published by the Free Software Foundation raise
     funds for GNU development.

This work is based on manual pages under the following copyright and license:

@noindent
Mach Operating System@*
Copyright @copyright{} 1991,1990 Carnegie Mellon University@*
All Rights Reserved.

Permission to use, copy, modify and distribute this software and its
documentation is hereby granted, provided that both the copyright
notice and this permission notice appear in all copies of the
software, derivative works or modified versions, and any portions
thereof, and that both notices appear in supporting documentation.

CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
@end ifinfo

@iftex
@shorttitlepage The GNU Mach Reference Manual
@end iftex
@titlepage
@center @titlefont{The GNU Mach}
@sp 1
@center @titlefont{Reference Manual}
@sp 2
@center Marcus Brinkmann
@center with
@center Gordon Matzigkeit, Gibran Hasnaoui,
@center Robert V. Baron, Richard P. Draves, Mary R. Thompson, Joseph S. Barrera
@sp 3
@center Edition @value{EDITION}
@sp 1
@center last updated @value{UPDATED}
@sp 1
@center for version @value{VERSION}
@page
@vskip 0pt plus 1filll
Copyright @copyright{} 2001, 2002, 2006, 2007 Free Software Foundation, Inc.
@c  @sp 2
@c  Published by the Free Software Foundation @*
@c  59 Temple Place -- Suite 330, @*
@c  Boston, MA 02111-1307 USA @*
@c  ISBN @value{ISBN} @*

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2 or
any later version published by the Free Software Foundation; with the
Invariant Sections being ``Free Software Needs Free Documentation'' and
``GNU Lesser General Public License'', the Front-Cover texts being (a)
(see below), and with the Back-Cover Texts being (b) (see below).  A
copy of the license is included in the section entitled ``GNU Free
Documentation License''.

(a) The FSF's Front-Cover Text is:

     A GNU Manual

(b) The FSF's Back-Cover Text is:

     You have freedom to copy and modify this GNU Manual, like GNU
     software.  Copies published by the Free Software Foundation raise
     funds for GNU development.

This work is based on manual pages under the following copyright and license:

@noindent
Mach Operating System@*
Copyright @copyright{} 1991,1990 Carnegie Mellon University@*
All Rights Reserved.

Permission to use, copy, modify and distribute this software and its
documentation is hereby granted, provided that both the copyright
notice and this permission notice appear in all copies of the
software, derivative works or modified versions, and any portions
thereof, and that both notices appear in supporting documentation.

CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
@end titlepage
@c  @titlepage
@c  @finalout
@c  @title The GNU Mach Reference Manual
@c  @author Marcus Brinkmann
@c  @author Gordon Matzigkeit
@c  @author Gibran Hasnaoui

@c  @author Robert V. Baron @c (rvb)
@c  @author Richard P. Draves @c (rpd)
@c  @author Mary R. Thompson @c (mrt)
@c  @author Joseph S. Barrera @c (jsb)
@c  @c The following occur rarely in the rcs commit logs of the man pages:
@c  @c Dan Stodolsky, (danner)
@c  @c David B. Golub, (dbg)
@c  @c Terri Watson, (elf)
@c  @c Lori Iannamico, (lli) [distribution coordinator]
@c  @c Further authors of kernel_interfaces.ps:
@c  @c David Black [OSF]
@c  @c William Bolosky
@c  @c Jonathan Chew
@c  @c Alessandro Forin
@c  @c Richard F. Rashid
@c  @c Avadis Tevanian Jr.
@c  @c Michael W. Young
@c  @c See also
@c  @c http://www.cs.cmu.edu/afs/cs/project/mach/public/www/people-former.html
@page

@ifnottex
@node Top
@top Main Menu
This is edition @value{EDITION}, last updated on @value{UPDATED}, of @cite{The
GNU Mach Reference Manual}, for version @value{VERSION} of the GNU Mach
microkernel.
@end ifnottex

@menu
* Introduction::                  How to use this manual.
* Installing::                    Setting up GNU Mach on your computer.
* Bootstrap::                     Running GNU Mach on your machine.
* Inter Process Communication::   Communication between process.
* Virtual Memory Interface::      Allocating and deallocating virtual memory.
* External Memory Management::    Handling memory pages in user space.
* Threads and Tasks::             Handling of threads and tasks.
* Host Interface::                Interface to a Mach host.
* Processors and Processor Sets:: Handling processors and sets of processors.
* Device Interface::              Accessing kernel devices.
* Kernel Debugger::               How to use the built-in kernel debugger.

Appendices

* Copying::                       The GNU General Public License says how you
                                  can copy and share the GNU Mach microkernel.
* Documentation License::         This manual is under the GNU Free
                                  Documentation License.

Indices

* Concept Index::                 Index of concepts and programs.
* Function and Data Index::       Index of functions, variables and data types.


@detailmenu
 --- The Detailed Node Listing ---

Introduction

* Audience::                      The people for whom this manual is written.
* Features::                      Reasons to install and use GNU Mach.
* Overview::                      Basic architecture of the Mach microkernel.
* History::                       The story about Mach.

Installing

* Binary Distributions::          Obtaining ready-to-run GNU distributions.
* Compilation::                   Building GNU Mach from its source code.
* Configuration::                 Configuration options at compilation time.
* Cross-Compilation::             Building GNU Mach from another system.

Bootstrap

* Bootloader::                    Starting the microkernel, or other OSes.
* Modules::                       Starting the first task of the OS.

Inter Process Communication

* Major Concepts::                The concepts behind the Mach IPC system.
* Messaging Interface::           Composing, sending and receiving messages.
* Port Manipulation Interface::   Manipulating ports, port rights, port sets.

Messaging Interface

* Mach Message Call::             Sending and receiving messages.
* Message Format::                The format of Mach messages.
* Exchanging Port Rights::        Sending and receiving port rights.
* Memory::                        Passing memory regions in messages.
* Message Send::                  Sending messages.
* Message Receive::               Receiving messages.
* Atomicity::                     Atomicity of port rights.

Port Manipulation Interface

* Port Creation::                 How to create new ports and port sets.
* Port Destruction::              How to destroy ports and port sets.
* Port Names::                    How to query and manipulate port names.
* Port Rights::                   How to work with port rights.
* Ports and other Tasks::         How to move rights between tasks.
* Receive Rights::                How to work with receive rights.
* Port Sets::                     How to work with port sets.
* Request Notifications::         How to request notifications for events.
@c  * Inherited Ports::               How to work with the inherited system ports.

Virtual Memory Interface

* Memory Allocation::             Allocation of new virtual memory.
* Memory Deallocation::           Freeing unused virtual memory.
* Data Transfer::                 Reading, writing and copying memory.
* Memory Attributes::             Tweaking memory regions.
* Mapping Memory Objects::        How to map memory objects.
* Memory Statistics::             How to get statistics about memory usage.

External Memory Management

* Memory Object Server::          The basics of external memory management.
* Memory Object Creation::        How new memory objects are created.
* Memory Object Termination::     How memory objects are terminated.
* Memory Objects and Data::       Data transfer to and from memory objects.
* Memory Object Locking::         How memory objects are locked.
* Memory Object Attributes::      Manipulating attributes of memory objects.
* Default Memory Manager::        Setting and using the default memory manager.

Threads and Tasks

* Thread Interface::              Manipulating threads.
* Task Interface::                Manipulating tasks.
* Profiling::                     Profiling threads and tasks.

Thread Interface

* Thread Creation::               Creating threads.
* Thread Termination::            Terminating threads.
* Thread Information::            How to get informations on threads.
* Thread Settings::               How to set threads related informations.
* Thread Execution::              How to control the thread's machine state.
* Scheduling::                    Operations on thread scheduling.
* Thread Special Ports::          How to handle the thread's special ports.
* Exceptions::                    Managing exceptions.

Scheduling

* Thread Priority::               Changing the priority of a thread.
* Hand-Off Scheduling::           Switch to a new thread.
* Scheduling Policy::             Setting the scheduling policy.

Task Interface

* Task Creation::                 Creating tasks.
* Task Termination::              Terminating tasks.
* Task Information::              Informations on tasks.
* Task Execution::                Thread scheduling in a task.
* Task Special Ports::            How to get and set the task's special ports.
* Syscall Emulation::             How to emulate system calls.

Host Interface

* Host Ports::                    Ports representing a host.
* Host Information::              Query information about a host.
* Host Time::                     Functions to query manipulate the host time.
* Host Reboot::                   Rebooting the system.

Processors and Processor Sets

* Processor Set Interface::       How to work with processor sets.
* Processor Interface::           How to work with individual processors.

Processor Set Interface

* Processor Set Ports::           Ports representing a processor set.
* Processor Set Access::          How the processor sets are accessed.
* Processor Set Creation::        How new processor sets are created.
* Processor Set Destruction::     How processor sets are destroyed.
* Tasks and Threads on Sets::     Assigning tasks or threads to processor sets.
* Processor Set Priority::        Specifying the priority of a processor set.
* Processor Set Policy::          Changing the processor set policies.
* Processor Set Info::            Obtaining information about a processor set.

Processor Interface

* Hosted Processors::             Getting a list of all processors on a host.
* Processor Control::             Starting, stopping, controlling processors.
* Processors and Sets::           Combining processors into processor sets.
* Processor Info::                Obtaining information on processors.

Device Interface

* Device Open::                   Opening hardware devices.
* Device Close::                  Closing hardware devices.
* Device Read::                   Reading data from the device.
* Device Write::                  Writing data to the device.
* Device Map::                    Mapping devices into virtual memory.
* Device Status::                 Querying and manipulating a device.
* Device Filter::                 Filtering packets arriving on a device.

Kernel Debugger

* Operation::                     Basic architecture of the kernel debugger.
* Commands::                      Available commands in the kernel debugger.
* Variables::                     Access of variables from the kernel debugger.
* Expressions::                   Usage of expressions in the kernel debugger.

Documentation License

* GNU Free Documentation License:: The GNU Free Documentation License.
* CMU License::                   The CMU license applies to the original Mach
                                  kernel and its documentation.

@end detailmenu
@end menu


@node Introduction
@chapter Introduction

GNU Mach is the microkernel of the GNU Project.  It is the base of the
operating system, and provides its functionality to the Hurd servers,
the GNU C Library and all user applications.  The microkernel itself
does not provide much functionality of the system, just enough to make
it possible for the Hurd servers and the C library to implement the missing
features you would expect from a POSIX compatible operating system.

@menu
* Audience::                      The people for whom this manual is written.
* Features::                      Reasons to install and use GNU Mach.
* Overview::                      Basic architecture of the Mach microkernel.
* History::                       The story about Mach.
@end menu


@node Audience
@section Audience

This manual is designed to be useful to everybody who is interested in
using, administering, or programming the Mach microkernel.

If you are an end-user and you are looking for help on running the Mach
kernel, the first few chapters of this manual describe the essential
parts of installing and using the kernel in the GNU operating system.

The rest of this manual is a technical discussion of the Mach
programming interface and its implementation, and would not be helpful
until you want to learn how to extend the system or modify the kernel.

This manual is organized according to the subsystems of Mach, and each
chapter begins with descriptions of conceptual ideas that are related to
that subsystem.  If you are a programmer and want to learn more about,
say, the Mach IPC subsystem, you can skip to the IPC chapter
(@pxref{Inter Process Communication}), and read about the related
concepts and interface definitions.


@node Features
@section Features

GNU Mach is not the most advanced microkernel known to the planet,
nor is it the fastest or smallest, but it has a rich set of interfaces and
some features which make it useful as the base of the Hurd system.

@table @asis
@item it's free software
Anybody can use, modify, and redistribute it under the terms of the GNU
General Public License (@pxref{Copying}).  GNU Mach is part of the GNU
system, which is a complete operating system licensed under the GPL.

@item it's built to survive
As a microkernel, GNU Mach doesn't implement a lot of the features
commonly found in an operating system, but only the bare minimum
that is required to implement a full operating system on top of it.
This means that a lot of the operating system code is maintained outside
of GNU Mach, and while this code may go through a complete redesign, the
code of the microkernel can remain comparatively stable.

@item it's scalable
Mach is particularly well suited for SMP and network cluster techniques.
Thread support is provided at the kernel level, and the kernel itself
takes advantage of that.  Network transparency at the IPC level makes
resources of the system available across machine boundaries (with NORMA
IPC, currently not available in GNU Mach).

@item it exists
The Mach microkernel is real software that works Right Now.
It is not a research or a proposal.  You don't have to wait at all
before you can start using and developing it.  Mach has been used in
many operating systems in the past, usually as the base for a single
UNIX server.  In the GNU system, Mach is the base of a functional
multi-server operating system, the Hurd.
@end table


@node Overview
@section Overview

@c This paragraph by Gordon Matzigkeit from the Hurd manual.
An operating system kernel provides a framework for programs to share a
computer's hardware resources securely and efficiently.  This requires
that the programs are separated and protected from each other.  To make
running multiple programs in parallel useful, there also needs to be a
facility for programs to exchange information by communication.

The Mach microkernel provides abstractions of the underlying hardware
resources like devices and memory.  It organizes the running programs
into tasks and threads (points of execution in the tasks).  In addition,
Mach provides a rich interface for inter-process communication.

What Mach does not provide is a POSIX compatible programming interface.
In fact, it has no understanding of file systems, POSIX process semantics,
network protocols and many more.  All this is implemented in tasks
running on top of the microkernel.  In the GNU operating system, the Hurd
servers and the C library share the responsibility to implement the POSIX
interface, and the additional interfaces which are specific to the GNU
system.


@node History
@section History

XXX A few lines about the history of Mach here.


@node Installing
@chapter Installing

Before you can use the Mach microkernel in your system you'll need to install
it and all components you want to use with it, e.g. the rest of the operating
system.  You also need a bootloader to load the kernel from the storage
medium and run it when the computer is started.

GNU Mach is only available for Intel i386-compatible architectures
(such as the Pentium) currently.  If you have a different architecture
and want to run the GNU Mach microkernel, you will need to port the
kernel and all other software of the system to your machine's architecture.
Porting is an involved process which requires considerable programming skills,
and it is not recommended for the faint-of-heart.
If you have the talent and desire to do a port, contact
@email{bug-hurd@@gnu.org} in order to coordinate the effort.

@menu
* Binary Distributions::          Obtaining ready-to-run GNU distributions.
* Compilation::                   Building GNU Mach from its source code.
* Configuration::                 Configuration options at compile time.
* Cross-Compilation::             Building GNU Mach from another system.
@end menu


@node Binary Distributions
@section Binary Distributions

By far the easiest and best way to install GNU Mach and the operating
system is to obtain a GNU binary distribution.  The GNU operating
system consists of GNU Mach, the Hurd, the C library and many applications.
Without the GNU operating system, you will only have a microkernel, which
is not very useful by itself, without the other programs.

Building the whole operating system takes a huge effort, and you are well
advised to not do it yourself, but to get a binary distribution of the
GNU operating system.  The distribution also includes a binary of the
GNU Mach microkernel.

Information on how to obtain the GNU system can be found in the Hurd
info manual.


@node Compilation
@section Compilation

If you already have a running GNU system, and only want to recompile
the kernel, for example to select a different set of included hardware
drivers, you can easily do this.  You need the GNU C compiler and
MIG, the Mach interface generator, which both come in their own
packages.

Building and installing the kernel is as easy as with any other GNU
software package.  The configure script is used to configure the source
and set the compile time options.  The compilation is done by running:

@example
make
@end example

To install the kernel and its header files, just enter the command:

@example
make install
@end example

This will install the kernel as @file{EXEC_PREFIX/boot/gnumach}, the header
files into @file{PREFIX/include/}, the list of message ids as
@file{PREFIX/share/msgids/gnumach.msgids} and the documentation into
@file{PREFIX/share/info/}.

Note that there is also a way to only install the header and documentation
files without having to actually build the whole package: run @command{make
install-data} after having ran @command{configure} to do so.  (This is needed
for bootstrapping a cross compiler and similar procedures.)

@node Configuration
@section Configuration

See the following tables for the options can be passed to the
@command{configure} script as command line arguments to control what components
are built into the kernel, how certain things are configured and so on.

See the top-level @file{INSTALL} file for information about generic
@command{configure} options, like under which paths to install the package's
components.  It also describes how to control the process by setting
environment variables.

The file @file{i386/README-Drivers} has some i386-specific information for
device drivers.  You should only need to consult this file in case a device
driver is not working for you.

@subsection Table of configure switches not related to device drivers

@table @code
@item --enable-kdb
In-kernel debugger.  This is only useful if you actually anticipate debugging
the kernel.  It is not enabled by default because it adds considerably to the
unpageable memory footprint of the kernel.  @xref{Kernel Debugger}.
@end table

@subsection Turning device drivers on or off

Each device driver has an associated configure switch.  The following table
indicates whether a device driver is enabled by default or not.  It also gives
--- if possible at all --- the configure switches to use for disabling or
enabling device drivers, in case you're not satisfied with the default choices.
You can specify @samp{--enable-device-drivers=WHICH} (where WHICH on
@samp{ix86-at} must be one of @samp{default}, @samp{qemu}, @samp{none}) to
preset a certain subset of all available device drivers.
@samp{--enable-device-drivers} is sugar for
@samp{--enable-device-drivers=default} (and is the implicit default
nevertheless) and @samp{--disable-device-drivers} is short for
@samp{--enable-device-drivers=none}.  @samp{qemu} will include only the set of
device drivers that is useful when using the resulting kernel binary to drive a
Hurd system in the @acronym{QEMU} system emulator.  This is only useful for
reducing the kernel build time and the kernel image size.

@subsection What the configure switches do

Each configure switch has two effects.  First, it defines a @acronym{CPP}
symbol that turns on or off the hooks that autoconfigure the device and add it
to the list of available devices.  Second, it adds the source code for the
driver to a make variable so that the code for the driver is compiled and
linked into the kernel.  Also follow this route to find the file(s) which are
implementing a certain device driver.

@subsection Table of configure switches related to device drivers

(@samp{%d} in the following denotes a unit number, starting with @samp{0}.)

@table @code
@item --disable-kmsg
Kernel message device @samp{kmsg}.

@item --enable-lpr
Parallel port device driver for the @samp{lpr%d} devices.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-floppy
PC floppy disk controller device driver for the @samp{fd%d} devices.  On
@samp{ix86-at} enabled by @samp{default} and for @samp{qemu}.

@item --enable-ide
IDE controller device driver for the @samp{hd%d} and @samp{hd%ds%d} (disks and
their partitions) devices.  On @samp{ix86-at} enabled by @samp{default} and for
@samp{qemu}.
@end table

The following options control drivers for various SCSI controller.  SCSI
devices are named @samp{sd%d} and @samp{sd%ds$d} (disks and their partitions)
or @samp{cd%d} (CD ROMs).

@table @code
@item --enable-advansys
AdvanSys SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-buslogic
BusLogic SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-flashpoint
Only meaningful in conjunction with the above BusLogic SCSI controller device
driver.  Enable the FlashPoint support.

@item --enable-u14-34f
UltraStor 14F/34F SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-ultrastor
UltraStor SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-aha152x
Adaptec AHA-152x/2825 SCSI controller device driver.  On @samp{ix86-at} enabled
by @samp{default}.

@item --enable-aha1542
Adaptec AHA-1542 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-aha1740
Adaptec AHA-1740 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-aic7xxx
Adaptec AIC7xxx SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-fdomain
Future Domain 16xx SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-in2000
Always IN 2000 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-g_NCR5380
Generic NCR5380/53c400 SCSI controller device driver.

@item --enable-NCR53c406a
NCR53c406a SCSI controller device driver.

@item --enable-pas16
PAS16 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-seagate
Seagate ST02 and Future Domain TMC-8xx SCSI controller device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-t128
Trantor T128/T128F/T228 SCSI controller device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-53c78xx
NCR53C7,8xx SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-eata_dma
EATA-DMA (DPT, NEC, AT&T, SNI, AST, Olivetti, Alphatronix) SCSI controller
device driver.

@item --enable-eata_pio
EATA-PIO (old DPT PM2001, PM2012A) SCSI controller device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-wd7000
WD 7000 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-eata
EATA ISA/EISA/PCI (DPT and generic EATA/DMA-compliant boards) SCSI controller
device driver.  On @samp{ix86-at} enabled by @samp{default}.

@item --enable-am53c974
AM53/79C974 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-dtc
DTC3180/3280 SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-ncr53c8xx
NCR53C8XX, dc390w, dc390u, dc390f SCSI controller device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-tmscsim
Tekram DC-390(T) SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-ppa
IOMEGA Parallel Port ZIP drive device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-qlogicfas
Qlogic FAS SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-qlogicisp
Qlogic ISP SCSI controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-gdth
GDT SCSI Disk Array controller device driver.  On @samp{ix86-at} enabled by
@samp{default}.
@end table

The following options enable drivers for various ethernet cards.  NIC devices
are usually named @samp{eth%d}, except for the pocket adaptors.

@c GNU Mach does only autodetect one ethernet card.  To enable any further
@c cards, the source code has to be edited.
@c XXX Reference to the source code.

@table @code
@item --enable-ne
NE2000/NE1000 ISA network card device driver.  On @samp{ix86-at} enabled by
@samp{default} and for @samp{qemu}.

@item --enable-3c503
3Com 503 (Etherlink II) network card device driver.  On @samp{ix86-at} enabled
by @samp{default}.

@item --enable-3c509
3Com 509/579 (Etherlink III) network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-wd
WD80X3 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-3c501
3COM 501/Etherlink I network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-smc-ultra
SMC Ultra network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-smc-ultra32
SMC Ultra 32 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-hp-plus
HP PCLAN+ (27247B and 27252A) network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-hp
HP PCLAN (27245 and other 27xxx series) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-3c59x
3Com 590/900 series (592/595/597/900/905) "Vortex/Boomerang" network card
device driver.  On @samp{ix86-at} enabled by @samp{default}.

@item --enable-seeq8005
Seeq8005 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-hp100
HP 10/100VG PCLAN (ISA, EISA, PCI) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-ac3200
Ansel Communications EISA 3200 network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-e2100
Cabletron E21xx network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-at1700
AT1700 (Fujitsu 86965) network card device driver.  On @samp{ix86-at} enabled
by @samp{default}.

@item --enable-eth16i
ICL EtherTeam 16i/32 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-znet
Zenith Z-Note network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-eexpress
EtherExpress 16 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-eepro
EtherExpressPro network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-eepro100
Intel EtherExpressPro PCI 10+/100B/100+ network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-depca
DEPCA, DE10x, DE200, DE201, DE202, DE210, DE422 network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-ewrk3
EtherWORKS 3 (DE203, DE204, DE205) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-de4x5
DE425, DE434, DE435, DE450, DE500 network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-apricot
Apricot XEN-II on board ethernet network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-wavelan
AT&T WaveLAN & DEC RoamAbout DS network card device driver.

@item --enable-3c507
3Com 507 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-3c505
3Com 505/Etherlink II network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-de600
D-Link DE-600 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-de620
D-Link DE-620 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-sk_g16
Schneider & Koch G16 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-ni52
NI5210 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-ni65
NI6510 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-atp
AT-LAN-TEC/RealTek pocket adaptor network card device driver for the
@samp{atp%d} devices.

@item --enable-lance
AMD LANCE and PCnet (AT1500 and NE2100) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@itemx --enable-tulip
DECchip Tulip (dc21x4x) PCI network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-fmv18x
FMV-181/182/183/184 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-3c515
3Com 515 ISA Fast EtherLink network card device driver.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-pcnet32
AMD PCI PCnet32 (PCI bus NE2100 cards) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-ne2k-pci
PCI NE2000 network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-yellowfin
Packet Engines Yellowfin Gigabit-NIC network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-rtl8139
RealTek 8129/8139 (not 8019/8029!) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@itemx --enable-epic100
SMC 83c170/175 EPIC/100 (EtherPower II) network card device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-tlan
TI ThunderLAN network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-via-rhine
VIA Rhine network card device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-hamachi
Packet Engines "Hamachi" GNIC-2 Gigabit Ethernet device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-intel-gige
Intel PCI Gigabit Ethernet device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-myson803
Myson MTD803 Ethernet adapter series device driver.  On @samp{ix86-at} enabled
by @samp{default}.

@item --enable-natsemi
National Semiconductor DP8381x series PCI Ethernet device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-ns820
National Semiconductor DP8382x series PCI Ethernet device driver.  On
@samp{ix86-at} enabled by @samp{default}.

@item --enable-starfire
Adaptec Starfire network adapter device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-sundance
Sundance ST201 "Alta" PCI Ethernet device driver.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-winbond-840
Winbond W89c840 PCI Ethernet device driver.  On @samp{ix86-at} enabled by
@samp{default}.
@end table

The following options either control device drivers for supported PCMCIA
bridges or control the overall behaviour of the GNU Mach PCMCIA core.  To make
use of GNU Mach PCMCIA support you need to have the corresponding userland
applications (GNU Mach Card Services) installed.

@table @code
@item --enable-i82365
Device driver for the Intel 82365 and compatible PC Card controllers, and
Yenta-compatible PCI-to-CardBus controllers.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-pcmcia-isa
ISA bus related bits in the GNU Mach PCMCIA core.  Keeping it enabled is
generally a good idea, since it does not only have effect if your PC Card
bridge is attached to the ISA bus, but provides more (ISA) interrupts to the
Card Services for it to assign to the cards in turn.  On @samp{ix86-at} enabled
by @samp{default}.
@end table

The following options control device drivers for supported PCMCIA Ethernet
controllers.  NIC devices are usually named @samp{eth%d}.

@table @code
@item --enable-3c574_cs
PCMCIA ethernet driver for the 3Com 3c574 ``RoadRunner''.  On @samp{ix86-at}
enabled by @samp{default}.

@item --enable-3c589_cs
Driver for the 3Com 3c589 PCMCIA card.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-axnet_cs
Driver for the Asix AX88190-based PCMCIA cards.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-fmvj18x_cs
Driver for PCMCIA cards with the fmvj18x chipset.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-nmclan_cs
Driver for the New Media Ethernet LAN PCMCIA cards.  On @samp{ix86-at} enabled
by @samp{default}.

@item --enable-pcnet_cs
Driver for NS8390-based PCMCIA cards.  This driver supports the D-Link DE-650
and Linksys EthernetCard cards, the newer D-Link and Linksys combo cards,
Accton EN2212 cards, the RPTI EP400, and the PreMax PE-200 in non-shared-memory
mode, and the IBM Credit Card Adapter, the NE4100, the Thomas Conrad ethernet
card, and the Kingston KNE-PCM/x in shared-memory mode.  It will also handle
the Socket EA card in either mode.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-smc91c92_cs
Driver for SMC91c92-based PCMCIA cards.  On @samp{ix86-at} enabled by
@samp{default}.

@item --enable-xirc2ps_cs
Driver for Xircom CreditCard and Realport PCMCIA ethernet adapters.  On
@samp{ix86-at} enabled by @samp{default}.
@end table

The following options control device drivers for supported PCMCIA Wireless LAN
network controllers.  NIC devices are usually named @samp{eth%d}.

Please mind, that you need to have some userland applications (the GNU Mach
Wireless Tools) installed, in order to make use of these devices.

@table @code
@item --enable-orinoco_cs
Driver for the Hermes or Prism 2 chipset based PCMCIA wireless adapters, with
Lucent/Agere, Intersil or Symbol firmware.  This driver is suitable for PCMCIA
wireless adapters, such as the Lucent WavelanIEEE/Orinoco cards and their OEM
(Cabletron/EnteraSys RoamAbout 802.11, ELSA Airlancer, Melco Buffalo and
others).  It should also be usable on various Prism II based cards such as the
Linksys, D-Link and Farallon Skyline. It should also work on Symbol cards such
as the 3Com AirConnect and Ericsson WLAN.  On @samp{ix86-at} enabled by
@samp{default}.
@end table


@node Cross-Compilation
@section Cross-Compilation

Another way to install the kernel is to use an existing operating system
in order to compile the kernel binary.
This is called @dfn{cross-compiling}, because it is done between two
different platforms.  If the pre-built kernels are not working for
you, and you can't ask someone to compile a custom kernel for your
machine, this is your last chance to get a kernel that boots on your
hardware.

Luckily, the kernel does have light dependencies.  You don't even
need a cross compiler if your build machine has a compiler and is
the same architecture as the system you want to run GNU Mach on.

You need a cross-mig, though.

XXX More info needed.


@node Bootstrap
@chapter Bootstrap

Bootstrapping@footnote{The term @dfn{bootstrapping} refers to a Dutch
legend about a boy who was able to fly by pulling himself up by his
bootstraps.  In computers, this term refers to any process where a
simple system activates a more complicated system.} is the procedure by
which your machine loads the microkernel and transfers control to the
operating system.


@menu
* Bootloader::                    Starting the microkernel, or other OSes.
* Modules::                       Starting the first task of the OS.
@end menu

@node Bootloader
@section Bootloader

The @dfn{bootloader} is the first software that runs on your machine.
Many hardware architectures have a very simple startup routine which
reads a very simple bootloader from the beginning of the internal hard
disk, then transfers control to it.  Other architectures have startup
routines which are able to understand more of the contents of the hard
disk, and directly start a more advanced bootloader.

@cindex GRUB
@cindex GRand Unified Bootloader
@dfn{GRUB}@footnote{The GRand Unified Bootloader, available
from @uref{http://gnu.org/software/grub/}.} is the GNU bootloader.
GRUB provides advanced functionality, and is capable of loading several
different kernels (such as Mach, Linux, DOS, and the *BSD family).
@xref{Top, , Introduction, grub, GRUB Manual}.

GNU Mach conforms to the Multiboot specification which defines an
interface between the bootloader and the components that run very early
at startup.  GNU Mach can be started by any bootloader which supports
the multiboot standard.  After the bootloader loaded the kernel image to
a designated address in the system memory, it jumps into the startup
code of the kernel.  This code initializes the kernel and detects the
available hardware devices.  Afterwards, the first system task is
started.  @xref{Top, , Overview, multiboot, Multiboot Specification}.


@node Modules
@section Modules
@pindex serverboot

This is outdated.

Because the microkernel does not provide filesystem support and other
features necessary to load the first system task from a storage medium,
the first task is loaded by the bootloader as a module to a specified
address.  In the GNU system, this first program is the @code{serverboot}
executable.  GNU Mach inserts the host control port and the device
master port into this task and appends the port numbers to the command
line before executing it.

The @code{serverboot} program is responsible for loading and executing
the rest of the Hurd servers.  Rather than containing specific
instructions for starting the Hurd, it follows general steps given in a
user-supplied boot script.

XXX More about boot scripts.


@node Inter Process Communication
@chapter Inter Process Communication

This chapter describes the details of the Mach IPC system.  First the
actual calls concerned with sending and receiving messages are
discussed, then the details of the port system are described in detail.

@menu
* Major Concepts::                The concepts behind the Mach IPC system.
* Messaging Interface::           Composing, sending and receiving messages.
* Port Manipulation Interface::   Manipulating ports, port rights, port sets.
@end menu


@node Major Concepts
@section Major Concepts
@cindex interprocess communication (IPC)
@cindex IPC (interprocess communication)
@cindex communication between tasks
@cindex remote procedure calls (RPC)
@cindex RPC (remote procedure calls)
@cindex messages

The Mach kernel provides message-oriented, capability-based interprocess
communication.  The interprocess communication (IPC) primitives
efficiently support many different styles of interaction, including
remote procedure calls (RPC), object-oriented distributed programming,
streaming of data, and sending very large amounts of data.

The IPC primitives operate on three abstractions: messages, ports, and
port sets.  User tasks access all other kernel services and abstractions
via the IPC primitives.

The message primitives let tasks send and receive messages.  Tasks send
messages to ports.  Messages sent to a port are delivered reliably
(messages may not be lost) and are received in the order in which they
were sent.  Messages contain a fixed-size header and a variable amount
of typed data following the header.  The header describes the
destination and size of the message.

The IPC implementation makes use of the VM system to efficiently
transfer large amounts of data.  The message body can contain the
address of a region in the sender's address space which should be
transferred as part of the message.  When a task receives a message
containing an out-of-line region of data, the data appears in an unused
portion of the receiver's address space.  This transmission of
out-of-line data is optimized so that sender and receiver share the
physical pages of data copy-on-write, and no actual data copy occurs
unless the pages are written.  Regions of memory up to the size of a
full address space may be sent in this manner.

Ports hold a queue of messages.  Tasks operate on a port to send and
receive messages by exercising capabilities for the port.  Multiple
tasks can hold send capabilities, or rights, for a port.  Tasks can also
hold send-once rights, which grant the ability to send a single message.
Only one task can hold the receive capability, or receive right, for a
port.  Port rights can be transferred between tasks via messages.  The
sender of a message can specify in the message body that the message
contains a port right.  If a message contains a receive right for a
port, then the receive right is removed from the sender of the message
and the right is transferred to the receiver of the message.  While the
receive right is in transit, tasks holding send rights can still send
messages to the port, and they are queued until a task acquires the
receive right and uses it to receive the messages.

Tasks can receive messages from ports and port sets.  The port set
abstraction allows a single thread to wait for a message from any of
several ports.  Tasks manipulate port sets with a capability, or
port-set right, which is taken from the same space as the port
capabilities.  The port-set right may not be transferred in a message.
A port set holds receive rights, and a receive operation on a port set
blocks waiting for a message sent to any of the constituent ports.  A
port may not belong to more than one port set, and if a port is a member
of a port set, the holder of the receive right can't receive directly
from the port.

Port rights are a secure, location-independent way of naming ports.  The
port queue is a protected data structure, only accessible via the
kernel's exported message primitives.  Rights are also protected by the
kernel; there is no way for a malicious user task to guess a port name
and send a message to a port to which it shouldn't have access.  Port
rights do not carry any location information.  When a receive right for
a port moves from task to task, and even between tasks on different
machines, the send rights for the port remain unchanged and continue to
function.

@node Messaging Interface
@section Messaging Interface

This section describes how messages are composed, sent and received
within the Mach IPC system.

@menu
* Mach Message Call::             Sending and receiving messages.
* Message Format::                The format of Mach messages.
* Exchanging Port Rights::        Sending and receiving port rights.
* Memory::                        Passing memory regions in messages.
* Message Send::                  Sending messages.
* Message Receive::               Receiving messages.
* Atomicity::                     Atomicity of port rights.
@end menu


@node Mach Message Call
@subsection Mach Message Call

To use the @code{mach_msg} call, you can include the header files
@file{mach/port.h} and @file{mach/message.h}.

@deftypefun mach_msg_return_t mach_msg (@w{mach_msg_header_t *@var{msg}}, @w{mach_msg_option_t @var{option}}, @w{mach_msg_size_t @var{send_size}}, @w{mach_msg_size_t @var{rcv_size}}, @w{mach_port_t @var{rcv_name}}, @w{mach_msg_timeout_t @var{timeout}}, @w{mach_port_t @var{notify}})
The @code{mach_msg} function is used to send and receive messages.  Mach
messages contain typed data, which can include port rights and
references to large regions of memory.

@var{msg} is the address of a buffer in the caller's address space.
Message buffers should be aligned on long-word boundaries.  The message
options @var{option} are bit values, combined with bitwise-or.  One or
both of @code{MACH_SEND_MSG} and @code{MACH_RCV_MSG} should be used.
Other options act as modifiers.  When sending a message, @var{send_size}
specifies the size of the message buffer.  Otherwise zero should be
supplied.  When receiving a message, @var{rcv_size} specifies the size
of the message buffer.  Otherwise zero should be supplied.  When
receiving a message, @var{rcv_name} specifies the port or port set.
Otherwise @code{MACH_PORT_NULL} should be supplied.  When using the
@code{MACH_SEND_TIMEOUT} and @code{MACH_RCV_TIMEOUT} options,
@var{timeout} specifies the time in milliseconds to wait before giving
up.  Otherwise @code{MACH_MSG_TIMEOUT_NONE} should be supplied.  When
using the @code{MACH_SEND_NOTIFY}, @code{MACH_SEND_CANCEL}, and
@code{MACH_RCV_NOTIFY} options, @var{notify} specifies the port used for
the notification.  Otherwise @code{MACH_PORT_NULL} should be supplied.

If the option argument is @code{MACH_SEND_MSG}, it sends a message.  The
@var{send_size} argument specifies the size of the message to send.  The
@code{msgh_remote_port} field of the message header specifies the
destination of the message.

If the option argument is @code{MACH_RCV_MSG}, it receives a message.
The @var{rcv_size} argument specifies the size of the message buffer
that will receive the message; messages larger than @var{rcv_size} are
not received.  The @var{rcv_name} argument specifies the port or port
set from which to receive.

If the option argument is @code{MACH_SEND_MSG|MACH_RCV_MSG}, then
@code{mach_msg} does both send and receive operations.  If the send
operation encounters an error (any return code other than
@code{MACH_MSG_SUCCESS}), then the call returns immediately without
attempting the receive operation.  Semantically the combined call is
equivalent to separate send and receive calls, but it saves a system
call and enables other internal optimizations.

If the option argument specifies neither @code{MACH_SEND_MSG} nor
@code{MACH_RCV_MSG}, then @code{mach_msg} does nothing.

Some options, like @code{MACH_SEND_TIMEOUT} and @code{MACH_RCV_TIMEOUT},
share a supporting argument.  If these options are used together, they
make independent use of the supporting argument's value.
@end deftypefun

@deftp {Data type} mach_msg_timeout_t
This is a @code{natural_t} used by the timeout mechanism.  The units are
milliseconds.  The value to be used when there is no timeout is
@code{MACH_MSG_TIMEOUT_NONE}.
@end deftp


@node Message Format
@subsection Message Format
@cindex message format
@cindex format of a message
@cindex composing messages
@cindex message composition

A Mach message consists of a fixed size message header, a
@code{mach_msg_header_t}, followed by zero or more data items.  Data
items are typed.  Each item has a type descriptor followed by the actual
data (or the address of the data, for out-of-line memory regions).

The following data types are related to Mach ports:

@deftp {Data type} mach_port_t
The @code{mach_port_t} data type is an unsigned integer type which
represents a port name in the task's port name space.  In GNU Mach, this
is an @code{unsigned int}.
@end deftp

@c This is defined elsewhere.
@c  @deftp {Data type} mach_port_seqno_t
@c  The @code{mach_port_seqno_t} data type is an unsigned integer type which
@c  represents a sequence number of a message.  In GNU Mach, this is an
@c  @code{unsigned int}.
@c  @end deftp

The following data types are related to Mach messages:

@deftp {Data type} mach_msg_bits_t
The @code{mach_msg_bits_t} data type is an @code{unsigned int} used to
store various flags for a message.
@end deftp

@deftp {Data type} mach_msg_size_t
The @code{mach_msg_size_t} data type is an @code{unsigned int} used to
store the size of a message.
@end deftp

@deftp {Data type} mach_msg_id_t
The @code{mach_msg_id_t} data type is an @code{integer_t} typically used to
convey a function or operation id for the receiver.
@end deftp

@deftp {Data type} mach_msg_header_t
This structure is the start of every message in the Mach IPC system.  It
has the following members:

@table @code
@item mach_msg_bits_t msgh_bits
The @code{msgh_bits} field has the following bits defined, all other
bits should be zero:

@table @code
@item MACH_MSGH_BITS_REMOTE_MASK
@itemx MACH_MSGH_BITS_LOCAL_MASK
The remote and local bits encode @code{mach_msg_type_name_t} values that
specify the port rights in the @code{msgh_remote_port} and
@code{msgh_local_port} fields.  The remote value must specify a send or
send-once right for the destination of the message.  If the local value
doesn't specify a send or send-once right for the message's reply port,
it must be zero and msgh_local_port must be @code{MACH_PORT_NULL}.

@item MACH_MSGH_BITS_COMPLEX
The complex bit must be specified if the message body contains port
rights or out-of-line memory regions.  If it is not specified, then the
message body carries no port rights or memory, no matter what the type
descriptors may seem to indicate.
@end table

@code{MACH_MSGH_BITS_REMOTE} and @code{MACH_MSGH_BITS_LOCAL} macros
return the appropriate @code{mach_msg_type_name_t} values, given a
@code{msgh_bits} value.  The @code{MACH_MSGH_BITS} macro constructs a
value for @code{msgh_bits}, given two @code{mach_msg_type_name_t}
values.

@item mach_msg_size_t msgh_size
The @code{msgh_size} field in the header of a received message contains
the message's size.  The message size, a byte quantity, includes the
message header, type descriptors, and in-line data.  For out-of-line
memory regions, the message size includes the size of the in-line
address, not the size of the actual memory region.  There are no
arbitrary limits on the size of a Mach message, the number of data items
in a message, or the size of the data items.

@item mach_port_t msgh_remote_port
The @code{msgh_remote_port} field specifies the destination port of the
message.  The field must carry a legitimate send or send-once right for
a port.

@item mach_port_t msgh_local_port
The @code{msgh_local_port} field specifies an auxiliary port right,
which is conventionally used as a reply port by the recipient of the
message.  The field must carry a send right, a send-once right,
@code{MACH_PORT_NULL}, or @code{MACH_PORT_DEAD}.

@item mach_port_seqno_t msgh_seqno
The @code{msgh_seqno} field provides a sequence number for the message.
It is only valid in received messages; its value in sent messages is
overwritten.
@c XXX The "MESSAGE RECEIVE" section discusses message sequence numbers.

@item mach_msg_id_t msgh_id
The @code{mach_msg} call doesn't use the @code{msgh_id} field, but it
conventionally conveys an operation or function id.
@end table
@end deftp

@deftypefn Macro mach_msg_bits_t MACH_MSGH_BITS (@w{mach_msg_type_name_t @var{remote}}, @w{mach_msg_type_name_t @var{local}})
This macro composes two @code{mach_msg_type_name_t} values that specify
the port rights in the @code{msgh_remote_port} and
@code{msgh_local_port} fields of a @code{mach_msg} call into an
appropriate @code{mach_msg_bits_t} value.
@end deftypefn

@deftypefn Macro mach_msg_type_name_t MACH_MSGH_BITS_REMOTE (@w{mach_msg_bits_t @var{bits}})
This macro extracts the @code{mach_msg_type_name_t} value for the remote
port right in a @code{mach_msg_bits_t} value.
@end deftypefn

@deftypefn Macro mach_msg_type_name_t MACH_MSGH_BITS_LOCAL (@w{mach_msg_bits_t @var{bits}})
This macro extracts the @code{mach_msg_type_name_t} value for the local
port right in a @code{mach_msg_bits_t} value.
@end deftypefn

@deftypefn Macro mach_msg_bits_t MACH_MSGH_BITS_PORTS (@w{mach_msg_bits_t @var{bits}})
This macro extracts the @code{mach_msg_bits_t} component consisting of
the @code{mach_msg_type_name_t} values for the remote and local port
right in a @code{mach_msg_bits_t} value.
@end deftypefn

@deftypefn Macro mach_msg_bits_t MACH_MSGH_BITS_OTHER (@w{mach_msg_bits_t @var{bits}})
This macro extracts the @code{mach_msg_bits_t} component consisting of
everything except the @code{mach_msg_type_name_t} values for the remote
and local port right in a @code{mach_msg_bits_t} value.
@end deftypefn

Each data item has a type descriptor, a @code{mach_msg_type_t} or a
@code{mach_msg_type_long_t}.  The @code{mach_msg_type_long_t} type
descriptor allows larger values for some fields.  The
@code{msgtl_header} field in the long descriptor is only used for its
inline, longform, and deallocate bits.

@deftp {Data type} mach_msg_type_name_t
This is an @code{unsigned int} and can be used to hold the
@code{msgt_name} component of the @code{mach_msg_type_t} and
@code{mach_msg_type_long_t} structure.
@end deftp

@deftp {Data type} mach_msg_type_size_t
This is an @code{unsigned int} and can be used to hold the
@code{msgt_size} component of the @code{mach_msg_type_t} and
@code{mach_msg_type_long_t} structure.
@end deftp

@deftp {Data type} mach_msg_type_number_t
This is an @code{natural_t} and can be used to hold the
@code{msgt_number} component of the @code{mach_msg_type_t} and
@code{mach_msg_type_long_t} structure.
@c XXX This is used for the size of arrays, too.  Mmh?
@end deftp

@deftp {Data type} mach_msg_type_t
This structure has the following members:

@table @code
@item unsigned int msgt_name : 8
The @code{msgt_name} field specifies the data's type.  The following
types are predefined:

@table @code
@item MACH_MSG_TYPE_UNSTRUCTURED
@item MACH_MSG_TYPE_BIT
@item MACH_MSG_TYPE_BOOLEAN
@item MACH_MSG_TYPE_INTEGER_16
@item MACH_MSG_TYPE_INTEGER_32
@item MACH_MSG_TYPE_CHAR
@item MACH_MSG_TYPE_BYTE
@item MACH_MSG_TYPE_INTEGER_8
@item MACH_MSG_TYPE_REAL
@item MACH_MSG_TYPE_STRING
@item MACH_MSG_TYPE_STRING_C
@item MACH_MSG_TYPE_PORT_NAME
@end table

The following predefined types specify port rights, and receive special
treatment.  The next section discusses these types in detail.  The type
@c XXX cross ref
@code{MACH_MSG_TYPE_PORT_NAME} describes port right names, when no
rights are being transferred, but just names.  For this purpose, it
should be used in preference to @code{MACH_MSG_TYPE_INTEGER_32}.

@table @code
@item MACH_MSG_TYPE_MOVE_RECEIVE
@item MACH_MSG_TYPE_MOVE_SEND
@item MACH_MSG_TYPE_MOVE_SEND_ONCE
@item MACH_MSG_TYPE_COPY_SEND
@item MACH_MSG_TYPE_MAKE_SEND
@item MACH_MSG_TYPE_MAKE_SEND_ONCE
@end table

@item msgt_size : 8
The @code{msgt_size} field specifies the size of each datum, in bits.  For
example, the msgt_size of @code{MACH_MSG_TYPE_INTEGER_32} data is 32.

@item msgt_number : 12
The @code{msgt_number} field specifies how many data elements comprise
the data item.  Zero is a legitimate number.

The total length specified by a type descriptor is @w{@code{(msgt_size *
msgt_number)}}, rounded up to an integral number of bytes.  In-line data
is then padded to an integral number of long-words.  This ensures that
type descriptors always start on long-word boundaries.  It implies that
message sizes are always an integral multiple of a long-word's size.

@item msgt_inline : 1
The @code{msgt_inline} bit specifies, when @code{FALSE}, that the data
actually resides in an out-of-line region.  The address of the memory
region (a @code{vm_offset_t} or @code{vm_address_t}) follows the type
descriptor in the message body.  The @code{msgt_name}, @code{msgt_size},
and @code{msgt_number} fields describe the memory region, not the
address.

@item msgt_longform : 1
The @code{msgt_longform} bit specifies, when @code{TRUE}, that this type
descriptor is a @code{mach_msg_type_long_t} instead of a
@code{mach_msg_type_t}.  The @code{msgt_name}, @code{msgt_size}, and
@code{msgt_number} fields should be zero.  Instead, @code{mach_msg} uses
the following @code{msgtl_name}, @code{msgtl_size}, and
@code{msgtl_number} fields.

@item msgt_deallocate : 1
The @code{msgt_deallocate} bit is used with out-of-line regions.  When
@code{TRUE}, it specifies that the memory region should be deallocated
from the sender's address space (as if with @code{vm_deallocate}) when
the message is sent.

@item msgt_unused : 1
The @code{msgt_unused} bit should be zero.
@end table
@end deftp

@deftypefn Macro boolean_t MACH_MSG_TYPE_PORT_ANY (mach_msg_type_name_t type)
This macro returns @code{TRUE} if the given type name specifies a port
type, otherwise it returns @code{FALSE}.
@end deftypefn

@deftypefn Macro boolean_t MACH_MSG_TYPE_PORT_ANY_SEND (mach_msg_type_name_t type)
This macro returns @code{TRUE} if the given type name specifies a port
type with a send or send-once right, otherwise it returns @code{FALSE}.
@end deftypefn

@deftypefn Macro boolean_t MACH_MSG_TYPE_PORT_ANY_RIGHT (mach_msg_type_name_t type)
This macro returns @code{TRUE} if the given type name specifies a port
right type which is moved, otherwise it returns @code{FALSE}.
@end deftypefn

@deftp {Data type} mach_msg_type_long_t
This structure has the following members:

@table @code
@item mach_msg_type_t msgtl_header
Same meaning as @code{msgt_header}.
@c XXX cross ref

@item unsigned short msgtl_name
Same meaning as @code{msgt_name}.

@item unsigned short msgtl_size
Same meaning as @code{msgt_size}.

@item unsigned int msgtl_number
Same meaning as @code{msgt_number}.
@end table
@end deftp


@node Exchanging Port Rights
@subsection Exchanging Port Rights
@cindex sending port rights
@cindex receiving port rights
@cindex moving port rights

Each task has its own space of port rights.  Port rights are named with
positive integers.  Except for the reserved values
@w{@code{MACH_PORT_NULL (0)}@footnote{In the Hurd system, we don't make
the assumption that @code{MACH_PORT_NULL} is zero and evaluates to
false, but rather compare port names to @code{MACH_PORT_NULL}
explicitly}} and @w{@code{MACH_PORT_DEAD (~0)}}, this is a full 32-bit
name space.  When the kernel chooses a name for a new right, it is free
to pick any unused name (one which denotes no right) in the space.

There are five basic kinds of rights: receive rights, send rights,
send-once rights, port-set rights, and dead names.  Dead names are not
capabilities.  They act as place-holders to prevent a name from being
otherwise used.

A port is destroyed, or dies, when its receive right is deallocated.
When a port dies, send and send-once rights for the port turn into dead
names.  Any messages queued at the port are destroyed, which deallocates
the port rights and out-of-line memory in the messages.

Tasks may hold multiple user-references for send rights and dead names.
When a task receives a send right which it already holds, the kernel
increments the right's user-reference count.  When a task deallocates a
send right, the kernel decrements its user-reference count, and the task
only loses the send right when the count goes to zero.

Send-once rights always have a user-reference count of one, although a
port can have multiple send-once rights, because each send-once right
held by a task has a different name.  In contrast, when a task holds
send rights or a receive right for a port, the rights share a single
name.

A message body can carry port rights; the @code{msgt_name}
(@code{msgtl_name}) field in a type descriptor specifies the type of
port right and how the port right is to be extracted from the caller.
The values @code{MACH_PORT_NULL} and @code{MACH_PORT_DEAD} are always
valid in place of a port right in a message body.  In a sent message,
the following @code{msgt_name} values denote port rights:

@table @code
@item MACH_MSG_TYPE_MAKE_SEND
The message will carry a send right, but the caller must supply a
receive right.  The send right is created from the receive right, and
the receive right's make-send count is incremented.

@item MACH_MSG_TYPE_COPY_SEND
The message will carry a send right, and the caller should supply a send
right.  The user reference count for the supplied send right is not
changed.  The caller may also supply a dead name and the receiving task
will get @code{MACH_PORT_DEAD}.

@item MACH_MSG_TYPE_MOVE_SEND
The message will carry a send right, and the caller should supply a send
right.  The user reference count for the supplied send right is
decremented, and the right is destroyed if the count becomes zero.
Unless a receive right remains, the name becomes available for
recycling.  The caller may also supply a dead name, which loses a user
reference, and the receiving task will get @code{MACH_PORT_DEAD}.

@item MACH_MSG_TYPE_MAKE_SEND_ONCE
The message will carry a send-once right, but the caller must supply a
receive right.  The send-once right is created from the receive right.

@item MACH_MSG_TYPE_MOVE_SEND_ONCE
The message will carry a send-once right, and the caller should supply a
send-once right.  The caller loses the supplied send-once right.  The
caller may also supply a dead name, which loses a user reference, and
the receiving task will get @code{MACH_PORT_DEAD}.

@item MACH_MSG_TYPE_MOVE_RECEIVE
The message will carry a receive right, and the caller should supply a
receive right.  The caller loses the supplied receive right, but retains
any send rights with the same name.
@end table

If a message carries a send or send-once right, and the port dies while
the message is in transit, then the receiving task will get
@code{MACH_PORT_DEAD} instead of a right.  The following
@code{msgt_name} values in a received message indicate that it carries
port rights:

@table @code
@item MACH_MSG_TYPE_PORT_SEND
This name is an alias for @code{MACH_MSG_TYPE_MOVE_SEND}.  The message
carried a send right.  If the receiving task already has send and/or
receive rights for the port, then that name for the port will be reused.
Otherwise, the new right will have a new name.  If the task already has
send rights, it gains a user reference for the right (unless this would
cause the user-reference count to overflow).  Otherwise, it acquires the
send right, with a user-reference count of one.

@item MACH_MSG_TYPE_PORT_SEND_ONCE
This name is an alias for @code{MACH_MSG_TYPE_MOVE_SEND_ONCE}.  The
message carried a send-once right.  The right will have a new name.

@item MACH_MSG_TYPE_PORT_RECEIVE
This name is an alias for @code{MACH_MSG_TYPE_MOVE_RECEIVE}.  The
message carried a receive right.  If the receiving task already has send
rights for the port, then that name for the port will be reused.
Otherwise, the right will have a new name.  The make-send count of the
receive right is reset to zero, but the port retains other attributes
like queued messages, extant send and send-once rights, and requests for
port-destroyed and no-senders notifications.
@end table

When the kernel chooses a new name for a port right, it can choose any
name, other than @code{MACH_PORT_NULL} and @code{MACH_PORT_DEAD}, which
is not currently being used for a port right or dead name.  It might
choose a name which at some previous time denoted a port right, but is
currently unused.


@node Memory
@subsection Memory
@cindex sending memory
@cindex receiving memory

A message body can contain the address of a region in the sender's
address space which should be transferred as part of the message.  The
message carries a logical copy of the memory, but the kernel uses VM
techniques to defer any actual page copies.  Unless the sender or the
receiver modifies the data, the physical pages remain shared.

An out-of-line transfer occurs when the data's type descriptor specifies
@code{msgt_inline} as @code{FALSE}.  The address of the memory region (a
@code{vm_offset_t} or @code{vm_address_t}) should follow the type
descriptor in the message body.  The type descriptor and the address
contribute to the message's size (@code{send_size}, @code{msgh_size}).
The out-of-line data does not contribute to the message's size.

The name, size, and number fields in the type descriptor describe the
type and length of the out-of-line data, not the in-line address.
Out-of-line memory frequently requires long type descriptors
(@code{mach_msg_type_long_t}), because the @code{msgt_number} field is
too small to describe a page of 4K bytes.

Out-of-line memory arrives somewhere in the receiver's address space as
new memory.  It has the same inheritance and protection attributes as
newly @code{vm_allocate}'d memory.  The receiver has the responsibility
of deallocating (with @code{vm_deallocate}) the memory when it is no
longer needed.  Security-conscious receivers should exercise caution
when using out-of-line memory from untrustworthy sources, because the
memory may be backed by an unreliable memory manager.

Null out-of-line memory is legal.  If the out-of-line region size is
zero (for example, because @code{msgtl_number} is zero), then the
region's specified address is ignored.  A received null out-of-line
memory region always has a zero address.

Unaligned addresses and region sizes that are not page multiples are
legal.  A received message can also contain memory with unaligned
addresses and funny sizes.  In the general case, the first and last
pages in the new memory region in the receiver do not contain only data
from the sender, but are partly zero.@footnote{Sending out-of-line
memory with a non-page-aligned address, or a size which is not a page
multiple, works but with a caveat.  The extra bytes in the first and
last page of the received memory are not zeroed, so the receiver can
peek at more data than the sender intended to transfer.  This might be a
security problem for the sender.}  The received address points to the
start of the data in the first page.  This possibility doesn't
complicate deallocation, because @code{vm_deallocate} does the right
thing, rounding the start address down and the end address up to
deallocate all arrived pages.

Out-of-line memory has a deallocate option, controlled by the
@code{msgt_deallocate} bit.  If it is @code{TRUE} and the out-of-line
memory region is not null, then the region is implicitly deallocated
from the sender, as if by @code{vm_deallocate}.  In particular, the
start and end addresses are rounded so that every page overlapped by the
memory region is deallocated.  The use of @code{msgt_deallocate}
effectively changes the memory copy into a memory movement.  In a
received message, @code{msgt_deallocate} is @code{TRUE} in type
descriptors for out-of-line memory.

Out-of-line memory can carry port rights.


@node Message Send
@subsection Message Send
@cindex sending messages

The send operation queues a message to a port.  The message carries a
copy of the caller's data.  After the send, the caller can freely modify
the message buffer or the out-of-line memory regions and the message
contents will remain unchanged.

Message delivery is reliable and sequenced.  Messages are not lost, and
messages sent to a port, from a single thread, are received in the order
in which they were sent.

If the destination port's queue is full, then several things can happen.
If the message is sent to a send-once right (@code{msgh_remote_port}
carries a send-once right), then the kernel ignores the queue limit and
delivers the message.  Otherwise the caller blocks until there is room
in the queue, unless the @code{MACH_SEND_TIMEOUT} or
@code{MACH_SEND_NOTIFY} options are used.  If a port has several blocked
senders, then any of them may queue the next message when space in the
queue becomes available, with the proviso that a blocked sender will not
be indefinitely starved.

These options modify @code{MACH_SEND_MSG}.  If @code{MACH_SEND_MSG} is
not also specified, they are ignored.

@table @code
@item MACH_SEND_TIMEOUT
The timeout argument should specify a maximum time (in milliseconds) for
the call to block before giving up.@footnote{If MACH_SEND_TIMEOUT is
used without MACH_SEND_INTERRUPT, then the timeout duration might not be
accurate.  When the call is interrupted and automatically retried, the
original timeout is used.  If interrupts occur frequently enough, the
timeout interval might never expire.}  If the message can't be queued
before the timeout interval elapses, then the call returns
@code{MACH_SEND_TIMED_OUT}.  A zero timeout is legitimate.

@item MACH_SEND_NOTIFY
The notify argument should specify a receive right for a notify port.
If the send were to block, then instead the message is queued,
@code{MACH_SEND_WILL_NOTIFY} is returned, and a msg-accepted
notification is requested.  If @code{MACH_SEND_TIMEOUT} is also
specified, then @code{MACH_SEND_NOTIFY} doesn't take effect until the
timeout interval elapses.

With @code{MACH_SEND_NOTIFY}, a task can forcibly queue to a send right
one message at a time.  A msg-accepted notification is sent to the
notify port when another message can be forcibly queued.  If an attempt
is made to use @code{MACH_SEND_NOTIFY} before then, the call returns a
@code{MACH_SEND_NOTIFY_IN_PROGRESS} error.

The msg-accepted notification carries the name of the send right.  If
the send right is deallocated before the msg-accepted notification is
generated, then the msg-accepted notification carries the value
@code{MACH_PORT_NULL}.  If the destination port is destroyed before the
notification is generated, then a send-once notification is generated
instead.

@item MACH_SEND_INTERRUPT
If specified, the @code{mach_msg} call will return
@code{MACH_SEND_INTERRUPTED} if a software interrupt aborts the call.
Otherwise, the send operation will be retried.

@item MACH_SEND_CANCEL
The notify argument should specify a receive right for a notify port.
If the send operation removes the destination port right from the
caller, and the removed right had a dead-name request registered for it,
and notify is the notify port for the dead-name request, then the
dead-name request may be silently canceled (instead of resulting in a
port-deleted notification).

This option is typically used to cancel a dead-name request made with
the @code{MACH_RCV_NOTIFY} option.  It should only be used as an optimization.
@end table

The send operation can generate the following return codes.  These
return codes imply that the call did nothing:

@table @code
@item MACH_SEND_MSG_TOO_SMALL
The specified send_size was smaller than the minimum size for a message.

@item MACH_SEND_NO_BUFFER
A resource shortage prevented the kernel from allocating a message
buffer.

@item MACH_SEND_INVALID_DATA
The supplied message buffer was not readable.

@item MACH_SEND_INVALID_HEADER
The @code{msgh_bits} value was invalid.

@item MACH_SEND_INVALID_DEST
The @code{msgh_remote_port} value was invalid.

@item MACH_SEND_INVALID_REPLY
The @code{msgh_local_port} value was invalid.

@item MACH_SEND_INVALID_NOTIFY
When using @code{MACH_SEND_CANCEL}, the notify argument did not denote a
valid receive right.
@end table

These return codes imply that some or all of the message was destroyed:

@table @code
@item MACH_SEND_INVALID_MEMORY
The message body specified out-of-line data that was not readable.

@item MACH_SEND_INVALID_RIGHT
The message body specified a port right which the caller didn't possess.

@item MACH_SEND_INVALID_TYPE
A type descriptor was invalid.

@item MACH_SEND_MSG_TOO_SMALL
The last data item in the message ran over the end of the message.
@end table

These return codes imply that the message was returned to the caller
with a pseudo-receive operation:

@table @code
@item MACH_SEND_TIMED_OUT
The timeout interval expired.

@item MACH_SEND_INTERRUPTED
A software interrupt occurred.

@item MACH_SEND_INVALID_NOTIFY
When using @code{MACH_SEND_NOTIFY}, the notify argument did not denote a
valid receive right.

@item MACH_SEND_NO_NOTIFY
A resource shortage prevented the kernel from setting up a msg-accepted
notification.

@item MACH_SEND_NOTIFY_IN_PROGRESS
A msg-accepted notification was already requested, and hasn't yet been
generated.
@end table

These return codes imply that the message was queued:

@table @code
@item MACH_SEND_WILL_NOTIFY
The message was forcibly queued, and a msg-accepted notification was
requested.

@item MACH_MSG_SUCCESS
The message was queued.
@end table

Some return codes, like @code{MACH_SEND_TIMED_OUT}, imply that the
message was almost sent, but could not be queued.  In these situations,
the kernel tries to return the message contents to the caller with a
pseudo-receive operation.  This prevents the loss of port rights or
memory which only exist in the message.  For example, a receive right
which was moved into the message, or out-of-line memory sent with the
deallocate bit.

The pseudo-receive operation is very similar to a normal receive
operation.  The pseudo-receive handles the port rights in the message
header as if they were in the message body.  They are not reversed.
After the pseudo-receive, the message is ready to be resent.  If the
message is not resent, note that out-of-line memory regions may have
moved and some port rights may have changed names.

The pseudo-receive operation may encounter resource shortages.  This is
similar to a @code{MACH_RCV_BODY_ERROR} return code from a receive
operation.  When this happens, the normal send return codes are
augmented with the @code{MACH_MSG_IPC_SPACE}, @code{MACH_MSG_VM_SPACE},
@code{MACH_MSG_IPC_KERNEL}, and @code{MACH_MSG_VM_KERNEL} bits to
indicate the nature of the resource shortage.

The queueing of a message carrying receive rights may create a circular
loop of receive rights and messages, which can never be received.  For
example, a message carrying a receive right can be sent to that receive
right.  This situation is not an error, but the kernel will
garbage-collect such loops, destroying the messages and ports involved.


@node Message Receive
@subsection Message Receive

The receive operation dequeues a message from a port.  The receiving
task acquires the port rights and out-of-line memory regions carried in
the message.

The @code{rcv_name} argument specifies a port or port set from which to
receive.  If a port is specified, the caller must possess the receive
right for the port and the port must not be a member of a port set.  If
no message is present, then the call blocks, subject to the
@code{MACH_RCV_TIMEOUT} option.

If a port set is specified, the call will receive a message sent to any
of the member ports.  It is permissible for the port set to have no
member ports, and ports may be added and removed while a receive from
the port set is in progress.  The received message can come from any of
the member ports which have messages, with the proviso that a member
port with messages will not be indefinitely starved.  The
@code{msgh_local_port} field in the received message header specifies
from which port in the port set the message came.

The @code{rcv_size} argument specifies the size of the caller's message
buffer.  The @code{mach_msg} call will not receive a message larger than
@code{rcv_size}.  Messages that are too large are destroyed, unless the
@code{MACH_RCV_LARGE} option is used.

The destination and reply ports are reversed in a received message
header.  The @code{msgh_local_port} field names the destination port,
from which the message was received, and the @code{msgh_remote_port}
field names the reply port right.  The bits in @code{msgh_bits} are also
reversed.  The @code{MACH_MSGH_BITS_LOCAL} bits have the value
@code{MACH_MSG_TYPE_PORT_SEND} if the message was sent to a send right,
and the value @code{MACH_MSG_TYPE_PORT_SEND_ONCE} if was sent to a
send-once right.  The @code{MACH_MSGH_BITS_REMOTE} bits describe the
reply port right.

A received message can contain port rights and out-of-line memory.  The
@code{msgh_local_port} field does not receive a port right; the act of
receiving the message destroys the send or send-once right for the
destination port.  The msgh_remote_port field does name a received port
right, the reply port right, and the message body can carry port rights
and memory if @code{MACH_MSGH_BITS_COMPLEX} is present in msgh_bits.
Received port rights and memory should be consumed or deallocated in
some fashion.

In almost all cases, @code{msgh_local_port} will specify the name of a
receive right, either @code{rcv_name} or if @code{rcv_name} is a port
set, a member of @code{rcv_name}.  If other threads are concurrently
manipulating the receive right, the situation is more complicated.  If
the receive right is renamed during the call, then
@code{msgh_local_port} specifies the right's new name.  If the caller
loses the receive right after the message was dequeued from it, then
@code{mach_msg} will proceed instead of returning
@code{MACH_RCV_PORT_DIED}.  If the receive right was destroyed, then
@code{msgh_local_port} specifies @code{MACH_PORT_DEAD}.  If the receive
right still exists, but isn't held by the caller, then
@code{msgh_local_port} specifies @code{MACH_PORT_NULL}.

Received messages are stamped with a sequence number, taken from the
port from which the message was received.  (Messages received from a
port set are stamped with a sequence number from the appropriate member
port.)  Newly created ports start with a zero sequence number, and the
sequence number is reset to zero whenever the port's receive right moves
between tasks.  When a message is dequeued from the port, it is stamped
with the port's sequence number and the port's sequence number is then
incremented.  The dequeue and increment operations are atomic, so that
multiple threads receiving messages from a port can use the
@code{msgh_seqno} field to reconstruct the original order of the
messages.

These options modify @code{MACH_RCV_MSG}.  If @code{MACH_RCV_MSG} is not
also specified, they are ignored.

@table @code
@item MACH_RCV_TIMEOUT
The timeout argument should specify a maximum time (in milliseconds) for
the call to block before giving up.@footnote{If MACH_RCV_TIMEOUT is used
without MACH_RCV_INTERRUPT, then the timeout duration might not be
accurate.  When the call is interrupted and automatically retried, the
original timeout is used.  If interrupts occur frequently enough, the
timeout interval might never expire.}  If no message arrives before the
timeout interval elapses, then the call returns
@code{MACH_RCV_TIMED_OUT}.  A zero timeout is legitimate.

@item MACH_RCV_NOTIFY
The notify argument should specify a receive right for a notify port.
If receiving the reply port creates a new port right in the caller, then
the notify port is used to request a dead-name notification for the new
port right.

@item MACH_RCV_INTERRUPT
If specified, the @code{mach_msg} call will return
@code{MACH_RCV_INTERRUPTED} if a software interrupt aborts the call.
Otherwise, the receive operation will be retried.

@item MACH_RCV_LARGE
If the message is larger than @code{rcv_size}, then the message remains
queued instead of being destroyed.  The call returns
@code{MACH_RCV_TOO_LARGE} and the actual size of the message is returned
in the @code{msgh_size} field of the message header.
@end table

The receive operation can generate the following return codes.  These
return codes imply that the call did not dequeue a message:

@table @code
@item MACH_RCV_INVALID_NAME
The specified @code{rcv_name} was invalid.

@item MACH_RCV_IN_SET
The specified port was a member of a port set.

@item MACH_RCV_TIMED_OUT
The timeout interval expired.

@item MACH_RCV_INTERRUPTED
A software interrupt occurred.

@item MACH_RCV_PORT_DIED
The caller lost the rights specified by @code{rcv_name}.

@item MACH_RCV_PORT_CHANGED
@code{rcv_name} specified a receive right which was moved into a port
set during the call.

@item MACH_RCV_TOO_LARGE
When using @code{MACH_RCV_LARGE}, and the message was larger than
@code{rcv_size}.  The message is left queued, and its actual size is
returned in the @code{msgh_size} field of the message buffer.
@end table

These return codes imply that a message was dequeued and destroyed:

@table @code
@item MACH_RCV_HEADER_ERROR
A resource shortage prevented the reception of the port rights in the
message header.

@item MACH_RCV_INVALID_NOTIFY
When using @code{MACH_RCV_NOTIFY}, the notify argument did not denote a
valid receive right.

@item MACH_RCV_TOO_LARGE
When not using @code{MACH_RCV_LARGE}, a message larger than
@code{rcv_size} was dequeued and destroyed.
@end table

In these situations, when a message is dequeued and then destroyed, the
reply port and all port rights and memory in the message body are
destroyed.  However, the caller receives the message's header, with all
fields correct, including the destination port but excepting the reply
port, which is @code{MACH_PORT_NULL}.

These return codes imply that a message was received:

@table @code
@item MACH_RCV_BODY_ERROR
A resource shortage prevented the reception of a port right or
out-of-line memory region in the message body.  The message header,
including the reply port, is correct.  The kernel attempts to transfer
all port rights and memory regions in the body, and only destroys those
that can't be transferred.

@item MACH_RCV_INVALID_DATA
The specified message buffer was not writable.  The calling task did
successfully receive the port rights and out-of-line memory regions in
the message.

@item MACH_MSG_SUCCESS
A message was received.
@end table

Resource shortages can occur after a message is dequeued, while
transferring port rights and out-of-line memory regions to the receiving
task.  The @code{mach_msg} call returns @code{MACH_RCV_HEADER_ERROR} or
@code{MACH_RCV_BODY_ERROR} in this situation.  These return codes always
carry extra bits (bitwise-ored) that indicate the nature of the resource
shortage:

@table @code
@item MACH_MSG_IPC_SPACE
There was no room in the task's IPC name space for another port name.

@item MACH_MSG_VM_SPACE
There was no room in the task's VM address space for an out-of-line
memory region.

@item MACH_MSG_IPC_KERNEL
A kernel resource shortage prevented the reception of a port right.

@item MACH_MSG_VM_KERNEL
A kernel resource shortage prevented the reception of an out-of-line
memory region.
@end table

If a resource shortage prevents the reception of a port right, the port
right is destroyed and the caller sees the name @code{MACH_PORT_NULL}.
If a resource shortage prevents the reception of an out-of-line memory
region, the region is destroyed and the caller receives a zero address.
In addition, the @code{msgt_size} (@code{msgtl_size}) field in the
data's type descriptor is changed to zero.  If a resource shortage
prevents the reception of out-of-line memory carrying port rights, then
the port rights are always destroyed if the memory region can not be
received.  A task never receives port rights or memory regions that it
isn't told about.


@node Atomicity
@subsection Atomicity

The @code{mach_msg} call handles port rights in a message header
atomically.  Port rights and out-of-line memory in a message body do not
enjoy this atomicity guarantee.  The message body may be processed
front-to-back, back-to-front, first out-of-line memory then port rights,
in some random order, or even atomically.

For example, consider sending a message with the destination port
specified as @code{MACH_MSG_TYPE_MOVE_SEND} and the reply port specified
as @code{MACH_MSG_TYPE_COPY_SEND}.  The same send right, with one
user-reference, is supplied for both the @code{msgh_remote_port} and
@code{msgh_local_port} fields.  Because @code{mach_msg} processes the
message header atomically, this succeeds.  If @code{msgh_remote_port}
were processed before @code{msgh_local_port}, then @code{mach_msg} would
return @code{MACH_SEND_INVALID_REPLY} in this situation.

On the other hand, suppose the destination and reply port are both
specified as @code{MACH_MSG_TYPE_MOVE_SEND}, and again the same send
right with one user-reference is supplied for both.  Now the send
operation fails, but because it processes the header atomically,
mach_msg can return either @code{MACH_SEND_INVALID_DEST} or
@code{MACH_SEND_INVALID_REPLY}.

For example, consider receiving a message at the same time another
thread is deallocating the destination receive right.  Suppose the reply
port field carries a send right for the destination port.  If the
deallocation happens before the dequeuing, then the receiver gets
@code{MACH_RCV_PORT_DIED}.  If the deallocation happens after the
receive, then the @code{msgh_local_port} and the @code{msgh_remote_port}
fields both specify the same right, which becomes a dead name when the
receive right is deallocated.  If the deallocation happens between the
dequeue and the receive, then the @code{msgh_local_port} and
@code{msgh_remote_port} fields both specify @code{MACH_PORT_DEAD}.
Because the header is processed atomically, it is not possible for just
one of the two fields to hold @code{MACH_PORT_DEAD}.

The @code{MACH_RCV_NOTIFY} option provides a more likely example.
Suppose a message carrying a send-once right reply port is received with
@code{MACH_RCV_NOTIFY} at the same time the reply port is destroyed.  If
the reply port is destroyed first, then @code{msgh_remote_port}
specifies @code{MACH_PORT_DEAD} and the kernel does not generate a
dead-name notification.  If the reply port is destroyed after it is
received, then @code{msgh_remote_port} specifies a dead name for which
the kernel generates a dead-name notification.  It is not possible to
receive the reply port right and have it turn into a dead name before
the dead-name notification is requested; as part of the message header
the reply port is received atomically.


@node Port Manipulation Interface
@section Port Manipulation Interface

This section describes the interface to create, destroy and manipulate
ports, port rights and port sets.

@cindex IPC space port
@cindex port representing an IPC space
@deftp {Data type} ipc_space_t
This is a @code{task_t} (and as such a @code{mach_port_t}), which holds
a port name associated with a port that represents an IPC space in the
kernel.  An IPC space is used by the kernel to manage the port names and
rights available to a task.  The IPC space doesn't get a port name of
its own.  Instead the port name of the task containing the IPC space is
used to name the IPC space of the task (as is indicated by the fact that
the type of @code{ipc_space_t} is actually @code{task_t}).

The IPC spaces of tasks are the only ones accessible outside of
the kernel.
@end deftp

@menu
* Port Creation::                 How to create new ports and port sets.
* Port Destruction::              How to destroy ports and port sets.
* Port Names::                    How to query and manipulate port names.
* Port Rights::                   How to work with port rights.
* Ports and other Tasks::         How to move rights between tasks.
* Receive Rights::                How to work with receive rights.
* Port Sets::                     How to work with port sets.
* Request Notifications::         How to request notifications for events.
@c  * Inherited Ports::               How to work with the inherited system ports.
@end menu


@node Port Creation
@subsection Port Creation

@deftypefun kern_return_t mach_port_allocate (@w{ipc_space_t @var{task}}, @w{mach_port_right_t @var{right}}, @w{mach_port_t *@var{name}})
The @code{mach_port_allocate} function creates a new right in the
specified task.  The new right's name is returned in @var{name}, which
may be any name that wasn't in use.

The @var{right} argument takes the following values:

@table @code
@item MACH_PORT_RIGHT_RECEIVE
@code{mach_port_allocate} creates a port.  The new port is not a member
of any port set.  It doesn't have any extant send or send-once rights.
Its make-send count is zero, its sequence number is zero, its queue
limit is @code{MACH_PORT_QLIMIT_DEFAULT}, and it has no queued messages.
@var{name} denotes the receive right for the new port.

@var{task} does not hold send rights for the new port, only the receive
right.  @code{mach_port_insert_right} and @code{mach_port_extract_right}
can be used to convert the receive right into a combined send/receive
right.

@item MACH_PORT_RIGHT_PORT_SET
@code{mach_port_allocate} creates a port set.  The new port set has no
members.

@item MACH_PORT_RIGHT_DEAD_NAME
@code{mach_port_allocate} creates a dead name.  The new dead name has
one user reference.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{right} was invalid, @code{KERN_NO_SPACE} if
there was no room in @var{task}'s IPC name space for another right and
@code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of memory.

The @code{mach_port_allocate} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun mach_port_t mach_reply_port ()
The @code{mach_reply_port} system call creates a reply port in the
calling task.

@code{mach_reply_port} creates a port, giving the calling task the
receive right for the port.  The call returns the name of the new
receive right.

This is very much like creating a receive right with the
@code{mach_port_allocate} call, with two differences.  First,
@code{mach_reply_port} is a system call and not an RPC (which requires a
reply port).  Second, the port created by @code{mach_reply_port} may be
optimized for use as a reply port.

The function returns @code{MACH_PORT_NULL} if a resource shortage
prevented the creation of the receive right.
@end deftypefun

@deftypefun kern_return_t mach_port_allocate_name (@w{ipc_space_t @var{task}}, @w{mach_port_right_t @var{right}}, @w{mach_port_t @var{name}})
The function @code{mach_port_allocate_name} creates a new right in the
specified task, with a specified name for the new right.  @var{name}
must not already be in use for some right, and it can't be the reserved
values @code{MACH_PORT_NULL} and @code{MACH_PORT_DEAD}.

The @var{right} argument takes the following values:

@table @code
@item MACH_PORT_RIGHT_RECEIVE
@code{mach_port_allocate_name} creates a port.  The new port is not a
member of any port set.  It doesn't have any extant send or send-once
rights.  Its make-send count is zero, its sequence number is zero, its
queue limit is @code{MACH_PORT_QLIMIT_DEFAULT}, and it has no queued
messages.  @var{name} denotes the receive right for the new port.

@var{task} does not hold send rights for the new port, only the receive
right.  @code{mach_port_insert_right} and @code{mach_port_extract_right}
can be used to convert the receive right into a combined send/receive
right.

@item MACH_PORT_RIGHT_PORT_SET
@code{mach_port_allocate_name} creates a port set.  The new port set has
no members.

@item MACH_PORT_RIGHT_DEAD_NAME
@code{mach_port_allocate_name} creates a new dead name.  The new dead
name has one user reference.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{right} was invalid or @var{name} was
@code{MACH_PORT_NULL} or @code{MACH_PORT_DEAD}, @code{KERN_NAME_EXISTS}
if @var{name} was already in use for a port right and
@code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of memory.

The @code{mach_port_allocate_name} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun


@node Port Destruction
@subsection Port Destruction

@deftypefun kern_return_t mach_port_deallocate (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}})
The function @code{mach_port_deallocate} releases a user reference for a
right in @var{task}'s IPC name space.  It allows a task to release a
user reference for a send or send-once right without failing if the port
has died and the right is now actually a dead name.

If @var{name} denotes a dead name, send right, or send-once right, then
the right loses one user reference.  If it only had one user reference,
then the right is destroyed.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right and
@code{KERN_INVALID_RIGHT} if @var{name} denoted an invalid right.

The @code{mach_port_deallocate} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_destroy (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}})
The function @code{mach_port_destroy} deallocates all rights denoted by
a name.  The name becomes immediately available for reuse.

For most purposes, @code{mach_port_mod_refs} and
@code{mach_port_deallocate} are preferable.

If @var{name} denotes a port set, then all members of the port set are
implicitly removed from the port set.

If @var{name} denotes a receive right that is a member of a port set,
the receive right is implicitly removed from the port set.  If there is
a port-destroyed request registered for the port, then the receive right
is not actually destroyed, but instead is sent in a port-destroyed
notification to the backup port.  If there is no registered
port-destroyed request, remaining messages queued to the port are
destroyed and extant send and send-once rights turn into dead names.  If
those send and send-once rights have dead-name requests registered, then
dead-name notifications are generated for them.

If @var{name} denotes a send-once right, then the send-once right is
used to produce a send-once notification for the port.

If @var{name} denotes a send-once, send, and/or receive right, and it
has a dead-name request registered, then the registered send-once right
is used to produce a port-deleted notification for the name.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right.

The @code{mach_port_destroy} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun


@node Port Names
@subsection Port Names

@deftypefun kern_return_t mach_port_names (@w{ipc_space_t @var{task}}, @w{mach_port_array_t *@var{names}}, @w{mach_msg_type_number_t *@var{ncount}}, @w{mach_port_type_array_t *@var{types}}, @w{mach_msg_type_number_t *@var{tcount}})
The function @code{mach_port_names} returns information about
@var{task}'s port name space.  For each name, it also returns what type
of rights @var{task} holds.  (The same information returned by
@code{mach_port_type}.)  @var{names} and @var{types} are arrays that are
automatically allocated when the reply message is received. The user
should @code{vm_deallocate} them when the data is no longer needed.

@code{mach_port_names} will return in @var{names} the names of the
ports, port sets, and dead names in the task's port name space, in no
particular order and in @var{ncount} the number of names returned.  It
will return in @var{types} the type of each corresponding name, which
indicates what kind of rights the task holds with that name.
@var{tcount} should be the same as @var{ncount}.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of memory.

The @code{mach_port_names} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_type (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_type_t *@var{ptype}})
The function @code{mach_port_type} returns information about
@var{task}'s rights for a specific name in its port name space.  The
returned @var{ptype} is a bitmask indicating what rights @var{task}
holds for the port, port set or dead name.  The bitmask is composed of
the following bits:

@table @code
@item MACH_PORT_TYPE_SEND
The name denotes a send right.

@item MACH_PORT_TYPE_RECEIVE
The name denotes a receive right.

@item MACH_PORT_TYPE_SEND_ONCE
The name denotes a send-once right.

@item MACH_PORT_TYPE_PORT_SET
The name denotes a port set.

@item MACH_PORT_TYPE_DEAD_NAME
The name is a dead name.

@item MACH_PORT_TYPE_DNREQUEST
A dead-name request has been registered for the right.

@item MACH_PORT_TYPE_MAREQUEST
A msg-accepted request for the right is pending.

@item MACH_PORT_TYPE_COMPAT
The port right was created in the compatibility mode.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid and
@code{KERN_INVALID_NAME} if @var{name} did not denote a right.

The @code{mach_port_type} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_rename (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{old_name}}, @w{mach_port_t @var{new_name}})
The function @code{mach_port_rename} changes the name by which a port,
port set, or dead name is known to @var{task}.  @var{old_name} is the
original name and @var{new_name} the new name for the port right.
@var{new_name} must not already be in use, and it can't be the
distinguished values @code{MACH_PORT_NULL} and @code{MACH_PORT_DEAD}.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{old_name} did not denote a right,
@code{KERN_INVALID_VALUE} if @var{new_name} was @code{MACH_PORT_NULL} or
@code{MACH_PORT_DEAD}, @code{KERN_NAME_EXISTS} if @code{new_name}
already denoted a right and @code{KERN_RESOURCE_SHORTAGE} if the kernel
ran out of memory.

The @code{mach_port_rename} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun


@node Port Rights
@subsection Port Rights

@deftypefun kern_return_t mach_port_get_refs (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_right_t @var{right}}, @w{mach_port_urefs_t *@var{refs}})
The function @code{mach_port_get_refs} returns the number of user
references a task has for a right.

The @var{right} argument takes the following values:
@itemize @bullet
@item @code{MACH_PORT_RIGHT_SEND}
@item @code{MACH_PORT_RIGHT_RECEIVE}
@item @code{MACH_PORT_RIGHT_SEND_ONCE}
@item @code{MACH_PORT_RIGHT_PORT_SET}
@item @code{MACH_PORT_RIGHT_DEAD_NAME}
@end itemize

If @var{name} denotes a right, but not the type of right specified, then
zero is returned.  Otherwise a positive number of user references is
returned.  Note that a name may simultaneously denote send and receive
rights.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{right} was invalid and
@code{KERN_INVALID_NAME} if @var{name} did not denote a right.

The @code{mach_port_get_refs} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_mod_refs (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_right_t @var{right}}, @w{mach_port_delta_t @var{delta}})
The function @code{mach_port_mod_refs} requests that the number of user
references a task has for a right be changed.  This results in the right
being destroyed, if the number of user references is changed to zero.
The task holding the right is @var{task}, @var{name} should denote the
specified right.  @var{right} denotes the type of right being modified.
@var{delta} is the signed change to the number of user references.

The @var{right} argument takes the following values:
@itemize @bullet
@item @code{MACH_PORT_RIGHT_SEND}
@item @code{MACH_PORT_RIGHT_RECEIVE}
@item @code{MACH_PORT_RIGHT_SEND_ONCE}
@item @code{MACH_PORT_RIGHT_PORT_SET}
@item @code{MACH_PORT_RIGHT_DEAD_NAME}
@end itemize

The number of user references for the right is changed by the amount
@var{delta}, subject to the following restrictions: port sets, receive
rights, and send-once rights may only have one user reference.  The
resulting number of user references can't be negative.  If the resulting
number of user references is zero, the effect is to deallocate the
right.  For dead names and send rights, there is an
implementation-defined maximum number of user references.

If the call destroys the right, then the effect is as described for
@code{mach_port_destroy}, with the exception that
@code{mach_port_destroy} simultaneously destroys all the rights denoted
by a name, while @code{mach_port_mod_refs} can only destroy one right.
The name will be available for reuse if it only denoted the one right.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{right} was invalid or the
user-reference count would become negative, @code{KERN_INVALID_NAME} if
@var{name} did not denote a right, @code{KERN_INVALID_RIGHT} if
@var{name} denoted a right, but not the specified right and
@code{KERN_UREFS_OVERFLOW} if the user-reference count would overflow.

The @code{mach_port_mod_refs} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun


@node Ports and other Tasks
@subsection Ports and other Tasks

@deftypefun kern_return_t mach_port_insert_right (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_t @var{right}}, @w{mach_msg_type_name_t @var{right_type}})
The function @var{mach_port_insert_right} inserts into @var{task} the
caller's right for a port, using a specified name for the right in the
target task.

The specified @var{name} can't be one of the reserved values
@code{MACH_PORT_NULL} or @code{MACH_PORT_DEAD}.  The @var{right} can't
be @code{MACH_PORT_NULL} or @code{MACH_PORT_DEAD}.

The argument @var{right_type} specifies a right to be inserted and how
that right should be extracted from the caller.  It should be a value
appropriate for @var{msgt_name}; see @code{mach_msg}. @c XXX cross ref

If @var{right_type} is @code{MACH_MSG_TYPE_MAKE_SEND},
@code{MACH_MSG_TYPE_MOVE_SEND}, or @code{MACH_MSG_TYPE_COPY_SEND}, then
a send right is inserted.  If the target already holds send or receive
rights for the port, then @var{name} should denote those rights in the
target.  Otherwise, @var{name} should be unused in the target.  If the
target already has send rights, then those send rights gain an
additional user reference.  Otherwise, the target gains a send right,
with a user reference count of one.

If @var{right_type} is @code{MACH_MSG_TYPE_MAKE_SEND_ONCE} or
@code{MACH_MSG_TYPE_MOVE_SEND_ONCE}, then a send-once right is inserted.
The name should be unused in the target.  The target gains a send-once
right.

If @var{right_type} is @code{MACH_MSG_TYPE_MOVE_RECEIVE}, then a receive
right is inserted.  If the target already holds send rights for the
port, then name should denote those rights in the target.  Otherwise,
name should be unused in the target.  The receive right is moved into
the target task.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{right} was not a port right or
@var{name} was @code{MACH_PORT_NULL} or @code{MACH_PORT_DEAD},
@code{KERN_NAME_EXISTS} if @var{name} already denoted a right,
@code{KERN_INVALID_CAPABILITY} if @var{right} was @code{MACH_PORT_NULL}
or @code{MACH_PORT_DEAD} @code{KERN_RIGHT_EXISTS} if @var{task} already
had rights for the port, with a different name,
@code{KERN_UREFS_OVERFLOW} if the user-reference count would overflow
and @code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of memory.

The @code{mach_port_insert_right} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_extract_right (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_msg_type_name_t @var{desired_type}}, @w{mach_port_t *@var{right}}, @w{mach_msg_type_name_t *@var{acquired_type}})
The function @var{mach_port_extract_right} extracts a port right from
the target @var{task} and returns it to the caller as if the task sent
the right voluntarily, using @var{desired_type} as the value of
@var{msgt_name}.  @xref{Mach Message Call}.

The returned value of @var{acquired_type} will be
@code{MACH_MSG_TYPE_PORT_SEND} if a send right is extracted,
@code{MACH_MSG_TYPE_PORT_RECEIVE} if a receive right is extracted, and
@code{MACH_MSG_TYPE_PORT_SEND_ONCE} if a send-once right is extracted.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right,
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but an invalid one,
@code{KERN_INVALID_VALUE} if @var{desired_type} was invalid.

The @code{mach_port_extract_right} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun


@node Receive Rights
@subsection Receive Rights

@deftp {Data type} mach_port_seqno_t
The @code{mach_port_seqno_t} data type is an @code{unsigned int} which
contains the sequence number of a port.
@end deftp

@deftp {Data type} mach_port_mscount_t
The @code{mach_port_mscount_t} data type is an @code{unsigned int} which
contains the make-send count for a port.
@end deftp

@deftp {Data type} mach_port_msgcount_t
The @code{mach_port_msgcount_t} data type is an @code{unsigned int} which
contains a number of messages.
@end deftp

@deftp {Data type} mach_port_rights_t
The @code{mach_port_rights_t} data type is an @code{unsigned int} which
contains a number of rights for a port.
@end deftp

@deftp {Data type} mach_port_status_t
This structure contains some status information about a port, which can
be queried with @code{mach_port_get_receive_status}.  It has the following
members:

@table @code
@item mach_port_t mps_pset
The containing port set.

@item mach_port_seqno_t mps_seqno
The sequence number.

@item mach_port_mscount_t mps_mscount
The make-send count.

@item mach_port_msgcount_t mps_qlimit
The maximum number of messages in the queue.

@item mach_port_msgcount_t mps_msgcount
The current number of messages in the queue.

@item mach_port_rights_t mps_sorights
The number of send-once rights that exist.

@item boolean_t mps_srights
@code{TRUE} if send rights exist.

@item boolean_t mps_pdrequest
@code{TRUE} if port-deleted notification is requested.

@item boolean_t mps_nsrequest
@code{TRUE} if no-senders notification is requested.
@end table
@end deftp

@deftypefun kern_return_t mach_port_get_receive_status (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_status_t *@var{status}})
The function @code{mach_port_get_receive_status} returns the current
status of the specified receive right.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right and
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but not a
receive right.

The @code{mach_port_get_receive_status} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_set_mscount (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_mscount_t @var{mscount}})
The function @code{mach_port_set_mscount} changes the make-send count of
@var{task}'s receive right named @var{name} to @var{mscount}.  All
values for @var{mscount} are valid.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right and
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but not a
receive right.

The @code{mach_port_set_mscount} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_set_qlimit (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_msgcount_t @var{qlimit}})
The function @code{mach_port_set_qlimit} changes the queue limit
@var{task}'s receive right named @var{name} to @var{qlimit}.  Valid
values for @var{qlimit} are between zero and
@code{MACH_PORT_QLIMIT_MAX}, inclusive.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right,
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but not a
receive right and @code{KERN_INVALID_VALUE} if @var{qlimit} was invalid.

The @code{mach_port_set_qlimit} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_set_seqno (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_seqno_t @var{seqno}})
The function @code{mach_port_set_seqno} changes the sequence number
@var{task}'s receive right named @var{name} to @var{seqno}.  All
sequence number values are valid.  The next message received from the
port will be stamped with the specified sequence number.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right and
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but not a
receive right.

The @code{mach_port_set_seqno} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun


@node Port Sets
@subsection Port Sets

@deftypefun kern_return_t mach_port_get_set_status (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_port_array_t *@var{members}}, @w{mach_msg_type_number_t *@var{count}})
The function @code{mach_port_get_set_status} returns the members of a
port set.  @var{members} is an array that is automatically allocated
when the reply message is received.  The user should
@code{vm_deallocate} it when the data is no longer needed.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right,
@code{KERN_INVALID_RIGHT} if @var{name} denoted a right, but not a
receive right and @code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of
memory.

The @code{mach_port_get_set_status} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun

@deftypefun kern_return_t mach_port_move_member (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{member}}, @w{mach_port_t @var{after}})
The function @var{mach_port_move_member} moves the receive right
@var{member} into the port set @var{after}.  If the receive right is
already a member of another port set, it is removed from that set first
(the whole operation is atomic).  If the port set is
@code{MACH_PORT_NULL}, then the receive right is not put into a port
set, but removed from its current port set.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_NAME} if @var{member} or @var{after} did not denote a
right, @code{KERN_INVALID_RIGHT} if @var{member} denoted a right, but
not a receive right or @var{after} denoted a right, but not a port set,
and @code{KERN_NOT_IN_SET} if @var{after} was @code{MACH_PORT_NULL}, but
@code{member} wasn't currently in a port set.

The @code{mach_port_move_member} call is actually an RPC to @var{task},
normally a send right for a task port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun


@node Request Notifications
@subsection Request Notifications

@deftypefun kern_return_t mach_port_request_notification (@w{ipc_space_t @var{task}}, @w{mach_port_t @var{name}}, @w{mach_msg_id_t @var{variant}}, @w{mach_port_mscount_t @var{sync}}, @w{mach_port_t @var{notify}}, @w{mach_msg_type_name_t @var{notify_type}}, @w{mach_port_t *@var{previous}})
The function @code{mach_port_request_notification} registers a request
for a notification and supplies the send-once right @var{notify} to
which the notification will be sent.  The @var{notify_type} denotes the
IPC type for the send-once right, which can be
@code{MACH_MSG_TYPE_MAKE_SEND_ONCE} or
@code{MACH_MSG_TYPE_MOVE_SEND_ONCE}.  It is an atomic swap, returning
the previously registered send-once right (or @code{MACH_PORT_NULL} for
none) in @var{previous}.  A previous notification request may be
cancelled by providing @code{MACH_PORT_NULL} for @var{notify}.

The @var{variant} argument takes the following values:

@table @code
@item MACH_NOTIFY_PORT_DESTROYED
@var{sync} must be zero.  The @var{name} must specify a receive right,
and the call requests a port-destroyed notification for the receive
right.  If the receive right were to have been destroyed, say by
@code{mach_port_destroy}, then instead the receive right will be sent in
a port-destroyed notification to the registered send-once right.

@item MACH_NOTIFY_DEAD_NAME
The call requests a dead-name notification.  @var{name} specifies send,
receive, or send-once rights for a port.  If the port is destroyed (and
the right remains, becoming a dead name), then a dead-name notification
which carries the name of the right will be sent to the registered
send-once right.  If @var{notify} is not null and sync is non-zero, the
name may specify a dead name, and a dead-name notification is
immediately generated.

Whenever a dead-name notification is generated, the user reference count
of the dead name is incremented.  For example, a send right with two
user refs has a registered dead-name request.  If the port is destroyed,
the send right turns into a dead name with three user refs (instead of
two), and a dead-name notification is generated.

If the name is made available for reuse, perhaps because of
@code{mach_port_destroy} or @code{mach_port_mod_refs}, or the name
denotes a send-once right which has a message sent to it, then the
registered send-once right is used to generate a port-deleted
notification.

@item MACH_NOTIFY_NO_SENDERS
The call requests a no-senders notification.  @var{name} must specify a
receive right.  If @var{notify} is not null, and the receive right's
make-send count is greater than or equal to the sync value, and it has
no extant send rights, than an immediate no-senders notification is
generated.  Otherwise the notification is generated when the receive
right next loses its last extant send right.  In either case, any
previously registered send-once right is returned.

The no-senders notification carries the value the port's make-send count
had when it was generated.  The make-send count is incremented whenever
@code{MACH_MSG_TYPE_MAKE_SEND} is used to create a new send right from
the receive right.  The make-send count is reset to zero when the
receive right is carried in a message.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_TASK} if @var{task} was invalid,
@code{KERN_INVALID_VALUE} if @var{variant} was invalid,
@code{KERN_INVALID_NAME} if @var{name} did not denote a right,
@code{KERN_INVALID_RIGHT} if @var{name} denoted an invalid right and
@code{KERN_INVALID_CAPABILITY} if @var{notify} was invalid.

When using @code{MACH_NOTIFY_PORT_DESTROYED}, the function returns
@code{KERN_INVALID_VALUE} if @var{sync} wasn't zero.

When using @code{MACH_NOTIFY_DEAD_NAME}, the function returns
@code{KERN_RESOURCE_SHORTAGE} if the kernel ran out of memory,
@code{KERN_INVALID_ARGUMENT} if @var{name} denotes a dead name, but
@var{sync} is zero or @var{notify} is @code{MACH_PORT_NULL}, and
@code{KERN_UREFS_OVERFLOW} if @var{name} denotes a dead name, but
generating an immediate dead-name notification would overflow the name's
user-reference count.

The @code{mach_port_request_notification} call is actually an RPC to
@var{task}, normally a send right for a task port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@end deftypefun

@c The inherited ports concept is not used in the Hurd,
@c and so the _SLOT macros are not defined in GNU Mach.

@c  @node Inherited Ports
@c  @subsection Inherited Ports

@c  @deftypefun kern_return_t mach_ports_register (@w{task_t @var{target_task}, @w{port_array_t @var{init_port_set}}, @w{int @var{init_port_array_count}})
@c  @deftypefunx kern_return_t mach_ports_lookup (@w{task_t @var{target_task}, @w{port_array_t *@var{init_port_set}}, @w{int *@var{init_port_array_count}})
@c  @code{mach_ports_register} manipulates the inherited ports array,
@c  @code{mach_ports_lookup} is used to acquire specific parent ports.
@c  @var{target_task} is the task to be affected.  @var{init_port_set} is an
@c  array of system ports to be registered, or returned.  Although the array
@c  size is given as variable, the kernel will only accept a limited number
@c  of ports.  @var{init_port_array_count} is the number of ports returned
@c  in @var{init_port_set}.

@c  @code{mach_ports_register} registers an array of well-known system ports
@c  with the kernel on behalf of a specific task.  Currently the ports to be
@c  registered are: the port to the Network Name Server, the port to the
@c  Environment Manager, and a port to the Service server.  These port
@c  values must be placed in specific slots in the init_port_set.  The slot
@c  numbers are given by the global constants defined in @file{mach_init.h}:
@c  @code{NAME_SERVER_SLOT}, @code{ENVIRONMENT_SLOT}, and
@c  @code{SERVICE_SLOT}.  These ports may later be retrieved with
@c  @code{mach_ports_lookup}.

@c  When a new task is created (see @code{task_create}), the child task will
@c  be given access to these ports.  Only port send rights may be
@c  registered.  Furthermore, the number of ports which may be registered is
@c  fixed and given by the global constant @code{MACH_PORT_SLOTS_USED}
@c  Attempts to register too many ports will fail.

@c  It is intended that this mechanism be used only for task initialization,
@c  and then only by runtime support modules.  A parent task has three
@c  choices in passing these system ports to a child task. Most commonly it
@c  can do nothing and its child will inherit access to the same
@c  @var{init_port_set} that the parent has; or a parent task may register a
@c  set of ports it wishes to have passed to all of its children by calling
@c  @code{mach_ports_register} using its task port; or it may make necessary
@c  modifications to the set of ports it wishes its child to see, and then
@c  register those ports using the child's task port prior to starting the
@c  child's thread(s).  The @code{mach_ports_lookup} call which is done by
@c  @code{mach_init} in the child task will acquire these initial ports for
@c  the child.

@c  Tasks other than the Network Name Server and the Environment Manager
@c  should not need access to the Service port. The Network Name Server port
@c  is the same for all tasks on a given machine. The Environment port is
@c  the only port likely to have different values for different tasks.

@c  Since the number of ports which may be registered is limited, ports
@c  other than those used by the runtime system to initialize a task should
@c  be passed to children either through an initial message, or through the
@c  Network Name Server for public ports, or the Environment Manager for
@c  private ports.

@c  The function returns @code{KERN_SUCCESS} if the memory was allocated,
@c  and @code{KERN_INVALID_ARGUMENT} if an attempt was made to register more
@c  ports than the current kernel implementation allows.
@c  @end deftypefun


@node Virtual Memory Interface
@chapter Virtual Memory Interface

@cindex virtual memory map port
@cindex port representing a virtual memory map
@deftp {Data type} vm_task_t
This is a @code{task_t} (and as such a @code{mach_port_t}), which holds
a port name associated with a port that represents a virtual memory map
in the kernel.  An virtual memory map is used by the kernel to manage
the address space of a task.  The virtual memory map doesn't get a port
name of its own.  Instead the port name of the task provided with the
virtual memory is used to name the virtual memory map of the task (as is
indicated by the fact that the type of @code{vm_task_t} is actually
@code{task_t}).

The virtual memory maps of tasks are the only ones accessible outside of
the kernel.
@end deftp

@menu
* Memory Allocation::             Allocation of new virtual memory.
* Memory Deallocation::           Freeing unused virtual memory.
* Data Transfer::                 Reading, writing and copying memory.
* Memory Attributes::             Tweaking memory regions.
* Mapping Memory Objects::        How to map memory objects.
* Memory Statistics::             How to get statistics about memory usage.
@end menu

@node Memory Allocation
@section Memory Allocation

@deftypefun kern_return_t vm_allocate (@w{vm_task_t @var{target_task}}, @w{vm_address_t *@var{address}}, @w{vm_size_t @var{size}}, @w{boolean_t @var{anywhere}})
The function @code{vm_allocate} allocates a region of virtual memory,
placing it in the specified @var{task}'s address space.

The starting address is @var{address}.  If the @var{anywhere} option is
false, an attempt is made to allocate virtual memory starting at this
virtual address.  If this address is not at the beginning of a virtual
page, it will be rounded down to one.  If there is not enough space at
this address, no memory will be allocated.  If the @var{anywhere} option
is true, the input value of this address will be ignored, and the space
will be allocated wherever it is available.  In either case, the address
at which memory was actually allocated will be returned in
@var{address}.

@var{size} is the number of bytes to allocate (rounded by the system in
a machine dependent way to an integral number of virtual pages).

If @var{anywhere} is true, the kernel should find and allocate any
region of the specified size, and return the address of the resulting
region in address address, rounded to a virtual page boundary if there
is sufficient space.

The physical memory is not actually allocated until the new virtual
memory is referenced.  By default, the kernel rounds all addresses down
to the nearest page boundary and all memory sizes up to the nearest page
size.  The global variable @code{vm_page_size} contains the page size.
@code{mach_task_self} returns the value of the current task port which
should be used as the @var{target_task} argument in order to allocate
memory in the caller's address space.  For languages other than C, these
values can be obtained by the calls @code{vm_statistics} and
@code{mach_task_self}.  Initially, the pages of allocated memory will be
protected to allow all forms of access, and will be inherited in child
tasks as a copy.  Subsequent calls to @code{vm_protect} and
@code{vm_inherit} may be used to change these properties.  The allocated
region is always zero-filled.

The function returns @code{KERN_SUCCESS} if the memory was successfully
allocated, @code{KERN_INVALID_ADDRESS} if an invalid address was
specified and @code{KERN_NO_SPACE} if there was not enough space left to
satisfy the request.
@end deftypefun


@node Memory Deallocation
@section Memory Deallocation

@deftypefun kern_return_t vm_deallocate (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}})
@code{vm_deallocate} relinquishes access to a region of a @var{task}'s
address space, causing further access to that memory to fail.  This
address range will be available for reallocation.  @var{address} is the
starting address, which will be rounded down to a page boundary.
@var{size} is the number of bytes to deallocate, which will be rounded
up to give a page boundary.  Note, that because of the rounding to
virtual page boundaries, more than @var{size} bytes may be deallocated.
Use @code{vm_page_size} or @code{vm_statistics} to find out the current
virtual page size.

This call may be used to deallocate memory that was passed to a task in a
message (via out of line data).  In that case, the rounding should cause
no trouble, since the region of memory was allocated as a set of pages.

The @code{vm_deallocate} call affects only the task specified by the
@var{target_task}.  Other tasks which may have access to this memory may
continue to reference it.

The function returns @code{KERN_SUCCESS} if the memory was successfully
deallocated and @code{KERN_INVALID_ADDRESS} if an invalid or
non-allocated address was specified.
@end deftypefun


@node Data Transfer
@section Data Transfer

@deftypefun kern_return_t vm_read (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}}, @w{vm_offset_t *@var{data}}, @w{mach_msg_type_number_t *@var{data_count}})
The function @code{vm_read} allows one task's virtual memory to be read
by another task.  The @var{target_task} is the task whose memory is to
be read.  @var{address} is the first address to be read and must be on a
page boundary.  @var{size} is the number of bytes of data to be read and
must be an integral number of pages.  @var{data} is the array of data
copied from the given task, and @var{data_count} is the size of the data
array in bytes (will be an integral number of pages).

Note that the data array is returned in a newly allocated region; the
task reading the data should @code{vm_deallocate} this region when it is
done with the data.

The function returns @code{KERN_SUCCESS} if the memory was successfully
read, @code{KERN_INVALID_ADDRESS} if an invalid or non-allocated address
was specified or there was not @var{size} bytes of data following the
address, @code{KERN_INVALID_ARGUMENT} if the address does not start on a
page boundary or the size is not an integral number of pages,
@code{KERN_PROTECTION_FAILURE} if the address region in the target task
is protected against reading and @code{KERN_NO_SPACE} if there was not
enough room in the callers virtual memory to allocate space for the data
to be returned.
@end deftypefun

@deftypefun kern_return_t vm_write (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_offset_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}})
The function @code{vm_write} allows a task to write to the virtual memory
of @var{target_task}.  @var{address} is the starting address in task to
be affected.  @var{data} is an array of bytes to be written, and
@var{data_count} the size of the @var{data} array.

The current implementation requires that @var{address}, @var{data} and
@var{data_count} all be page-aligned.  Otherwise,
@code{KERN_INVALID_ARGUMENT} is returned.

The function returns @code{KERN_SUCCESS} if the memory was successfully
written, @code{KERN_INVALID_ADDRESS} if an invalid or non-allocated
address was specified or there was not @var{data_count} bytes of
allocated memory starting at @var{address} and
@code{KERN_PROTECTION_FAILURE} if the address region in the target task
is protected against writing.
@end deftypefun

@deftypefun kern_return_t vm_copy (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{source_address}}, @w{vm_size_t @var{count}}, @w{vm_offset_t @var{dest_address}})
The function @code{vm_copy} causes the source memory range to be copied
to the destination address.  The source and destination memory ranges
may overlap.  The destination address range must already be allocated
and writable; the source range must be readable.

@code{vm_copy} is equivalent to @code{vm_read} followed by
@code{vm_write}.

The current implementation requires that @var{address}, @var{data} and
@var{data_count} all be page-aligned.  Otherwise,
@code{KERN_INVALID_ARGUMENT} is returned.

The function returns @code{KERN_SUCCESS} if the memory was successfully
written, @code{KERN_INVALID_ADDRESS} if an invalid or non-allocated
address was specified or there was insufficient memory allocated at one
of the addresses and @code{KERN_PROTECTION_FAILURE} if the destination
region was not writable or the source region was not readable.
@end deftypefun


@node Memory Attributes
@section Memory Attributes

@deftypefun kern_return_t vm_region (@w{vm_task_t @var{target_task}}, @w{vm_address_t *@var{address}}, @w{vm_size_t *@var{size}}, @w{vm_prot_t *@var{protection}}, @w{vm_prot_t *@var{max_protection}}, @w{vm_inherit_t *@var{inheritance}}, @w{boolean_t *@var{shared}}, @w{memory_object_name_t *@var{object_name}}, @w{vm_offset_t *@var{offset}})
The function @code{vm_region} returns a description of the specified
region of @var{target_task}'s virtual address space.  @code{vm_region}
begins at @var{address} and looks forward through memory until it comes
to an allocated region.  If address is within a region, then that region
is used.  Various bits of information about the region are returned.  If
@var{address} was not within a region, then @var{address} is set to the
start of the first region which follows the incoming value.  In this way
an entire address space can be scanned.

The @var{size} returned is the size of the located region in bytes.
@var{protection} is the current protection of the region,
@var{max_protection} is the maximum allowable protection for this
region.  @var{inheritance} is the inheritance attribute for this region.
@var{shared} tells if the region is shared or not.  The port
@var{object_name} identifies the memory object associated with this
region, and @var{offset} is the offset into the pager object that this
region begins at.
@c XXX cross ref pager_init

The function returns @code{KERN_SUCCESS} if the memory region was
successfully located and the information returned and @code{KERN_NO_SPACE} if
there is no region at or above @var{address} in the specified task.
@end deftypefun

@deftypefun kern_return_t vm_protect (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}}, @w{boolean_t @var{set_maximum}}, @w{vm_prot_t @var{new_protection}})
The function @code{vm_protect} sets the virtual memory access privileges
for a range of allocated addresses in @var{target_task}'s virtual
address space.  The protection argument describes a combination of read,
write, and execute accesses that should be @emph{permitted}.

@var{address} is the starting address, which will be rounded down to a
page boundary.  @var{size} is the size in bytes of the region for which
protection is to change, and will be rounded up to give a page boundary.
If @var{set_maximum} is set, make the protection change apply to the
maximum protection associated with this address range; otherwise, the
current protection on this range is changed.  If the maximum protection
is reduced below the current protection, both will be changed to reflect
the new maximum.  @var{new_protection} is the new protection value for
this region; a set of: @code{VM_PROT_READ}, @code{VM_PROT_WRITE},
@code{VM_PROT_EXECUTE}.

The enforcement of virtual memory protection is machine-dependent.
Nominally read access requires @code{VM_PROT_READ} permission, write
access requires @code{VM_PROT_WRITE} permission, and execute access
requires @code{VM_PROT_EXECUTE} permission.  However, some combinations
of access rights may not be supported.  In particular, the kernel
interface allows write access to require @code{VM_PROT_READ} and
@code{VM_PROT_WRITE} permission and execute access to require
@code{VM_PROT_READ} permission.

The function returns @code{KERN_SUCCESS} if the memory was successfully
protected, @code{KERN_INVALID_ADDRESS} if an invalid or non-allocated
address was specified and @code{KERN_PROTECTION_FAILURE} if an attempt
was made to increase the current or maximum protection beyond the
existing maximum protection value.
@end deftypefun

@deftypefun kern_return_t vm_inherit (@w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}}, @w{vm_inherit_t @var{new_inheritance}})
The function @code{vm_inherit} specifies how a region of
@var{target_task}'s address space is to be passed to child tasks at the
time of task creation.  Inheritance is an attribute of virtual pages, so
@var{address} to start from will be rounded down to a page boundary and
@var{size}, the size in bytes of the region for which inheritance is to
change, will be rounded up to give a page boundary.  How this memory is
to be inherited in child tasks is specified by @var{new_inheritance}.
Inheritance is specified by using one of these following three values:

@table @code
@item VM_INHERIT_SHARE
Child tasks will share this memory with this task.

@item VM_INHERIT_COPY
Child tasks will receive a copy of this region.

@item VM_INHERIT_NONE
This region will be absent from child tasks.
@end table

Setting @code{vm_inherit} to @code{VM_INHERIT_SHARE} and forking a child
task is the only way two Mach tasks can share physical memory.  Remember
that all the threads of a given task share all the same memory.

The function returns @code{KERN_SUCCESS} if the memory inheritance was
successfully set and @code{KERN_INVALID_ADDRESS} if an invalid or
non-allocated address was specified.
@end deftypefun

@deftypefun kern_return_t vm_wire (@w{host_priv_t @var{host_priv}}, @w{vm_task_t @var{target_task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}}, @w{vm_prot_t @var{access}})
The function @code{vm_wire} allows privileged applications to control
memory pageability.  @var{host_priv} is the privileged host port for the
host on which @var{target_task} resides.  @var{address} is the starting
address, which will be rounded down to a page boundary.  @var{size} is
the size in bytes of the region for which protection is to change, and
will be rounded up to give a page boundary.  @var{access} specifies the
types of accesses that must not cause page faults.

The semantics of a successful @code{vm_wire} operation are that memory
in the specified range will not cause page faults for any accesses
included in access.  Data memory can be made non-pageable (wired) with a
access argument of @code{VM_PROT_READ | VM_PROT_WRITE}.  A special case
is that @code{VM_PROT_NONE} makes the memory pageable.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_HOST} if @var{host_priv} was not the privileged host
port, @code{KERN_INVALID_TASK} if @var{task} was not a valid task,
@code{KERN_INVALID_VALUE} if @var{access} specified an invalid access
mode, @code{KERN_FAILURE} if some memory in the specified range is not
present or has an inappropriate protection value, and
@code{KERN_INVALID_ARGUMENT} if unwiring (@var{access} is
@code{VM_PROT_NONE}) and the memory is not already wired.

The @code{vm_wire} call is actually an RPC to @var{host_priv}, normally
a send right for a privileged host port, but potentially any send right.
In addition to the normal diagnostic return codes from the call's server
(normally the kernel), the call may return @code{mach_msg} return codes.
@end deftypefun

@deftypefun kern_return_t vm_machine_attribute (@w{vm_task_t @var{task}}, @w{vm_address_t @var{address}}, @w{vm_size_t @var{size}}, @w{vm_prot_t @var{access}}, @w{vm_machine_attribute_t @var{attribute}}, @w{vm_machine_attribute_val_t @var{value}})
The function @code{vm_machine_attribute} specifies machine-specific
attributes for a VM mapping, such as cachability, migrability,
replicability.  This is used on machines that allow the user control
over the cache (this is the case for MIPS architectures) or placement of
memory pages as in NUMA architectures (Non-Uniform Memory Access time)
such as the IBM ACE multiprocessor.

Machine-specific attributes can be consider additions to the
machine-independent ones such as protection and inheritance, but they
are not guaranteed to be supported by any given machine.  Moreover,
implementations of Mach on new architectures might find the need for new
attribute types and or values besides the ones defined in the initial
implementation.

The types currently defined are
@table @code
@item MATTR_CACHE
Controls caching of memory pages

@item MATTR_MIGRATE
Controls migrability of memory pages

@item MATTR_REPLICATE
Controls replication of memory pages
@end table

Corresponding values, and meaning of a specific call to
@code{vm_machine_attribute}
@table @code
@item MATTR_VAL_ON
Enables the attribute.  Being enabled is the default value for any
applicable attribute.

@item MATTR_VAL_OFF
Disables the attribute, making memory non-cached, or non-migratable, or
non-replicatable.

@item MATTR_VAL_GET
Returns the current value of the attribute for the memory segment.  If
the attribute does not apply uniformly to the given range the value
returned applies to the initial portion of the segment only.

@item MATTR_VAL_CACHE_FLUSH
Flush the memory pages from the Cache.  The size value in this case
might be meaningful even if not a multiple of the page size, depending
on the implementation.

@item MATTR_VAL_ICACHE_FLUSH
Same as above, applied to the Instruction Cache alone.

@item MATTR_VAL_DCACHE_FLUSH
Same as above, applied to the Data Cache alone.
@end table

The function returns @code{KERN_SUCCESS} if call succeeded, and
@code{KERN_INVALID_ARGUMENT} if @var{task} is not a task, or
@var{address} and @var{size} do not define a valid address range in
task, or @var{attribute} is not a valid attribute type, or it is not
implemented, or @var{value} is not a permissible value for attribute.
@end deftypefun


@node Mapping Memory Objects
@section Mapping Memory Objects

@deftypefun kern_return_t vm_map (@w{vm_task_t @var{target_task}}, @w{vm_address_t *@var{address}}, @w{vm_size_t @var{size}}, @w{vm_address_t @var{mask}}, @w{boolean_t @var{anywhere}}, @w{memory_object_t @var{memory_object}}, @w{vm_offset_t @var{offset}}, @w{boolean_t @var{copy}}, @w{vm_prot_t @var{cur_protection}}, @w{vm_prot_t @var{max_protection}}, @w{vm_inherit_t @var{inheritance}})
The function @code{vm_map} maps a region of virtual memory at the
specified address, for which data is to be supplied by the given memory
object, starting at the given offset within that object.  In addition to
the arguments used in @code{vm_allocate}, the @code{vm_map} call allows
the specification of an address alignment parameter, and of the initial
protection and inheritance values.
@c XXX See the descriptions of vm_allocate, vm_protect , and vm_inherit

If the memory object in question is not currently in use, the kernel
will perform a @code{memory_object_init} call at this time.  If the copy
parameter is asserted, the specified region of the memory object will be
copied to this address space; changes made to this object by other tasks
will not be visible in this mapping, and changes made in this mapping
will not be visible to others (or returned to the memory object).

The @code{vm_map} call returns once the mapping is established.
Completion of the call does not require any action on the part of the
memory manager.

Warning: Only memory objects that are provided by bona fide memory
managers should be used in the @code{vm_map} call.  A memory manager
must implement the memory object interface described elsewhere in this
manual.  If other ports are used, a thread that accesses the mapped
virtual memory may become permanently hung or may receive a memory
exception.

@var{target_task} is the task to be affected.  The starting address is
@var{address}.  If the @var{anywhere} option is used, this address is
ignored.  The address actually allocated will be returned in
@var{address}. @var{size} is the number of bytes to allocate (rounded by
the system in a machine dependent way).  The alignment restriction is
specified by @var{mask}.  Bits asserted in this mask must not be
asserted in the address returned.  If @var{anywhere} is set, the kernel
should find and allocate any region of the specified size, and return
the address of the resulting region in @var{address}.

@var{memory_object} is the port that represents the memory object: used
by user tasks in @code{vm_map}; used by the make requests for data or
other management actions.  If this port is @code{MEMORY_OBJECT_NULL},
then zero-filled memory is allocated instead.  Within a memory object,
@var{offset} specifies an offset in bytes.  This must be page aligned.
If @var{copy} is set, the range of the memory object should be copied to
the target task, rather than mapped read-write.

The function returns @code{KERN_SUCCESS} if the object is mapped,
@code{KERN_NO_SPACE} if no unused region of the task's virtual address
space that meets the address, size, and alignment criteria could be
found, and @code{KERN_INVALID_ARGUMENT} if an invalid argument was provided.
@end deftypefun


@node Memory Statistics
@section Memory Statistics

@deftp {Data type} vm_statistics_data_t
This structure is returned in @var{vm_stats} by the @code{vm_statistics}
function and provides virtual memory statistics for the system.  It has
the following members:

@table @code
@item long pagesize
The page size in bytes.

@item long free_count
The number of free pages.

@item long active_count
The umber of active pages.

@item long inactive_count
The number of inactive pages.

@item long wire_count
The number of pages wired down.

@item long zero_fill_count
The number of zero filled pages.

@item long reactivations
The number of reactivated pages.

@item long pageins
The number of pageins.

@item long pageouts
The number of pageouts.

@item long faults
The number of faults.

@item long cow_faults
The number of copy-on-writes.

@item long lookups
The number of object cache lookups.

@item long hits
The number of object cache hits.
@end table
@end deftp

@deftypefun kern_return_t vm_statistics (@w{vm_task_t @var{target_task}}, @w{vm_statistics_data_t *@var{vm_stats}})
The function @code{vm_statistics} returns the statistics about the
kernel's use of virtual memory since the kernel was booted.
@code{pagesize} can also be found as a global variable
@code{vm_page_size} which is set at task initialization and remains
constant for the life of the task.
@end deftypefun


@node External Memory Management
@chapter External Memory Management

@menu
* Memory Object Server::          The basics of external memory management.
* Memory Object Creation::        How new memory objects are created.
* Memory Object Termination::     How memory objects are terminated.
* Memory Objects and Data::       Data transfer to and from memory objects.
* Memory Object Locking::         How memory objects are locked.
* Memory Object Attributes::      Manipulating attributes of memory objects.
* Default Memory Manager::        Setting and using the default memory manager.
@end menu


@node Memory Object Server
@section Memory Object Server

@deftypefun boolean_t memory_object_server (@w{msg_header_t *@var{in_msg}}, @w{msg_header_t *@var{out_msg}})
@deftypefunx boolean_t memory_object_default_server (@w{msg_header_t *@var{in_msg}}, @w{msg_header_t *@var{out_msg}})
@deftypefunx boolean_t seqnos_memory_object_server (@w{msg_header_t *@var{in_msg}}, @w{msg_header_t *@var{out_msg}})
@deftypefunx boolean_t seqnos_memory_object_default_server (@w{msg_header_t *@var{in_msg}}, @w{msg_header_t *@var{out_msg}})
A memory manager is a server task that responds to specific messages
from the kernel in order to handle memory management functions for the
kernel.

In order to isolate the memory manager from the specifics of message
formatting, the remote procedure call generator produces a procedure,
@code{memory_object_server}, to handle a received message.  This
function does all necessary argument handling, and actually calls one of
the following functions: @code{memory_object_init},
@code{memory_object_data_write}, @code{memory_object_data_return},
@code{memory_object_data_request}, @code{memory_object_data_unlock},
@code{memory_object_lock_completed}, @code{memory_object_copy},
@code{memory_object_terminate}.  The @strong{default memory manager} may
get two additional requests from the kernel: @code{memory_object_create}
and @code{memory_object_data_initialize}.  The remote procedure call
generator produces a procedure @code{memory_object_default_server} to
handle those functions specific to the default memory manager.

The @code{seqnos_memory_object_server} and
@code{seqnos_memory_object_default_server} differ from
@code{memory_object_server} and @code{memory_object_default_server} in
that they supply message sequence numbers to the server interfaces.
They call the @code{seqnos_memory_object_*} functions, which complement
the @code{memory_object_*} set of functions.

The return value from the @code{memory_object_server} function indicates
that the message was appropriate to the memory management interface
(returning @code{TRUE}), or that it could not handle this message
(returning @code{FALSE}).

The @var{in_msg} argument is the message that has been received from the
kernel.  The @var{out_msg} is a reply message, but this is not used for
this server.

The function returns @code{TRUE} to indicate that the message in
question was applicable to this interface, and that the appropriate
routine was called to interpret the message.  It returns @code{FALSE} to
indicate that the message did not apply to this interface, and that no
other action was taken.
@end deftypefun


@node Memory Object Creation
@section Memory Object Creation

@deftypefun kern_return_t memory_object_init (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{memory_object_name_t @var{memory_object_name}}, @w{vm_size_t @var{memory_object_page_size}})
@deftypefunx kern_return_t seqnos_memory_object_init (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{memory_object_name_t @var{memory_object_name}}, @w{vm_size_t @var{memory_object_page_size}})
The function @code{memory_object_init} serves as a notification that the
kernel has been asked to map the given memory object into a task's
virtual address space.  Additionally, it provides a port on which the
memory manager may issue cache management requests, and a port which the
kernel will use to name this data region.  In the event that different
each will perform a @code{memory_object_init} call with new request and
name ports.  The virtual page size that is used by the calling kernel is
included for planning purposes.

When the memory manager is prepared to accept requests for data for this
object, it must call @code{memory_object_ready} with the attribute.
Otherwise the kernel will not process requests on this object.  To
reject all mappings of this object, the memory manager may use
@code{memory_object_destroy}.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)
@var{memory_object_name} is a port used by the kernel to refer to the
memory object data in response to @code{vm_region} calls.
@code{memory_object_page_size} is the page size to be used by this
kernel.  All data sizes in calls involving this kernel must be an
integral multiple of the page size.  Note that different kernels,
indicated by a different @code{memory_control}, may have different page
sizes.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_ready (@w{memory_object_control_t @var{memory_control}}, @w{boolean_t @var{may_cache_object}}, @w{memory_object_copy_strategy_t @var{copy_strategy}})
The function @code{memory_object_ready} informs the kernel that the
memory manager is ready to receive data or unlock requests on behalf of
the clients.  The argument @var{memory_control} is the port, provided by
the kernel in a @code{memory_object_init} call, to which cache
management requests may be issued.  If @var{may_cache_object} is set,
the kernel may keep data associated with this memory object, even after
virtual memory references to it are gone.

@var{copy_strategy} tells how the kernel should copy regions of the
associated memory object.  There are three possible caching strategies:
@code{MEMORY_OBJECT_COPY_NONE} which specifies that nothing special
should be done when data in the object is copied;
@code{MEMORY_OBJECT_COPY_CALL} which specifies that the memory manager
should be notified via a @code{memory_object_copy} call before any part
of the object is copied; and @code{MEMORY_OBJECT_COPY_DELAY} which
guarantees that the memory manager does not externally modify the data
so that the kernel can use its normal copy-on-write algorithms.
@code{MEMORY_OBJECT_COPY_DELAY} is the strategy most commonly used.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun


@node Memory Object Termination
@section Memory Object Termination

@deftypefun kern_return_t memory_object_terminate (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{memory_object_name_t @var{memory_object_name}})
@deftypefunx kern_return_t seqnos_memory_object_terminate (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{memory_object_name_t @var{memory_object_name}})
The function @code{memory_object_terminate} indicates that the kernel
has completed its use of the given memory object.  All rights to the
memory object control and name ports are included, so that the memory
manager can destroy them (using @code{mach_port_deallocate}) after doing
appropriate bookkeeping.  The kernel will terminate a memory object only
after all address space mappings of that memory object have been
deallocated, or upon explicit request by the memory manager.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)
@var{memory_object_name} is a port used by the kernel to refer to the
memory object data in response to @code{vm_region} calls.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_destroy (@w{memory_object_control_t @var{memory_control}}, @w{kern_return_t @var{reason}})
The function @code{memory_object_destroy} tells the kernel to shut down
the memory object.  As a result of this call the kernel will no longer
support paging activity or any @code{memory_object} calls on this
object, and all rights to the memory object port, the memory control
port and the memory name port will be returned to the memory manager in
a memory_object_terminate call.  If the memory manager is concerned that
any modified cached data be returned to it before the object is
terminated, it should call @code{memory_object_lock_request} with
@var{should_flush} set and a lock value of @code{VM_PROT_WRITE} before
making this call.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{reason} is an error code indicating why the object
must be destroyed.
@c The error code is currently ignored.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun


@node Memory Objects and Data
@section Memory Objects and Data

@deftypefun kern_return_t memory_object_data_return (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}}, @w{boolean_t @var{dirty}}, @w{boolean_t @var{kernel_copy}})
@deftypefunx kern_return_t seqnos_memory_object_data_return (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}}, @w{boolean_t @var{dirty}}, @w{boolean_t @var{kernel_copy}})
The function @code{memory_object_data_return} provides the memory
manager with data that has been modified while cached in physical
memory.  Once the memory manager no longer needs this data (e.g., it has
been written to another storage medium), it should be deallocated using
@code{vm_deallocate}.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  This will be
page aligned.  @var{data} is the data which has been modified while
cached in physical memory.  @var{data_count} is the amount of data to be
written, in bytes.  This will be an integral number of memory object
pages.

The kernel will also use this call to return precious pages.  If an
unmodified precious age is returned, @var{dirty} is set to @code{FALSE},
otherwise it is @code{TRUE}.  If @var{kernel_copy} is @code{TRUE}, the
kernel kept a copy of the page.  Precious data remains precious if the
kernel keeps a copy.  The indication that the kernel kept a copy is only
a hint if the data is not precious; the cleaned copy may be discarded
without further notifying the manager.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_data_request (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{length}}, @w{vm_prot_t @var{desired_access}})
@deftypefunx kern_return_t seqnos_memory_object_data_request (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{length}}, @w{vm_prot_t @var{desired_access}})
The function @code{memory_object_data_request} is a request for data
from the specified memory object, for at least the access specified.
The memory manager is expected to return at least the specified data,
with as much access as it can allow, using
@code{memory_object_data_supply}.  If the memory manager is unable to
provide the data (for example, because of a hardware error), it may use
the @code{memory_object_data_error} call.  The
@code{memory_object_data_unavailable} call may be used to tell the
kernel to supply zero-filled memory for this region.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  This will be
page aligned.  @var{length} is the number of bytes of data, starting at
@var{offset}, to which this call refers.  This will be an integral
number of memory object pages.  @var{desired_access} is a protection
value describing the memory access modes which must be permitted on the
specified cached data.  One or more of: @code{VM_PROT_READ},
@code{VM_PROT_WRITE} or @code{VM_PROT_EXECUTE}.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_data_supply (@w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}}, @w{vm_prot_t @var{lock_value}}, @w{boolean_t @var{precious}},  @w{mach_port_t @var{reply}})
The function @code{memory_object_data_supply} supplies the kernel with
data for the specified memory object.  Ordinarily, memory managers
should only provide data in response to @code{memory_object_data_request}
calls from the kernel (but they may provide data in advance as desired).
When data already held by this kernel is provided again, the new data is
ignored.  The kernel may not provide any data (or protection)
consistency among pages with different virtual page alignments within
the same object.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{offset} is an offset within a memory object in bytes.
This must be page aligned.  @var{data} is the data that is being
provided to the kernel.  This is a pointer to the data.
@var{data_count} is the amount of data to be provided.  Only whole
virtual pages of data can be accepted; partial pages will be discarded.

@var{lock_value} is a protection value indicating those forms of access
that should @strong{not} be permitted to the specified cached data.  The
lock values must be one or more of the set: @code{VM_PROT_NONE},
@code{VM_PROT_READ}, @code{VM_PROT_WRITE}, @code{VM_PROT_EXECUTE} and
@code{VM_PROT_ALL} as defined in @file{mach/vm_prot.h}.

If @var{precious} is @code{FALSE}, the kernel treats the data as a
temporary and may throw it away if it hasn't been changed.  If the
@var{precious} value is @code{TRUE}, the kernel treats its copy as a
data repository and promises to return it to the manager; the manager
may tell the kernel to throw it away instead by flushing and not
cleaning the data (see @code{memory_object_lock_request}).

If @var{reply_to} is not @code{MACH_PORT_NULL}, the kernel will send a
completion message to the provided port (see
@code{memory_object_supply_completed}).

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_supply_completed (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{kern_return_t @var{result}}, @w{vm_offset_t @var{error_offset}})
@deftypefunx kern_return_t seqnos_memory_object_supply_completed (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{kern_return_t @var{result}}, @w{vm_offset_t @var{error_offset}})
The function @code{memory_object_supply_completed} indicates that a
previous @code{memory_object_data_supply} has been completed.  Note that
this call is made on whatever port was specified in the
@code{memory_object_data_supply} call; that port need not be the memory
object port itself.  No reply is expected after this call.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  @var{length}
is the length of the data covered by the lock request.  The @var{result}
parameter indicates what happened during the supply.  If it is not
@code{KERN_SUCCESS}, then @var{error_offset} identifies the first offset
at which a problem occurred.  The pagein operation stopped at this
point.  Note that the only failures reported by this mechanism are
@code{KERN_MEMORY_PRESENT}.  All other failures (invalid argument, error
on pagein of supplied data in manager's address space) cause the entire
operation to fail.


@end deftypefun

@deftypefun kern_return_t memory_object_data_error (@w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{size}}, @w{kern_return_t @var{reason}})
The function @code{memory_object_data_error} indicates that the memory
manager cannot return the data requested for the given region,
specifying a reason for the error.  This is typically used when a
hardware error is encountered.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{offset} is an offset within a memory object in bytes.
This must be page aligned.  @var{data} is the data that is being
provided to the kernel.  This is a pointer to the data.  @var{size} is
the amount of cached data (starting at @var{offset}) to be handled.
This must be an integral number of the memory object page size.
@var{reason} is an error code indicating what type of error occurred.
@c The error code is currently ignored.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_data_unavailable (@w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{size}}, @w{kern_return_t @var{reason}})
The function @code{memory_object_data_unavailable} indicates that the
memory object does not have data for the given region and that the
kernel should provide the data for this range.  The memory manager may
use this call in three different situations.

@enumerate
@item
The object was created by @code{memory_object_create} and the kernel has
not yet provided data for this range (either via a
@code{memory_object_data_initialize}, @code{memory_object_data_write} or
a @code{memory_object_data_return} for the object.

@item
The object was created by an @code{memory_object_data_copy} and the
kernel should copy this region from the original memory object.

@item
The object is a normal user-created memory object and the kernel should
supply unlocked zero-filled pages for the range.
@end enumerate

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{offset} is an offset within a memory object, in bytes.
This must be page aligned.  @var{size} is the amount of cached data
(starting at @var{offset}) to be handled.  This must be an integral
number of the memory object page size.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_copy (@w{memory_object_t @var{old_memory_object}}, @w{memory_object_control_t @var{old_memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{memory_object_t @var{new_memory_object}})
@deftypefunx kern_return_t seqnos_memory_object_copy (@w{memory_object_t @var{old_memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{old_memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{memory_object_t @var{new_memory_object}})
The function @code{memory_object_copy} indicates that a copy has been
made of the specified range of the given original memory object.  This
call includes only the new memory object itself; a
@code{memory_object_init} call will be made on the new memory object
after the currently cached pages of the original object are prepared.
After the memory manager receives the init call, it must reply with the
@code{memory_object_ready} call to assert the "ready" attribute.  The
kernel will use the new memory object, control and name ports to refer
to the new copy.

This call is made when the original memory object had the caching
parameter set to @code{MEMORY_OBJECT_COPY_CALL} and a user of the object
has asked the kernel to copy it.

Cached pages from the original memory object at the time of the copy
operation are handled as follows: Readable pages may be silently copied
to the new memory object (with all access permissions).  Pages not
copied are locked to prevent write access.

The new memory object is @strong{temporary}, meaning that the memory
manager should not change its contents or allow the memory object to be
mapped in another client.  The memory manager may use the
@code{memory_object_data_unavailable} call to indicate that the
appropriate pages of the original memory object may be used to fulfill
the data request.

The argument @var{old_memory_object} is the port that represents the old
memory object data.  @var{old_memory_control} is the kernel port for the
old object.  @var{offset} is the offset within a memory object to which
this call refers.  This will be page aligned.  @var{length} is the
number of bytes of data, starting at @var{offset}, to which this call
refers.  This will be an integral number of memory object pages.
@var{new_memory_object} is a new memory object created by the kernel;
see synopsis for further description.  Note that all port rights
(including receive rights) are included for the new memory object.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

The remaining interfaces in this section are obsolete.

@deftypefun kern_return_t memory_object_data_write (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}})
@deftypefunx kern_return_t seqnos_memory_object_data_write (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}})
The function @code{memory_object_data_write} provides the memory manager
with data that has been modified while cached in physical memory.  It is the old form of @code{memory_object_data_return}.  Once
the memory manager no longer needs this data (e.g., it has been written
to another storage medium), it should be deallocated using
@code{vm_deallocate}.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  This will be
page aligned.  @var{data} is the data which has been modified while
cached in physical memory.  @var{data_count} is the amount of data to be
written, in bytes.  This will be an integral number of memory object
pages.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_data_provided (@w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}}, @w{vm_prot_t @var{lock_value}})
The function @code{memory_object_data_provided} supplies the kernel with
data for the specified memory object.  It is the old form of
@code{memory_object_data_supply}.  Ordinarily, memory managers should
only provide data in response to @code{memory_object_data_request} calls
from the kernel.  The @var{lock_value} specifies what type of access
will not be allowed to the data range.  The lock values must be one or
more of the set: @code{VM_PROT_NONE}, @code{VM_PROT_READ},
@code{VM_PROT_WRITE}, @code{VM_PROT_EXECUTE} and @code{VM_PROT_ALL} as
defined in @file{mach/vm_prot.h}.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{offset} is an offset within a memory object in bytes.
This must be page aligned.  @var{data} is the data that is being
provided to the kernel.  This is a pointer to the data.
@var{data_count} is the amount of data to be provided.  This must be an
integral number of memory object pages.  @var{lock_value} is a
protection value indicating those forms of access that should
@strong{not} be permitted to the specified cached data.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun


@node Memory Object Locking
@section Memory Object Locking

@deftypefun kern_return_t memory_object_lock_request (@w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{size}}, @w{memory_object_return_t @var{should_clean}}, @w{boolean_t @var{should_flush}}, @w{vm_prot_t @var{lock_value}}, @w{mach_port_t @var{reply_to}})
The function @code{memory_object_lock_request} allows a memory manager
to make cache management requests.  As specified in arguments to the
call, the kernel will:
@itemize
@item
clean (i.e., write back using @code{memory_object_data_supply} or
@code{memory_object_data_write}) any cached data which has been modified
since the last time it was written

@item
flush (i.e., remove any uses of) that data from memory

@item
lock (i.e., prohibit the specified uses of) the cached data
@end itemize

Locks applied to cached data are not cumulative; new lock values
override previous ones.  Thus, data may also be unlocked using this
primitive.  The lock values must be one or more of the following values:
@code{VM_PROT_NONE}, @code{VM_PROT_READ}, @code{VM_PROT_WRITE},
@code{VM_PROT_EXECUTE} and @code{VM_PROT_ALL} as defined in
@file{mach/vm_prot.h}.

Only data which is cached at the time of this call is affected.  When a
running thread requires a prohibited access to cached data, the kernel
will issue a @code{memory_object_data_unlock} call specifying the forms
of access required. 

Once all of the actions requested by this call have been completed, the
kernel issues a @code{memory_object_lock_completed} call on the
specified reply port.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  @var{offset} is an offset within a memory object, in bytes.
This must be page aligned.  @var{size} is the amount of cached data
(starting at @var{offset}) to be handled.  This must be an integral
number of the memory object page size.  If @var{should_clean} is set,
modified data should be written back to the memory manager.  If
@var{should_flush} is set, the specified cached data should be
invalidated, and all uses of that data should be revoked.
@var{lock_value} is a protection value indicating those forms of access
that should @strong{not} be permitted to the specified cached data.
@var{reply_to} is a port on which a @code{memory_object_lock_completed}
call should be issued, or @code{MACH_PORT_NULL} if no acknowledgement is
desired.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_lock_completed (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}})
@deftypefunx kern_return_t seqnos_memory_object_lock_completed (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}})
The function @code{memory_object_lock_completed} indicates that a
previous @code{memory_object_lock_request} has been completed.  Note
that this call is made on whatever port was specified in the
@code{memory_object_lock_request} call; that port need not be the memory
object port itself.  No reply is expected after this call.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  @var{length}
is the length of the data covered by the lock request.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_data_unlock (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{vm_prot_t @var{desired_access}})
@deftypefunx kern_return_t seqnos_memory_object_data_unlock (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{length}}, @w{vm_prot_t @var{desired_access}})
The function @code{memory_object_data_unlock} is a request that the
memory manager permit at least the desired access to the specified data
cached by the kernel.  A call to @code{memory_object_lock_request} is
expected in response.

The argument @var{memory_object} is the port that represents the memory
object data, as supplied to the kernel in a @code{vm_map} call.
@var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  This will be
page aligned.  @var{length} is the number of bytes of data, starting at
@var{offset}, to which this call refers.  This will be an integral
number of memory object pages.  @var{desired_access} a protection value
describing the memory access modes which must be permitted on the
specified cached data.  One or more of: @code{VM_PROT_READ},
@code{VM_PROT_WRITE} or @code{VM_PROT_EXECUTE}.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun


@node Memory Object Attributes
@section Memory Object Attributes

@deftypefun kern_return_t memory_object_get_attributes (@w{memory_object_control_t @var{memory_control}}, @w{boolean_t *@var{object_ready}}, @w{boolean_t *@var{may_cache_object}}, @w{memory_object_copy_strategy_t *@var{copy_strategy}})
The function @code{memory_object_get_attribute} retrieves the current
attributes associated with the memory object.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  If @var{object_ready} is set, the kernel may issue new data
and unlock requests on the associated memory object.  If
@var{may_cache_object} is set, the kernel may keep data associated with
this memory object, even after virtual memory references to it are gone.
@var{copy_strategy} tells how the kernel should copy regions of the
associated memory object.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_change_attributes (@w{memory_object_control_t @var{memory_control}}, @w{boolean_t @var{may_cache_object}}, @w{memory_object_copy_strategy_t @var{copy_strategy}}, @w{mach_port_t @var{reply_to}})
The function @code{memory_object_change_attribute} sets
performance-related attributes for the specified memory object.  If the
caching attribute is asserted, the kernel is permitted (and encouraged)
to maintain cached data for this memory object even after no virtual
address space contains this data.

There are three possible caching strategies:
@code{MEMORY_OBJECT_COPY_NONE} which specifies that nothing special
should be done when data in the object is copied;
@code{MEMORY_OBJECT_COPY_CALL} which specifies that the memory manager
should be notified via a @code{memory_object_copy} call before any part
of the object is copied; and @code{MEMORY_OBJECT_COPY_DELAY} which
guarantees that the memory manager does not externally modify the data
so that the kernel can use its normal copy-on-write algorithms.
@code{MEMORY_OBJECT_COPY_DELAY} is the strategy most commonly used.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  If @var{may_cache_object} is set, the kernel may keep data
associated with this memory object, even after virtual memory references
to it are gone.  @var{copy_strategy} tells how the kernel should copy
regions of the associated memory object.  @var{reply_to} is a port on
which a @code{memory_object_change_completed} call will be issued upon
completion of the attribute change, or @code{MACH_PORT_NULL} if no
acknowledgement is desired.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun

@deftypefun kern_return_t memory_object_change_completed (@w{memory_object_t @var{memory_object}}, @w{boolean_t @var{may_cache_object}}, @w{memory_object_copy_strategy_t @var{copy_strategy}})
@deftypefunx kern_return_t seqnos_memory_object_change_completed (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{boolean_t @var{may_cache_object}}, @w{memory_object_copy_strategy_t @var{copy_strategy}})
The function @code{memory_object_change_completed} indicates the
completion of an attribute change call.

@c  Warning: This routine does NOT contain a memory_object_control_t because
@c  the memory_object_change_attributes call may cause memory object
@c  termination (by uncaching the object).  This would yield an invalid
@c  port.
@end deftypefun

The following interface is obsoleted by @code{memory_object_ready} and
@code{memory_object_change_attributes}.  If the old form
@code{memory_object_set_attributes} is used to make a memory object
ready, the kernel will write back data using the old
@code{memory_object_data_write} interface rather than
@code{memory_object_data_return}..

@deftypefun kern_return_t memory_object_set_attributes (@w{memory_object_control_t @var{memory_control}}, @w{boolean @var{object_ready}}, @w{boolean_t @var{may_cache_object}}, @w{memory_object_copy_strategy_t @var{copy_strategy}})
The function @code{memory_object_set_attribute} controls how the
memory object.  The kernel will only make data or unlock requests when
the ready attribute is asserted.  If the caching attribute is asserted,
the kernel is permitted (and encouraged) to maintain cached data for
this memory object even after no virtual address space contains this
data.

There are three possible caching strategies:
@code{MEMORY_OBJECT_COPY_NONE} which specifies that nothing special
should be done when data in the object is copied;
@code{MEMORY_OBJECT_COPY_CALL} which specifies that the memory manager
should be notified via a @code{memory_object_copy} call before any part
of the object is copied; and @code{MEMORY_OBJECT_COPY_DELAY} which
guarantees that the memory manager does not externally modify the data
so that the kernel can use its normal copy-on-write algorithms.
@code{MEMORY_OBJECT_COPY_DELAY} is the strategy most commonly used.

The argument @var{memory_control} is the port, provided by the kernel in
a @code{memory_object_init} call, to which cache management requests may
be issued.  If @var{object_ready} is set, the kernel may issue new data
and unlock requests on the associated memory object.  If
@var{may_cache_object} is set, the kernel may keep data associated with
this memory object, even after virtual memory references to it are gone.
@var{copy_strategy} tells how the kernel should copy regions of the
associated memory object.

This routine does not receive a reply message (and consequently has no
return value), so only message transmission errors apply.
@end deftypefun


@node Default Memory Manager
@section Default Memory Manager

@deftypefun kern_return_t vm_set_default_memory_manager (@w{host_t @var{host}}, @w{mach_port_t *@var{default_manager}})
The function @code{vm_set_default_memory_manager} sets the kernel's
default memory manager.  It sets the port to which newly-created
temporary memory objects are delivered by @code{memory_object_create} to
the host.  The old memory manager port is returned.  If
@var{default_manager} is @code{MACH_PORT_NULL} then this routine just returns
the current default manager port without changing it.

The argument @var{host} is a task port to the kernel whose default
memory manager is to be changed.  @var{default_manager} is an in/out
parameter. As input, @var{default_manager} is the port that the new
memory manager is listening on for @code{memory_object_create} calls.
As output, it is the old default memory manager's port.

The function returns @code{KERN_SUCCESS} if the new memory manager is
installed, and @code{KERN_INVALID_ARGUMENT} if this task does not have
the privileges required for this call.
@end deftypefun

@deftypefun kern_return_t memory_object_create (@w{memory_object_t @var{old_memory_object}}, @w{memory_object_t @var{new_memory_object}}, @w{vm_size_t @var{new_object_size}}, @w{memory_object_control_t @var{new_control}}, @w{memory_object_name_t @var{new_name}}, @w{vm_size_t @var{new_page_size}})
@deftypefunx kern_return_t seqnos_memory_object_create (@w{memory_object_t @var{old_memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_t @var{new_memory_object}}, @w{vm_size_t @var{new_object_size}}, @w{memory_object_control_t @var{new_control}}, @w{memory_object_name_t @var{new_name}}, @w{vm_size_t @var{new_page_size}})
The function @code{memory_object_create} is a request that the given
memory manager accept responsibility for the given memory object created
by the kernel.  This call will only be made to the system
@strong{default memory manager}.  The memory object in question
initially consists of zero-filled memory; only memory pages that are
actually written will ever be provided to
@code{memory_object_data_request} calls, the default memory manager must
use @code{memory_object_data_unavailable} for any pages that have not
previously been written.

No reply is expected after this call.  Since this call is directed to
the default memory manager, the kernel assumes that it will be ready to
handle data requests to this object and does not need the confirmation
of a @code{memory_object_set_attributes} call.

The argument @var{old_memory_object} is a memory object provided by the
default memory manager on which the kernel can make
@code{memory_object_create} calls.  @var{new_memory_object} is a new
memory object created by the kernel; see synopsis for further
description.  Note that all port rights (including receive rights) are
included for the new memory object.  @var{new_object_size} is the
maximum size of the new object.  @var{new_control} is a port, created by
the kernel, on which a memory manager may issue cache management
requests for the new object.  @var{new_name} a port used by the kernel
to refer to the new memory object data in response to @code{vm_region}
calls.  @var{new_page_size} is the page size to be used by this kernel.
All data sizes in calls involving this kernel must be an integral
multiple of the page size.  Note that different kernels, indicated by
different a @code{memory_control}, may have different page sizes.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun

@deftypefun kern_return_t memory_object_data_initialize (@w{memory_object_t @var{memory_object}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}})
@deftypefunx kern_return_t seqnos_memory_object_data_initialize (@w{memory_object_t @var{memory_object}}, @w{mach_port_seqno_t @var{seqno}}, @w{memory_object_control_t @var{memory_control}}, @w{vm_offset_t @var{offset}}, @w{vm_offset_t @var{data}}, @w{vm_size_t @var{data_count}})
The function @code{memory_object_data_initialize} provides the memory
manager with initial data for a kernel-created memory object.  If the
memory manager already has been supplied data (by a previous
@code{memory_object_data_initialize}, @code{memory_object_data_write} or
@code{memory_object_data_return}), then this data should be ignored.
Otherwise, this call behaves exactly as does
@code{memory_object_data_return} on memory objects created by the kernel
via @code{memory_object_create} and thus will only be made to default
memory managers.  This call will not be made on objects created via
@code{memory_object_copy}.

The argument @var{memory_object} the port that represents the memory
object data, as supplied by the kernel in a @code{memory_object_create}
call.  @var{memory_control} is the request port to which a response is
requested.  (In the event that a memory object has been supplied to more
than one the kernel that has made the request.)  @var{offset} is the
offset within a memory object to which this call refers.  This will be
page aligned.  @var{data} is the data which has been modified while
cached in physical memory.  @var{data_count} is the amount of data to be
written, in bytes.  This will be an integral number of memory object
pages.

The function should return @code{KERN_SUCCESS}, but since this routine
is called by the kernel, which does not wait for a reply message, this
value is ignored.
@end deftypefun


@node Threads and Tasks
@chapter Threads and Tasks

@menu
* Thread Interface::              Manipulating threads.
* Task Interface::                Manipulating tasks.
* Profiling::                     Profiling threads and tasks.
@end menu


@node Thread Interface
@section Thread Interface

@cindex thread port
@cindex port representing a thread
@deftp {Data type} thread_t 
This is a @code{mach_port_t} and used to hold the port name of a
thread port that represents the thread.  Manipulations of the thread are
implemented as remote procedure calls to the thread port.  A thread can
get a port to itself with the @code{mach_thread_self} system call.
@end deftp 

@menu
* Thread Creation::               Creating new threads.
* Thread Termination::            Terminating existing threads.
* Thread Information::            How to get informations on threads.
* Thread Settings::               How to set threads related informations.
* Thread Execution::              How to control the thread's machine state.
* Scheduling::                    Operations on thread scheduling.
* Thread Special Ports::          How to handle the thread's special ports.
* Exceptions::                    Managing exceptions.
@end menu


@node Thread Creation
@subsection Thread Creation

@deftypefun kern_return_t thread_create (@w{task_t @var{parent_task}}, @w{thread_t *@var{child_thread}})
The function @code{thread_create} creates a new thread within the task
specified by @var{parent_task}.  The new thread has no processor state,
and has a suspend count of 1.  To get a new thread to run, first
@code{thread_create} is called to get the new thread's identifier,
(@var{child_thread}).  Then @code{thread_set_state} is called to set a
processor state, and finally @code{thread_resume} is called to get the
thread scheduled to execute.

When the thread is created send rights to its thread kernel port are
given to it and returned to the caller in @var{child_thread}.  The new
thread's exception port is set to @code{MACH_PORT_NULL}.

The function returns @code{KERN_SUCCESS} if a new thread has been
created, @code{KERN_INVALID_ARGUMENT} if @var{parent_task} is not a
valid task and @code{KERN_RESOURCE_SHORTAGE} if some critical kernel
resource is not available.
@end deftypefun


@node Thread Termination
@subsection Thread Termination

@deftypefun kern_return_t thread_terminate (@w{thread_t @var{target_thread}})
The function @code{thread_terminate} destroys the thread specified by
@var{target_thread}.

The function returns @code{KERN_SUCCESS} if the thread has been killed
and @code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not a thread.
@end deftypefun


@node Thread Information
@subsection Thread Information

@deftypefun thread_t mach_thread_self ()
The @code{mach_thread_self} system call returns the calling thread's
thread port.

@code{mach_thread_self} has an effect equivalent to receiving a send
right for the thread port.  @code{mach_thread_self} returns the name of
the send right.  In particular, successive calls will increase the
calling task's user-reference count for the send right.

@c author{marcus}
As a special exception, the kernel will overrun the user reference count
of the thread name port, so that this function can not fail for that
reason.  Because of this, the user should not deallocate the port right
if an overrun might have happened.  Otherwise the reference count could
drop to zero and the send right be destroyed while the user still
expects to be able to use it.  As the kernel does not make use of the
number of extant send rights anyway, this is safe to do (the thread port
itself is not destroyed, even when there are no send rights anymore).

The function returns @code{MACH_PORT_NULL} if a resource shortage
prevented the reception of the send right or if the thread port is
currently null and @code{MACH_PORT_DEAD} if the thread port is currently
dead.
@end deftypefun

@deftypefun kern_return_t thread_info (@w{thread_t @var{target_thread}}, @w{int @var{flavor}}, @w{thread_info_t @var{thread_info}}, @w{mach_msg_type_number_t *@var{thread_infoCnt}})
The function @code{thread_info} returns the selected information array
for a thread, as specified by @var{flavor}.

@var{thread_info} is an array of integers that is supplied by the caller
and returned filled with specified information.  @var{thread_infoCnt} is
supplied as the maximum number of integers in @var{thread_info}.  On
return, it contains the actual number of integers in @var{thread_info}.
The maximum number of integers returned by any flavor is
@code{THREAD_INFO_MAX}.

The type of information returned is defined by @var{flavor}, which can
be one of the following:

@table @code
@item THREAD_BASIC_INFO
The function returns basic information about the thread, as defined by
@code{thread_basic_info_t}.  This includes the user and system time, the
run state, and scheduling priority.  The number of integers returned is
@code{THREAD_BASIC_INFO_COUNT}.

@item THREAD_SCHED_INFO
The function returns information about the scheduling policy for the
thread as defined by @code{thread_sched_info_t}.  The number of integers
returned is @code{THREAD_SCHED_INFO_COUNT}.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not a thread or
@var{flavor} is not recognized.  The function returns
@code{MIG_ARRAY_TOO_LARGE} if the returned info array is too large for
@var{thread_info}.  In this case, @var{thread_info} is filled as much as
possible and @var{thread_infoCnt} is set to the number of elements that
would have been returned if there were enough room.
@end deftypefun

@deftp {Data type} {struct thread_basic_info}
This structure is returned in @var{thread_info} by the
@code{thread_info} function and provides basic information about the
thread.  You can cast a variable of type @code{thread_info_t} to a
pointer of this type if you provided it as the @var{thread_info}
parameter for the @code{THREAD_BASIC_INFO} flavor of @code{thread_info}.
It has the following members:

@table @code
@item time_value_t user_time
user run time

@item time_value_t system_time
system run time
@item int cpu_usage
Scaled cpu usage percentage.  The scale factor is @code{TH_USAGE_SCALE}.

@item int base_priority
The base scheduling priority of the thread.

@item int cur_priority
The current scheduling priority of the thread.

@item integer_t run_state
The run state of the thread.  The possible values of this field are:
@table @code
@item TH_STATE_RUNNING
The thread is running normally.

@item TH_STATE_STOPPED
The thread is suspended.

@item TH_STATE_WAITING
The thread is waiting normally.

@item TH_STATE_UNINTERRUPTIBLE
The thread is in an uninterruptible wait.

@item TH_STATE_HALTED
The thread is halted at a clean point.
@end table

@item flags
Various flags.  The possible values of this field are:
@table @code
@item TH_FLAGS_SWAPPED
The thread is swapped out.

@item TH_FLAGS_IDLE
The thread is an idle thread.
@end table

@item int suspend_count
The suspend count for the thread.

@item int sleep_time
The number of seconds that the thread has been sleeping.

@item time_value_t creation_time
The time stamp of creation.
@end table
@end deftp

@deftp {Data type} thread_basic_info_t
This is a pointer to a @code{struct thread_basic_info}.
@end deftp

@deftp {Data type} {struct thread_sched_info}
This structure is returned in @var{thread_info} by the
@code{thread_info} function and provides schedule information about the
thread.  You can cast a variable of type @code{thread_info_t} to a
pointer of this type if you provided it as the @var{thread_info}
parameter for the @code{THREAD_SCHED_INFO} flavor of @code{thread_info}.
It has the following members:

@table @code
@item int policy
The scheduling policy of the thread, @ref{Scheduling Policy}.

@item integer_t data
Policy-dependent scheduling information, @ref{Scheduling Policy}.

@item int base_priority
The base scheduling priority of the thread.

@item int max_priority
The maximum scheduling priority of the thread.

@item int cur_priority
The current scheduling priority of the thread.

@item int depressed
@code{TRUE} if the thread is depressed.

@item int depress_priority
The priority the thread was depressed from.
@end table
@end deftp

@deftp {Data type} thread_sched_info_t
This is a pointer to a @code{struct thread_sched_info}.
@end deftp


@node Thread Settings
@subsection Thread Settings

@deftypefun kern_return_t thread_wire (@w{host_priv_t @var{host_priv}}, @w{thread_t @var{thread}}, @w{boolean_t @var{wired}})
The function @code{thread_wire} controls the VM privilege level of the
thread @var{thread}.  A VM-privileged thread never waits inside the
kernel for memory allocation from the kernel's free list of pages or for
allocation of a kernel stack.

Threads that are part of the default pageout path should be
VM-privileged, to prevent system deadlocks.  Threads that are not part
of the default pageout path should not be VM-privileged, to prevent the
kernel's free list of pages from being exhausted.

The functions returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{host_priv} or @var{thread} was
invalid.

The @code{thread_wire} call is actually an RPC to @var{host_priv},
normally a send right for a privileged host port, but potentially any
send right.  In addition to the normal diagnostic return codes from the
call's server (normally the kernel), the call may return @code{mach_msg}
return codes.
@c See also: vm_wire(2), vm_set_default_memory_manager(2).
@end deftypefun


@node Thread Execution
@subsection Thread Execution

@deftypefun kern_return_t thread_suspend (@w{thread_t @var{target_thread}})
Increments the thread's suspend count and prevents the thread from
executing any more user level instructions.  In this context a user
level instruction is either a machine instruction executed in user mode
or a system trap instruction including page faults.  Thus if a thread is
currently executing within a system trap the kernel code may continue to
execute until it reaches the system return code or it may suspend within
the kernel code.  In either case, when the thread is resumed the system
trap will return.  This could cause unpredictable results if the user
did a suspend and then altered the user state of the thread in order to
change its direction upon a resume.  The call @code{thread_abort} is
provided to allow the user to abort any system call that is in progress
in a predictable way.

The suspend count may become greater than one with the effect that it
will take more than one resume call to restart the thread.

The function returns @code{KERN_SUCCESS} if the thread has been
suspended and @code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not
a thread.
@end deftypefun

@deftypefun kern_return_t thread_resume (@w{thread_t @var{target_thread}})
Decrements the thread's suspend count.  If the count becomes zero the
thread is resumed.  If it is still positive, the thread is left
suspended.  The suspend count may not become negative.

The function returns @code{KERN_SUCCESS} if the thread has been resumed,
@code{KERN_FAILURE} if the suspend count is already zero and
@code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not a thread.
@end deftypefun

@deftypefun kern_return_t thread_abort (@w{thread_t @var{target_thread}})
The function @code{thread_abort} aborts the kernel primitives:
@code{mach_msg}, @code{msg_send}, @code{msg_receive} and @code{msg_rpc}
and page-faults, making the call return a code indicating that it was
interrupted.  The call is interrupted whether or not the thread (or task
containing it) is currently suspended.  If it is suspended, the thread
receives the interrupt when it is resumed.

A thread will retry an aborted page-fault if its state is not modified
before it is resumed.  @code{msg_send} returns @code{SEND_INTERRUPTED};
@code{msg_receive} returns @code{RCV_INTERRUPTED}; @code{msg_rpc}
returns either @code{SEND_INTERRUPTED} or @code{RCV_INTERRUPTED},
depending on which half of the RPC was interrupted.

The main reason for this primitive is to allow one thread to cleanly
stop another thread in a manner that will allow the future execution of
the target thread to be controlled in a predictable way.
@code{thread_suspend} keeps the target thread from executing any further
instructions at the user level, including the return from a system call.
@code{thread_get_state}/@code{thread_set_state} allows the examination
or modification of the user state of a target thread.  However, if a
suspended thread was executing within a system call, it also has
associated with it a kernel state.  This kernel state can not be
modified by @code{thread_set_state} with the result that when the thread
is resumed the system call may return changing the user state and
possibly user memory.  @code{thread_abort} aborts the kernel call from
the target thread's point of view by resetting the kernel state so that
the thread will resume execution at the system call return with the
return code value set to one of the interrupted codes.  The system call
itself will either be entirely completed or entirely aborted, depending
on the precise moment at which the abort was received.  Thus if the
thread's user state has been changed by @code{thread_set_state}, it will
not be modified by any unexpected system call side effects.

For example to simulate a Unix signal, the following sequence of calls
may be used:

@enumerate
@item
@code{thread_suspend}: Stops the thread.

@item
@code{thread_abort}: Interrupts any system call in progress, setting the
return value to `interrupted'.  Since the thread is stopped, it will not
return to user code.

@item
@code{thread_set_state}: Alters thread's state to simulate a procedure
call to the signal handler

@item
@code{thread_resume}: Resumes execution at the signal handler.  If the
thread's stack has been correctly set up, the thread may return to the
interrupted system call.  (Of course, the code to push an extra stack
frame and change the registers is VERY machine-dependent.)
@end enumerate

Calling @code{thread_abort} on a non-suspended thread is pretty risky,
since it is very difficult to know exactly what system trap, if any, the
thread might be executing and whether an interrupt return would cause
the thread to do something useful.

The function returns @code{KERN_SUCCESS} if the thread received an
interrupt and @code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not
a thread.
@end deftypefun

@deftypefun kern_return_t thread_get_state (@w{thread_t @var{target_thread}}, @w{int @var{flavor}}, @w{thread_state_t @var{old_state}}, @w{mach_msg_type_number_t *@var{old_stateCnt}})
The function @code{thread_get_state} returns the execution state
(e.g. the machine registers) of @var{target_thread} as specified by
@var{flavor}.  The @var{old_state} is an array of integers that is
provided by the caller and returned filled with the specified
information.  @var{old_stateCnt} is input set to the maximum number of
integers in @var{old_state} and returned equal to the actual number of
integers in @var{old_state}.

@var{target_thread} may not be @code{mach_thread_self()}.

The definition of the state structures can be found in
@file{machine/thread_status.h}.

The function returns @code{KERN_SUCCESS} if the state has been returned,
@code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not a thread or
is @code{mach_thread_self} or @var{flavor} is unrecognized for this machine.
The function returns @code{MIG_ARRAY_TOO_LARGE} if the returned state is
too large for @var{old_state}.  In this case, @var{old_state} is filled
as much as possible and @var{old_stateCnt} is set to the number of
elements that would have been returned if there were enough room.
@end deftypefun

@deftypefun kern_return_t thread_set_state (@w{thread_t @var{target_thread}}, @w{int @var{flavor}}, @w{thread_state_t @var{new_state}}, @w{mach_msg_type_number_t @var{new_state_count}})
The function @code{thread_set_state} sets the execution state (e.g. the
machine registers) of @var{target_thread} as specified by @var{flavor}.
The @var{new_state} is an array of integers.  @var{new_state_count} is
the number of elements in @var{new_state}.  The entire set of registers
is reset.  This will do unpredictable things if @var{target_thread} is
not suspended.

@var{target_thread} may not be @code{mach_thread_self}.

The definition of the state structures can be found in
@file{machine/thread_status.h}.

The function returns @code{KERN_SUCCESS} if the state has been set and
@code{KERN_INVALID_ARGUMENT} if @var{target_thread} is not a thread or
is @code{mach_thread_self} or @var{flavor} is unrecognized for this
machine.
@end deftypefun


@node Scheduling
@subsection Scheduling

@menu
* Thread Priority::               Changing the priority of a thread.
* Hand-Off Scheduling::           Switching to a new thread.
* Scheduling Policy::             Setting the scheduling policy.
@end menu


@node Thread Priority
@subsubsection Thread Priority

Threads have three priorities associated with them by the system, a
priority, a maximum priority, and a scheduled priority.  The scheduled
priority is used to make scheduling decisions about the thread.  It is
determined from the priority by the policy (for timesharing, this means
adding an increment derived from cpu usage).  The priority can be set
under user control, but may never exceed the maximum priority.  Changing
the maximum priority requires presentation of the control port for the
thread's processor set; since the control port for the default processor
set is privileged, users cannot raise their maximum priority to unfairly
compete with other users on that set.  Newly created threads obtain
their priority from their task and their max priority from the thread.

@deftypefun kern_return_t thread_priority (@w{thread_t @var{thread}}, @w{int @var{prority}}, @w{boolean_t @var{set_max}})
The function @code{thread_priority} changes the priority and optionally
the maximum priority of @var{thread}.  Priorities range from 0 to 31,
where lower numbers denote higher priorities.  If the new priority is
higher than the priority of the current thread, preemption may occur as
a result of this call.  The maximum priority of the thread is also set
if @var{set_max} is @code{TRUE}.  This call will fail if @var{priority}
is greater than the current maximum priority of the thread.  As a
result, this call can only lower the value of a thread's maximum
priority.

The functions returns @code{KERN_SUCCESS} if the operation completed
successfully, @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a
thread or @var{priority} is out of range (not in 0..31), and
@code{KERN_FAILURE} if the requested operation would violate the
thread's maximum priority (thread_priority).
@end deftypefun

@deftypefun kern_return_t thread_max_priority (@w{thread_t @var{thread}}, @w{processor_set_t @var{processor_set}}, @w{int @var{priority}})
The function @code{thread_max_priority} changes the maximum priority of
the thread.  Because it requires presentation of the corresponding
processor set port, this call can reset the maximum priority to any
legal value.

The functions returns @code{KERN_SUCCESS} if the operation completed
successfully, @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a
thread or @var{processor_set} is not a control port for a processor set
or @var{priority} is out of range (not in 0..31), and
@code{KERN_FAILURE} if the thread is not assigned to the processor set
whose control port was presented.
@end deftypefun


@node Hand-Off Scheduling
@subsubsection Hand-Off Scheduling

@deftypefun kern_return_t thread_switch (@w{thread_t @var{new_thread}}, @w{int @var{option}}, @w{int @var{time}})
The function @code{thread_switch} provides low-level access to the
scheduler's context switching code.  @var{new_thread} is a hint that
implements hand-off scheduling.  The operating system will attempt to
switch directly to the new thread (bypassing the normal logic that
selects the next thread to run) if possible.  Since this is a hint, it
may be incorrect; it is ignored if it doesn't specify a thread on the
same host as the current thread or if that thread can't be switched to
(i.e., not runnable or already running on another processor or giving
a plainly invalid hint, such as @code{MACH_PORT_NULL}).  In this case,
the normal logic to select the next thread to run is used; the current
thread may continue running if there is no other appropriate thread to
run.

Options for @var{option} are defined in @file{mach/thread_switch.h} and
specify the interpretation of @var{time}.  The possible values for
@var{option} are:

@table @code
@item SWITCH_OPTION_NONE
No options, the time argument is ignored.

@item SWITCH_OPTION_WAIT
The thread is blocked for the specified time (in milliseconds;
specifying @code{0} will wait for the next tick).  This can be aborted
by @code{thread_abort}.

@item SWITCH_OPTION_DEPRESS
The thread's priority is depressed to the lowest possible value for the
specified time.  This can be aborted by @code{thread_depress_abort}.
This depression is independent of operations that change the thread's
priority (e.g. @code{thread_priority} will not abort the depression).
The minimum time and units of time can be obtained as the
@code{min_timeout} value from @code{host_info}.  The depression is also
aborted when the current thread is next run (either via hand-off
scheduling or because the processor set has nothing better to do).
@end table

@code{thread_switch} is often called when the current thread can proceed
no further for some reason; the various options and arguments allow
information about this reason to be transmitted to the kernel.  The
@var{new_thread} argument (handoff scheduling) is useful when the
identity of the thread that must make progress before the current thread
runs again is known.  The @code{WAIT} option is used when the amount of
time that the current thread must wait before it can do anything useful
can be estimated and is fairly long.  The @code{DEPRESS} option is used
when the amount of time that must be waited is fairly short, especially
when the identity of the thread that is being waited for is not known.

Users should beware of calling @code{thread_switch} with an invalid hint
(e.g. @code{MACH_PORT_NULL}) and no option.  Because the time-sharing
scheduler varies the priority of threads based on usage, this may result
in a waste of cpu time if the thread that must be run is of lower
priority.  The use of the @code{DEPRESS} option in this situation is
highly recommended.

@code{thread_switch} ignores policies.  Users relying on the preemption
semantics of a fixed time policy should be aware that
@code{thread_switch} ignores these semantics; it will run the specified
@var{new_thread} independent of its priority and the priority of any other
threads that could be run instead.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a thread or
@var{option} is not a recognized option, and @code{KERN_FAILURE} if
@code{kern_depress_abort} failed because the thread was not depressed.
@end deftypefun

@deftypefun kern_return_t thread_depress_abort (@w{thread_t @var{thread}})
The function @code{thread_depress_abort} cancels any priority depression
for @var{thread} caused by a @code{swtch_pri} or @code{thread_switch}
call.

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a valid thread.
@end deftypefun

@deftypefun boolean_t swtch ()
@c XXX Clear up wording.
The system trap @code{swtch} attempts to switch the current thread off
the processor.  The return value indicates if more than the current
thread is running in the processor set.  This is useful for lock
management routines.

The call returns @code{FALSE} if the thread is justified in becoming a
resource hog by continuing to spin because there's nothing else useful
that the processor could do.  @code{TRUE} is returned if the thread
should make one more check on the lock and then be a good citizen and
really suspend.
@end deftypefun

@deftypefun boolean_t swtch_pri (@w{int @var{priority}})
The system trap @code{swtch_pri} attempts to switch the current thread
off the processor as @code{swtch} does, but depressing the priority of
the thread to the minimum possible value during the time.
@var{priority} is not used currently.

The return value is as for @code{swtch}.
@end deftypefun


@node Scheduling Policy
@subsubsection Scheduling Policy

@deftypefun kern_return_t thread_policy (@w{thread_t @var{thread}}, @w{int @var{policy}}, @w{int @var{data}})
The function @code{thread_policy} changes the scheduling policy for
@var{thread} to @var{policy}.

@var{data} is policy-dependent scheduling information.  There are
currently two supported policies: @code{POLICY_TIMESHARE} and
@code{POLICY_FIXEDPRI} defined in @file{mach/policy.h}; this file is
included by @file{mach.h}.  @var{data} is meaningless for timesharing,
but is the quantum to be used (in milliseconds) for the fixed priority
policy.  To be meaningful, this quantum must be a multiple of the basic
system quantum (min_quantum) which can be obtained from
@code{host_info}.  The system will always round up to the next multiple
of the quantum.

Processor sets may restrict the allowed policies, so this call will fail
if the processor set to which @var{thread} is currently assigned does
not permit @var{policy}.

The function returns @code{KERN_SUCCESS} if the call succeeded.
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a thread or
@var{policy} is not a recognized policy, and @code{KERN_FAILURE} if the
processor set to which @var{thread} is currently assigned does not
permit @var{policy}.
@end deftypefun


@node Thread Special Ports
@subsection Thread Special Ports

@deftypefun kern_return_t thread_get_special_port (@w{thread_t @var{thread}}, @w{int @var{which_port}}, @w{mach_port_t *@var{special_port}})
The function @code{thread_get_special_port} returns send rights to one
of a set of special ports for the thread specified by @var{thread}.

The possible values for @var{which_port} are @code{THREAD_KERNEL_PORT}
and @code{THREAD_EXCEPTION_PORT}.  A thread also has access to its
task's special ports.

The function returns @code{KERN_SUCCESS} if the port was returned and
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a thread or
@var{which_port} is an invalid port selector.
@end deftypefun

@deftypefun kern_return_t thread_get_kernel_port (@w{thread_t @var{thread}}, @w{mach_port_t *@var{kernel_port}})
The function @code{thread_get_kernel_port} is equivalent to the function
@code{thread_get_special_port} with the @var{which_port} argument set to
@code{THREAD_KERNEL_PORT}.
@end deftypefun

@deftypefun kern_return_t thread_get_exception_port (@w{thread_t @var{thread}}, @w{mach_port_t *@var{exception_port}})
The function @code{thread_get_exception_port} is equivalent to the
function @code{thread_get_special_port} with the @var{which_port}
argument set to @code{THREAD_EXCEPTION_PORT}.
@end deftypefun

@deftypefun kern_return_t thread_set_special_port (@w{thread_t @var{thread}}, @w{int @var{which_port}}, @w{mach_port_t @var{special_port}})
The function @code{thread_set_special_port} sets one of a set of special
ports for the thread specified by @var{thread}.

The possible values for @var{which_port} are @code{THREAD_KERNEL_PORT}
and @code{THREAD_EXCEPTION_PORT}.  A thread also has access to its
task's special ports.

The function returns @code{KERN_SUCCESS} if the port was set and
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a thread or
@var{which_port} is an invalid port selector.
@end deftypefun

@deftypefun kern_return_t thread_set_kernel_port (@w{thread_t @var{thread}}, @w{mach_port_t @var{kernel_port}})
The function @code{thread_set_kernel_port} is equivalent to the function
@code{thread_set_special_port} with the @var{which_port} argument set to
@code{THREAD_KERNEL_PORT}.
@end deftypefun

@deftypefun kern_return_t thread_set_exception_port (@w{thread_t @var{thread}}, @w{mach_port_t @var{exception_port}})
The function @code{thread_set_exception_port} is equivalent to the
function @code{thread_set_special_port} with the @var{which_port}
argument set to @code{THREAD_EXCEPTION_PORT}.
@end deftypefun


@node Exceptions
@subsection Exceptions

@deftypefun kern_return_t catch_exception_raise (@w{mach_port_t @var{exception_port}}, @w{thread_t @var{thread}}, @w{task_t @var{task}}, @w{int @var{exception}}, @w{int @var{code}}, @w{int @var{subcode}})
XXX Fixme
@end deftypefun

@deftypefun kern_return_t exception_raise (@w{mach_port_t @var{exception_port}}, @w{mach_port_t @var{thread}}, @w{mach_port_t @var{task}}, @w{integer_t @var{exception}}, @w{integer_t @var{code}}, @w{integer_t @var{subcode}})
XXX Fixme
@end deftypefun

@deftypefun kern_return_t evc_wait (@w{unsigned int @var{event}})
@c XXX This is for user space drivers, the description is incomplete.
The system trap @code{evc_wait} makes the calling thread wait for the
event specified by @var{event}.

The call returns @code{KERN_SUCCESS} if the event has occurred,
@code{KERN_NO_SPACE} if another thread is waiting for the same event and
@code{KERN_INVALID_ARGUMENT} if the event object is invalid.
@end deftypefun


@node Task Interface
@section Task Interface

@cindex task port
@cindex port representing a task
@deftp {Data type} task_t
This is a @code{mach_port_t} and used to hold the port name of a task
port that represents the thread.  Manipulations of the task are
implemented as remote procedure calls to the task port.  A task can get
a port to itself with the @code{mach_task_self} system call.

The task port name is also used to identify the task's IPC space
(@pxref{Port Manipulation Interface}) and the task's virtual memory map
(@pxref{Virtual Memory Interface}).
@end deftp

@menu
* Task Creation::                 Creating tasks.
* Task Termination::              Terminating tasks.
* Task Information::              Informations on tasks.
* Task Execution::                Thread scheduling in a task.
* Task Special Ports::            How to get and set the task's special ports.
* Syscall Emulation::             How to emulate system calls.
@end menu


@node Task Creation
@subsection Task Creation

@deftypefun kern_return_t task_create (@w{task_t @var{parent_task}}, @w{boolean_t @var{inherit_memory}}, @w{task_t *@var{child_task}})
The function @code{task_create} creates a new task from
@var{parent_task}; the resulting task (@var{child_task}) acquires shared
or copied parts of the parent's address space (see @code{vm_inherit}).
The child task initially contains no threads.

If @var{inherit_memory} is set, the child task's address space is built
from the parent task according to its memory inheritance values;
otherwise, the child task is given an empty address space.

The child task gets the three special ports created or copied for it at
task creation.  The @code{TASK_KERNEL_PORT} is created and send rights
for it are given to the child and returned to the caller.
@c The following is only relevant if MACH_IPC_COMPAT is used.
@c  The @code{TASK_NOTIFY_PORT} is created and receive, ownership and send rights
@c  for it are given to the child.  The caller has no access to it.
The @code{TASK_BOOTSTRAP_PORT} and the @code{TASK_EXCEPTION_PORT} are
inherited from the parent task.  The new task can get send rights to
these ports with the call @code{task_get_special_port}.

The function returns @code{KERN_SUCCESS} if a new task has been created,
@code{KERN_INVALID_ARGUMENT} if @var{parent_task} is not a valid task
port and @code{KERN_RESOURCE_SHORTAGE} if some critical kernel resource
is unavailable.
@end deftypefun


@node Task Termination
@subsection Task Termination

@deftypefun kern_return_t task_terminate (@w{task_t @var{target_task}})
The function @code{task_terminate} destroys the task specified by
@var{target_task} and all its threads.  All resources that are used only
by this task are freed.  Any port to which this task has receive and
ownership rights is destroyed.

The function returns @code{KERN_SUCCESS} if the task has been killed,
@code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a task.
@end deftypefun


@node Task Information
@subsection Task Information
@deftypefun task_t mach_task_self ()
The @code{mach_task_self} system call returns the calling thread's task
port.

@code{mach_task_self} has an effect equivalent to receiving a send right
for the task port.  @code{mach_task_self} returns the name of the send
right.  In particular, successive calls will increase the calling task's
user-reference count for the send right.

As a special exception, the kernel will overrun the user reference count
of the task name port, so that this function can not fail for that
reason.  Because of this, the user should not deallocate the port right
if an overrun might have happened.  Otherwise the reference count could
drop to zero and the send right be destroyed while the user still
expects to be able to use it.  As the kernel does not make use of the
number of extant send rights anyway, this is safe to do (the task port
itself is not destroyed, even when there are no send rights anymore).

The function returns @code{MACH_PORT_NULL} if a resource shortage
prevented the reception of the send right, @code{MACH_PORT_NULL} if the
task port is currently null, @code{MACH_PORT_DEAD} if the task port is
currently dead.
@end deftypefun

@deftypefun kern_return_t task_threads (@w{task_t @var{target_task}}, @w{thread_array_t *@var{thread_list}}, @w{mach_msg_type_number_t *@var{thread_count}})
The function @code{task_threads} gets send rights to the kernel port for
each thread contained in @var{target_task}.  @var{thread_list} is an
array that is created as a result of this call.  The caller may wish to
@code{vm_deallocate} this array when the data is no longer needed.

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a task.
@end deftypefun

@deftypefun kern_return_t task_info (@w{task_t @var{target_task}}, @w{int @var{flavor}}, @w{task_info_t @var{task_info}}, @w{mach_msg_type_number_t *@var{task_info_count}})
The function @code{task_info} returns the selected information array for
a task, as specified by @var{flavor}.  @var{task_info} is an array of
integers that is supplied by the caller, and filled with specified
information.  @var{task_info_count} is supplied as the maximum number of
integers in @var{task_info}.  On return, it contains the actual number
of integers in @var{task_info}.  The maximum number of integers returned
by any flavor is @code{TASK_INFO_MAX}.

The type of information returned is defined by @var{flavor}, which can
be one of the following:

@table @code
@item TASK_BASIC_INFO
The function returns basic information about the task, as defined by
@code{task_basic_info_t}.  This includes the user and system time and
memory consumption.  The number of integers returned is
@code{TASK_BASIC_INFO_COUNT}.

@item TASK_EVENTS_INFO
The function returns information about events for the task as defined by
@code{thread_sched_info_t}.  This includes statistics about virtual
memory and IPC events like pageouts, pageins and messages sent and
received.  The number of integers returned is
@code{TASK_EVENTS_INFO_COUNT}.

@item TASK_THREAD_TIMES_INFO
The function returns information about the total time for live threads
as defined by @code{task_thread_times_info_t}.  The number of integers
returned is @code{TASK_THREAD_TIMES_INFO_COUNT}.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a thread or
@var{flavor} is not recognized.  The function returns
@code{MIG_ARRAY_TOO_LARGE} if the returned info array is too large for
@var{task_info}.  In this case, @var{task_info} is filled as much as
possible and @var{task_infoCnt} is set to the number of elements that
would have been returned if there were enough room.
@end deftypefun

@deftp {Data type} {struct task_basic_info}
This structure is returned in @var{task_info} by the @code{task_info}
function and provides basic information about the task.  You can cast a
variable of type @code{task_info_t} to a pointer of this type if you
provided it as the @var{task_info} parameter for the
@code{TASK_BASIC_INFO} flavor of @code{task_info}.  It has the following
members:

@table @code
@item integer_t suspend_count
suspend count for task

@item integer_t base_priority
base scheduling priority

@item vm_size_t virtual_size
number of virtual pages

@item vm_size_t resident_size
number of resident pages

@item time_value_t user_time
total user run time for terminated threads

@item time_value_t system_time
total system run time for terminated threads

@item time_value_t creation_time
creation time stamp
@end table
@end deftp

@deftp {Data type} task_basic_info_t
This is a pointer to a @code{struct task_basic_info}.
@end deftp

@deftp {Data type} {struct task_events_info}
This structure is returned in @var{task_info} by the @code{task_info}
function and provides event statistics for the task.  You can cast a
variable of type @code{task_info_t} to a pointer of this type if you
provided it as the @var{task_info} parameter for the
@code{TASK_EVENTS_INFO} flavor of @code{task_info}.  It has the
following members:

@table @code
@item natural_t faults
number of page faults

@item natural_t zero_fills
number of zero fill pages

@item natural_t reactivations
number of reactivated pages

@item natural_t pageins
number of actual pageins

@item natural_t cow_faults
number of copy-on-write faults

@item natural_t messages_sent
number of messages sent

@item natural_t messages_received
number of messages received
@end table
@end deftp

@deftp {Data type} task_events_info_t
This is a pointer to a @code{struct task_events_info}.
@end deftp

@deftp {Data type} {struct task_thread_times_info}
This structure is returned in @var{task_info} by the @code{task_info}
function and provides event statistics for the task.  You can cast a
variable of type @code{task_info_t} to a pointer of this type if you
provided it as the @var{task_info} parameter for the
@code{TASK_THREAD_TIMES_INFO} flavor of @code{task_info}.  It has the
following members:

@table @code
@item time_value_t user_time
total user run time for live threads

@item time_value_t system_time
total system run time for live threads
@end table
@end deftp

@deftp {Data type} task_thread_times_info_t
This is a pointer to a @code{struct task_thread_times_info}.
@end deftp


@node Task Execution
@subsection Task Execution

@deftypefun kern_return_t task_suspend (@w{task_t @var{target_task}})
The function @code{task_suspend} increments the task's suspend count and
stops all threads in the task.  As long as the suspend count is positive
newly created threads will not run.  This call does not return until all
threads are suspended.

The count may become greater than one, with the effect that it will take
more than one resume call to restart the task.

The function returns @code{KERN_SUCCESS} if the task has been suspended
and @code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a task.
@end deftypefun

@deftypefun kern_return_t task_resume (@w{task_t @var{target_task}})
The function @code{task_resume} decrements the task's suspend count.  If
it becomes zero, all threads with zero suspend counts in the task are
resumed.  The count may not become negative.

The function returns @code{KERN_SUCCESS} if the task has been resumed,
@code{KERN_FAILURE} if the suspend count is already at zero and
@code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a task.
@end deftypefun

@c XXX Should probably be in the "Scheduling" node of the Thread Interface.
@deftypefun kern_return_t task_priority (@w{task_t @var{task}}, @w{int @var{priority}}, @w{boolean_t @var{change_threads}})
The priority of a task is used only for creation of new threads; a new
thread's priority is set to the enclosing task's priority.
@code{task_priority} changes this task priority.  It also sets the
priorities of all threads in the task to this new priority if
@var{change_threads} is @code{TRUE}.  Existing threads are not affected
otherwise.  If this priority change violates the maximum priority of
some threads, as many threads as possible will be changed and an error
code will be returned.

The function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{task} is not a task, or
@var{priority} is not a valid priority and @code{KERN_FAILURE} if
@var{change_threads} was @code{TRUE} and the attempt to change the
priority of at least one existing thread failed because the new priority
would have exceeded that thread's maximum priority.
@end deftypefun

@deftypefun kern_return_t task_ras_control (@w{task_t @var{target_task}}, @w{vm_address_t @var{start_pc}}, @w{vm_address_t @var{end_pc}}, @w{int @var{flavor}})
The function @code{task_ras_control} manipulates a task's set of
restartable atomic sequences.  If a sequence is installed, and any
thread in the task is preempted within the range
[@var{start_pc},@var{end_pc}], then the thread is resumed at
@var{start_pc}.  This enables applications to build atomic sequences
which, when executed to completion, will have executed atomically.
Restartable atomic sequences are intended to be used on systems that do
not have hardware support for low-overhead atomic primitives.

As a thread can be rolled-back, the code in the sequence should have no
side effects other than a final store at @var{end_pc}. The kernel does
not guarantee that the sequence is restartable.  It assumes the
application knows what it's doing.

A task may have a finite number of atomic sequences that is defined at
compile time.

The flavor specifies the particular operation that should be applied to
this restartable atomic sequence.  Possible values for flavor can be:

@table @code
@item TASK_RAS_CONTROL_PURGE_ALL
Remove all registered sequences for this task.

@item TASK_RAS_CONTROL_PURGE_ONE
Remove the named registered sequence for this task.

@item TASK_RAS_CONTROL_PURGE_ALL_AND_INSTALL_ONE
Atomically remove all registered sequences and install the named
sequence.

@item TASK_RAS_CONTROL_INSTALL_ONE
Install this sequence.
@end table

The function returns @code{KERN_SUCCESS} if the operation has been
performed, @code{KERN_INVALID_ADDRESS} if the @var{start_pc} or
@var{end_pc} values are not a valid address for the requested operation
(for example, it is invalid to purge a sequence that has not been
registered), @code{KERN_RESOURCE_SHORTAGE} if an attempt was made to
install more restartable atomic sequences for a task than can be
supported by the kernel, @code{KERN_INVALID_VALUE} if a bad flavor was
specified, @code{KERN_INVALID_ARGUMENT} if @var{target_task} is not a
task and @code{KERN_FAILURE} if the call is not not supported on this
configuration.
@end deftypefun


@node Task Special Ports
@subsection Task Special Ports

@deftypefun kern_return_t task_get_special_port (@w{task_t @var{task}}, @w{int @var{which_port}}, @w{mach_port_t *@var{special_port}})
The function @code{task_get_special_port} returns send rights to one of
a set of special ports for the task specified by @var{task}.

The special ports associated with a task are the kernel port
(@code{TASK_KERNEL_PORT}), the bootstrap port
(@code{TASK_BOOTSTRAP_PORT}) and the exception port
(@code{TASK_EXCEPTION_PORT}).  The bootstrap port is a port to which a
task may send a message requesting other system service ports.  This
port is not used by the kernel.  The task's exception port is the port
to which messages are sent by the kernel when an exception occurs and
the thread causing the exception has no exception port of its own.

The following macros to call @code{task_get_special_port} for a specific
port are defined in @code{mach/task_special_ports.h}:
@code{task_get_exception_port} and @code{task_get_bootstrap_port}.

The function returns @code{KERN_SUCCESS} if the port was returned and
@code{KERN_INVALID_ARGUMENT} if @var{task} is not a task or
@var{which_port} is an invalid port selector.
@end deftypefun

@deftypefun kern_return_t task_get_kernel_port (@w{task_t @var{task}}, @w{mach_port_t *@var{kernel_port}})
The function @code{task_get_kernel_port} is equivalent to the function
@code{task_get_special_port} with the @var{which_port} argument set to
@code{TASK_KERNEL_PORT}.
@end deftypefun

@deftypefun kern_return_t task_get_exception_port (@w{task_t @var{task}}, @w{mach_port_t *@var{exception_port}})
The function @code{task_get_exception_port} is equivalent to the
function @code{task_get_special_port} with the @var{which_port} argument
set to @code{TASK_EXCEPTION_PORT}.
@end deftypefun

@deftypefun kern_return_t task_get_bootstrap_port (@w{task_t @var{task}}, @w{mach_port_t *@var{bootstrap_port}})
The function @code{task_get_bootstrap_port} is equivalent to the
function @code{task_get_special_port} with the @var{which_port} argument
set to @code{TASK_BOOTSTRAP_PORT}.
@end deftypefun

@deftypefun kern_return_t task_set_special_port (@w{task_t @var{task}}, @w{int @var{which_port}}, @w{mach_port_t @var{special_port}})
The function @code{thread_set_special_port} sets one of a set of special
ports for the task specified by @var{task}.

The special ports associated with a task are the kernel port
(@code{TASK_KERNEL_PORT}), the bootstrap port
(@code{TASK_BOOTSTRAP_PORT}) and the exception port
(@code{TASK_EXCEPTION_PORT}).  The bootstrap port is a port to which a
thread may send a message requesting other system service ports.  This
port is not used by the kernel.  The task's exception port is the port
to which messages are sent by the kernel when an exception occurs and
the thread causing the exception has no exception port of its own.

The function returns @code{KERN_SUCCESS} if the port was set and
@code{KERN_INVALID_ARGUMENT} if @var{task} is not a task or
@var{which_port} is an invalid port selector.
@end deftypefun

@deftypefun kern_return_t task_set_kernel_port (@w{task_t @var{task}}, @w{mach_port_t @var{kernel_port}})
The function @code{task_set_kernel_port} is equivalent to the function
@code{task_set_special_port} with the @var{which_port} argument set to
@code{TASK_KERNEL_PORT}.
@end deftypefun

@deftypefun kern_return_t task_set_exception_port (@w{task_t @var{task}}, @w{mach_port_t @var{exception_port}})
The function @code{task_set_exception_port} is equivalent to the
function @code{task_set_special_port} with the @var{which_port} argument
set to @code{TASK_EXCEPTION_PORT}.
@end deftypefun

@deftypefun kern_return_t task_set_bootstrap_port (@w{task_t @var{task}}, @w{mach_port_t @var{bootstrap_port}})
The function @code{task_set_bootstrap_port} is equivalent to the
function @code{task_set_special_port} with the @var{which_port} argument
set to @code{TASK_BOOTSTRAP_PORT}.
@end deftypefun


@node Syscall Emulation
@subsection Syscall Emulation

@deftypefun kern_return_t task_get_emulation_vector (@w{task_t @var{task}}, @w{int *@var{vector_start}}, @w{emulation_vector_t *@var{emulation_vector}}, @w{mach_msg_type_number_t *@var{emulation_vector_count}})
The function @code{task_get_emulation_vector} gets the user-level
handler entry points for all emulated system calls.
@c XXX Fixme
@end deftypefun

@deftypefun kern_return_t task_set_emulation_vector (@w{task_t @var{task}}, @w{int @var{vector_start}}, @w{emulation_vector_t @var{emulation_vector}}, @w{mach_msg_type_number_t @var{emulation_vector_count}})
The function @code{task_set_emulation_vector} establishes user-level
handlers for the specified system calls.  Non-emulated system calls are
specified with an entry of @code{EML_ROUTINE_NULL}.  System call
emulation handlers are inherited by the children of @var{task}.
@c XXX Fixme
@end deftypefun

@deftypefun kern_return_t task_set_emulation (@w{task_t @var{task}}, @w{vm_address_t @var{routine_entry_pt}}, @w{int @var{routine_number}})
The function @code{task_set_emulation} establishes a user-level handler
for the specified system call.  System call emulation handlers are
inherited by the children of @var{task}.
@c XXX Fixme
@end deftypefun

@c XXX Fixme datatype emulation_vector_t


@node Profiling
@section Profiling

@deftypefun kern_return_t task_enable_pc_sampling (@w{task_t @var{task}}, @w{int *@var{ticks}}, @w{sampled_pc_flavor_t @var{flavor}})
@deftypefunx kern_return_t thread_enable_pc_sampling (@w{thread_t @var{thread}}, @w{int *@var{ticks}}, @w{sampled_pc_flavor_t @var{flavor}})
The function @code{task_enable_pc_sampling} enables PC sampling for
@var{task}, the function @code{thread_enable_pc_sampling} enables PC
sampling for @var{thread}.  The kernel's idea of clock granularity is
returned in @var{ticks} in usecs. (this value should not be trusted).  The
sampling flavor is specified by @var{flavor}.

The function returns @code{KERN_SUCCESS} if the operation is completed successfully
and @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a valid thread.
@end deftypefun

@deftypefun kern_return_t task_disable_pc_sampling (@w{task_t @var{task}}, @w{int *@var{sample_count}})
@deftypefunx kern_return_t thread_disable_pc_sampling (@w{thread_t @var{thread}}, @w{int *@var{sample_count}})
The function @code{task_disable_pc_sampling} disables PC sampling for
@var{task}, the function @code{thread_disable_pc_sampling} disables PC
sampling for @var{thread}.  The number of sample elements in the kernel
for the thread is returned in @var{sample_count}.

The function returns @code{KERN_SUCCESS} if the operation is completed successfully
and @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a valid thread.
@end deftypefun

@deftypefun kern_return_t task_get_sampled_pcs (@w{task_t @var{task}}, @w{sampled_pc_seqno_t *@var{seqno}}, @w{sampled_pc_array_t @var{sampled_pcs}}, @w{mach_msg_type_number_t *@var{sample_count}})
@deftypefunx kern_return_t thread_get_sampled_pcs (@w{thread_t @var{thread}}, @w{sampled_pc_seqno_t *@var{seqno}}, @w{sampled_pc_array_t @var{sampled_pcs}}, @w{int *@var{sample_count}})
The function @code{task_get_sampled_pcs} extracts the PC samples for
@var{task}, the function @code{thread_get_sampled_pcs} extracts the PC
samples for @var{thread}.  @var{seqno} is the sequence number of the
sampled PCs.  This is useful for determining when a collector thread has
missed a sample.  The sampled PCs for the thread are returned in
@var{sampled_pcs}.  @var{sample_count} contains the number of sample
elements returned.

The function returns @code{KERN_SUCCESS} if the operation is completed successfully,
@code{KERN_INVALID_ARGUMENT} if @var{thread} is not a valid thread and
@code{KERN_FAILURE} if @var{thread} is not sampled.
@end deftypefun


@deftp {Data type} sampled_pc_t
This structure is returned in @var{sampled_pcs} by the
@code{thread_get_sampled_pcs} and @code{task_get_sampled_pcs} functions
and provides pc samples for threads or tasks.  It has the following
members:

@table @code
@item natural_t id
A thread-specific unique identifier.

@item vm_offset_t pc
A pc value.

@item sampled_pc_flavor_t sampletype
The type of the sample as per flavor.
@end table
@end deftp


@deftp {Data type} sampled_pc_flavor_t
This data type specifies a pc sample flavor, either as argument passed
in @var{flavor} to the @code{thread_enable_pc_sample} and
@code{thread_disable_pc_sample} functions, or as member
@code{sampletype} in the @code{sample_pc_t} data type.  The flavor is a
bitwise-or of the possible flavors defined in @file{mach/pc_sample.h}:

@table @code
@item SAMPLED_PC_PERIODIC
default
@item SAMPLED_PC_VM_ZFILL_FAULTS
zero filled fault
@item SAMPLED_PC_VM_REACTIVATION_FAULTS
reactivation fault
@item SAMPLED_PC_VM_PAGEIN_FAULTS
pagein fault
@item SAMPLED_PC_VM_COW_FAULTS
copy-on-write fault
@item SAMPLED_PC_VM_FAULTS_ANY
any fault
@item SAMPLED_PC_VM_FAULTS
the bitwise-or of @code{SAMPLED_PC_VM_ZFILL_FAULTS},
@code{SAMPLED_PC_VM_REACTIVATION_FAULTS},
@code{SAMPLED_PC_VM_PAGEIN_FAULTS} and @code{SAMPLED_PC_VM_COW_FAULTS}.
@end table
@end deftp

@c XXX sampled_pc_array_t, sampled_pc_seqno_t


@node Host Interface
@chapter Host Interface
@cindex host interface

This section describes the Mach interface to a host executing a Mach
kernel.  The interface allows to query statistics about a host and
control its behaviour.

A host is represented by two ports, a name port @var{host} used to query
information about the host accessible to everyone, and a control port
@var{host_priv} used to manipulate it.  For example, you can query the
current time using the name port, but to change the time you need to
send a message to the host control port.

Everything described in this section is declared in the header file
@file{mach.h}.

@menu
* Host Ports::                    Ports representing a host.
* Host Information::              Retrieval of information about a host.
* Host Time::                     Operations on the time as seen by a host.
* Host Reboot::                   Rebooting the system.
@end menu


@node Host Ports
@section Host Ports
@cindex host ports
@cindex ports representing a host

@cindex host name port
@deftp {Data type} host_t 
This is a @code{mach_port_t} and used to hold the port name of a host
name port (or short: host port).  Any task can get a send right to the
name port of the host running the task using the @code{mach_host_self}
system call.  The name port can be used query information about the
host, for example the current time.
@end deftp

@deftypefun host_t mach_host_self ()
The @code{mach_host_self} system call returns the calling thread's host
name port.  It has an effect equivalent to receiving a send right for
the host port.  @code{mach_host_self} returns the name of the send
right.  In particular, successive calls will increase the calling task's
user-reference count for the send right.

As a special exception, the kernel will overrun the user reference count
of the host name port, so that this function can not fail for that
reason.  Because of this, the user should not deallocate the port right
if an overrun might have happened.  Otherwise the reference count could
drop to zero and the send right be destroyed while the user still
expects to be able to use it.  As the kernel does not make use of the
number of extant send rights anyway, this is safe to do (the host port
itself is never destroyed).

The function returns @code{MACH_PORT_NULL} if a resource shortage
prevented the reception of the send right.

This function is also available in @file{mach/mach_traps.h}.
@end deftypefun

@cindex host control port
@deftp {Data type} host_priv_t
This is a @code{mach_port_t} and used to hold the port name of a
privileged host control port.  A send right to the host control port is
inserted into the first task at bootstrap (@pxref{Modules}).  This is
the only way to get access to the host control port in Mach, so the
initial task has to preserve the send right carefully, moving a copy of
it to other privileged tasks if necessary and denying access to
unprivileged tasks.
@end deftp


@node Host Information
@section Host Information

@deftypefun kern_return_t host_info (@w{host_t @var{host}}, @w{int @var{flavor}}, @w{host_info_t @var{host_info}}, @w{mach_msg_type_number_t *@var{host_info_count}})
The @code{host_info} function returns various information about
@var{host}.  @var{host_info} is an array of integers that is supplied by
the caller.  It will be filled with the requested information.
@var{host_info_count} is supplied as the maximum number of integers in
@var{host_info}.  On return, it contains the actual number of integers
in @var{host_info}.  The maximum number of integers returned by any
flavor is @code{HOST_INFO_MAX}.

The type of information returned is defined by @var{flavor}, which can
be one of the following:

@table @code
@item HOST_BASIC_INFO
The function returns basic information about the host, as defined by
@code{host_basic_info_t}.  This includes the number of processors, their
type, and the amount of memory installed in the system.  The number of
integers returned is @code{HOST_BASIC_INFO_COUNT}.  For how to get more
information about the processor, see @ref{Processor Interface}.

@item HOST_PROCESSOR_SLOTS
The function returns the numbers of the slots with active processors in
them.  The number of integers returned can be up to @code{max_cpus}, as
returned by the @code{HOST_BASIC_INFO} flavor of @code{host_info}.

@item HOST_SCHED_INFO
The function returns information of interest to schedulers as defined by
@code{host_sched_info_t}.  The number of integers returned is
@code{HOST_SCHED_INFO_COUNT}.
@end table

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{host} is not a host or @var{flavor}
is not recognized.  The function returns @code{MIG_ARRAY_TOO_LARGE} if
the returned info array is too large for @var{host_info}.  In this case,
@var{host_info} is filled as much as possible and @var{host_info_count}
is set to the number of elements that would be returned if there were
enough room.
@c BUGS Availability limited.  Systems without this call support a
@c host_info call with an incompatible calling sequence.
@end deftypefun

@deftp {Data type} {struct host_basic_info}
A pointer to this structure is returned in @var{host_info} by the
@code{host_info} function and provides basic information about the host.
You can cast a variable of type @code{host_info_t} to a pointer of this
type if you provided it as the @var{host_info} parameter for the
@code{HOST_BASIC_INFO} flavor of @code{host_info}.  It has the following
members:

@table @code
@item int max_cpus
The maximum number of possible processors for which the kernel is
configured.

@item int avail_cpus
The number of cpus currently available.

@item vm_size_t memory_size
The size of physical memory in bytes.

@item cpu_type_t cpu_type
The type of the master processor.

@item cpu_subtype_t cpu_subtype
The subtype of the master processor.
@end table

The type and subtype of the individual processors are also available
by @code{processor_info}, see @ref{Processor Interface}.
@end deftp

@deftp {Data type} host_basic_info_t
This is a pointer to a @code{struct host_basic_info}.
@end deftp

@deftp {Data type} {struct host_sched_info}
A pointer to this structure is returned in @var{host_info} by the
@code{host_info} function and provides information of interest to
schedulers.  You can cast a variable of type @code{host_info_t} to a
pointer of this type if you provided it as the @var{host_info} parameter
for the @code{HOST_SCHED_INFO} flavor of @code{host_info}.  It has the
following members:

@table @code
@item int min_timeout
The minimum timeout and unit of time in milliseconds.

@item int min_quantum
The minimum quantum and unit of quantum in milliseconds.
@end table
@end deftp

@deftp {Data type} host_sched_info_t
This is a pointer to a @code{struct host_sched_info}.
@end deftp

@deftypefun kern_return_t host_kernel_version (@w{host_t @var{host}}, @w{kernel_version_t *@var{version}})
The @code{host_kernel_version} function returns the version string
compiled into the kernel executing on @var{host} at the time it was
built in the character string @var{version}.  This string describes the
version of the kernel.  The constant @code{KERNEL_VERSION_MAX} should be
used to dimension storage for the returned string if the
@code{kernel_version_t} declaration is not used.

If the version string compiled into the kernel is longer than
@code{KERNEL_VERSION_MAX}, the result is truncated and not necessarily
null-terminated.

If @var{host} is not a valid send right to a host port, the function
returns @code{KERN_INVALID_ARGUMENT}.  If @var{version} points to
inaccessible memory, it returns @code{KERN_INVALID_ADDRESS}, and
@code{KERN_SUCCESS} otherwise.
@end deftypefun

@deftypefun kern_return_t host_get_boot_info (@w{host_priv_t @var{host_priv}}, @w{kernel_boot_info_t @var{boot_info}})
The @code{host_get_boot_info} function returns the boot-time information
string supplied by the operator to the kernel executing on
@var{host_priv} in the character string @var{boot_info}.  The constant
@code{KERNEL_BOOT_INFO_MAX} should be used to dimension storage for the
returned string if the @code{kernel_boot_info_t} declaration is not
used.

If the boot-time information string supplied by the operator is longer
than @code{KERNEL_BOOT_INFO_MAX}, the result is truncated and not
necessarily null-terminated.
@end deftypefun


@node Host Time
@section Host Time

@deftp {Data type} time_value_t
This is the representation of a time in Mach.  It is a @code{struct
time_value} and consists of the following members:

@table @code
@item integer_t seconds
The number of seconds.
@item integer_t microseconds
The number of microseconds.
@end table
@end deftp

The number of microseconds should always be smaller than
@code{TIME_MICROS_MAX} (100000).  A time with this property is
@dfn{normalized}.  Normalized time values can be manipulated with the
following macros:

@defmac time_value_add_usec (@w{time_value_t *@var{val}}, @w{integer_t *@var{micros}})
Add @var{micros} microseconds to @var{val}.  If @var{val} is normalized
and @var{micros} smaller than @code{TIME_MICROS_MAX}, @var{val} will be
normalized afterwards.
@end defmac

@defmac time_value_add (@w{time_value_t *@var{result}}, @w{time_value_t *@var{addend}})
Add the values in @var{addend} to @var{result}.  If both are normalized,
@var{result} will be normalized afterwards.
@end defmac

A variable of type @code{time_value_t} can either represent a duration
or a fixed point in time.  In the latter case, it shall be interpreted as
the number of seconds and microseconds after the epoch 1. Jan 1970.

@deftypefun kern_return_t host_get_time (@w{host_t @var{host}}, @w{time_value_t *@var{current_time}})
Get the current time as seen by @var{host}.  On success, the time passed
since the epoch is returned in @var{current_time}.
@end deftypefun

@deftypefun kern_return_t host_set_time (@w{host_priv_t @var{host_priv}}, @w{time_value_t @var{new_time}})
Set the time of @var{host_priv} to @var{new_time}.
@end deftypefun

@deftypefun kern_return_t host_adjust_time (@w{host_priv_t @var{host_priv}}, @w{time_value_t @var{new_adjustment}}, @w{time_value_t *@var{old_adjustment}})
Arrange for the current time as seen by @var{host_priv} to be gradually
changed by the adjustment value @var{new_adjustment}, and return the old
adjustment value in @var{old_adjustment}.
@end deftypefun

For efficiency, the current time is available through a mapped-time
interface.

@deftp {Data type} mapped_time_value_t
This structure defines the mapped-time interface.  It has the following
members:

@table @code
@item integer_t seconds
The number of seconds.

@item integer_t microseconds
The number of microseconds.

@item integer_t check_seconds
This is a copy of the seconds value, which must be checked to protect
against a race condition when reading out the two time values.
@end table
@end deftp

Here is an example how to read out the current time using the
mapped-time interface:

@c XXX Complete the example.
@example
do 
  @{
    secs = mtime->seconds;
    usecs = mtime->microseconds;
  @}
while (secs != mtime->check_seconds);
@end example


@node Host Reboot
@section Host Reboot

@deftypefun kern_return_t host_reboot (@w{host_priv_t @var{host_priv}}, @w{int @var{options}})
Reboot the host specified by @var{host_priv}.  The argument
@var{options} specifies the flags.  The available flags are defined in
@file{sys/reboot.h}:

@table @code
@item RB_HALT
Do not reboot, but halt the machine.

@item RB_DEBUGGER
Do not reboot, but enter kernel debugger from user space.
@end table

If successful, the function might not return.
@end deftypefun


@node Processors and Processor Sets
@chapter Processors and Processor Sets

This section describes the Mach interface to processor sets and
individual processors.  The interface allows to group processors into
sets and control the processors and processor sets.

A processor is not a central part of the interface.  It is mostly of
relevance as a part of a processor set.  Threads are always assigned to
processor sets, and all processors in a set are equally involved in
executing all threads assigned to that set. 

The processor set is represented by two ports, a name port
@var{processor_set_name} used to query information about the host
accessible to everyone, and a control port @var{processor_set} used to
manipulate it.

@menu
* Processor Set Interface::       How to work with processor sets.
* Processor Interface::           How to work with individual processors.
@end menu


@node Processor Set Interface
@section Processor Set Interface

@menu
* Processor Set Ports::           Ports representing a processor set.
* Processor Set Access::          How the processor sets are accessed.
* Processor Set Creation::        How new processor sets are created.
* Processor Set Destruction::     How processor sets are destroyed.
* Tasks and Threads on Sets::     Assigning tasks, threads to processor sets.
* Processor Set Priority::        Specifying the priority of a processor set.
* Processor Set Policy::          Changing the processor set policies.
* Processor Set Info::            Obtaining information about a processor set.
@end menu


@node Processor Set Ports
@subsection Processor Set Ports
@cindex processor set ports
@cindex ports representing a processor set

@cindex processor set name port
@cindex port representing a processor set name
@deftp {Data type} processor_set_name_t 
This is a @code{mach_port_t} and used to hold the port name of a
processor set name port that names the processor set.  Any task can get
a send right to name port of a processor set.  The processor set name
port allows to get information about the processor set.
@end deftp

@cindex processor set port
@deftp {Data type} processor_set_t 
This is a @code{mach_port_t} and used to hold the port name of a
privileged processor set control port that represents the processor set.
Operations on the processor set are implemented as remote procedure
calls to the processor set port.  The processor set port allows to
manipulate the processor set.
@end deftp


@node Processor Set Access
@subsection Processor Set Access

@deftypefun kern_return_t host_processor_sets (@w{host_t @var{host}}, @w{processor_set_name_array_t *@var{processor_sets}}, @w{mach_msg_type_number_t *@var{processor_sets_count}})
The function @code{host_processor_sets} gets send rights to the name
port for each processor set currently assigned to @var{host}.

@code{host_processor_set_priv} can be used to obtain the control ports
from these if desired.  @var{processor_sets} is an array that is
created as a result of this call.  The caller may wish to
@code{vm_deallocate} this array when the data is no longer needed.
@var{processor_sets_count} is set to the number of processor sets in the
@var{processor_sets}.

This function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{host} is not a host.
@end deftypefun

@deftypefun kern_return_t host_processor_set_priv (@w{host_priv_t @var{host_priv}}, @w{processor_set_name_t @var{set_name}}, @w{processor_set_t *@var{set}})
The function @code{host_processor_set_priv} allows a privileged
application to obtain the control port @var{set} for an existing
processor set from its name port @var{set_name}.  The privileged host
port @var{host_priv} is required.

This function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{host_priv} is not a valid host
control port.
@end deftypefun

@deftypefun kern_return_t processor_set_default (@w{host_t @var{host}}, @w{processor_set_name_t *@var{default_set}})
The function @code{processor_set_default} returns the default processor
set of @var{host} in @var{default_set}.  The default processor set is
used by all threads, tasks, and processors that are not explicitly
assigned to other sets.  processor_set_default returns a port that can
be used to obtain information about this set (e.g. how many threads are
assigned to it).  This port cannot be used to perform operations on that
set.

This function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{host} is not a host and
@code{KERN_INVALID_ADDRESS} if @var{default_set} points to
inaccessible memory.
@end deftypefun


@node Processor Set Creation
@subsection Processor Set Creation

@deftypefun kern_return_t processor_set_create (@w{host_t @var{host}}, @w{processor_set_t *@var{new_set}}, @w{processor_set_name_t *@var{new_name}})
The function @code{processor_set_create} creates a new processor set on
@var{host} and returns the two ports associated with it.  The port
returned in @var{new_set} is the actual port representing the set.  It
is used to perform operations such as assigning processors, tasks, or
threads.  The port returned in @var{new_name} identifies the set, and is
used to obtain information about the set.

This function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{host} is not a host,
@code{KERN_INVALID_ADDRESS} if @var{new_set} or @var{new_name} points to
inaccessible memory and @code{KERN_FAILURE} is the operating system does
not support processor allocation.
@end deftypefun


@node Processor Set Destruction
@subsection Processor Set Destruction

@deftypefun kern_return_t processor_set_destroy (@w{processor_set_t @var{processor_set}})
The function @code{processor_set_destroy} destroys the specified
processor set.  Any assigned processors, tasks, or threads are
reassigned to the default set.  The object port for the processor set is
required (not the name port).  The default processor set cannot be
destroyed.

This function returns @code{KERN_SUCCESS} if the set was destroyed,
@code{KERN_FAILURE} if an attempt was made to destroy the default
processor set, or the operating system does not support processor
allocation, and @code{KERN_INVALID_ARGUMENT} if @var{processor_set} is
not a valid processor set control port.
@end deftypefun


@node Tasks and Threads on Sets
@subsection Tasks and Threads on Sets

@deftypefun kern_return_t processor_set_tasks (@w{processor_set_t @var{processor_set}}, @w{task_array_t *@var{task_list}}, @w{mach_msg_type_number_t *@var{task_count}})
The function @code{processor_set_tasks} gets send rights to the kernel
port for each task currently assigned to @var{processor_set}.

@var{task_list} is an array that is created as a result of this call.
The caller may wish to @code{vm_deallocate} this array when the data is
no longer needed.  @var{task_count} is set to the number of tasks in the
@var{task_list}.

This function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{processor_set} is not a processor
set.
@end deftypefun

@deftypefun kern_return_t processor_set_threads (@w{processor_set_t @var{processor_set}}, @w{thread_array_t *@var{thread_list}}, @w{mach_msg_type_number_t *@var{thread_count}})
The function @code{processor_set_thread} gets send rights to the kernel
port for each thread currently assigned to @var{processor_set}.

@var{thread_list} is an array that is created as a result of this call.
The caller may wish to @code{vm_deallocate} this array when the data is
no longer needed.  @var{thread_count} is set to the number of threads in
the @var{thread_list}.

This function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{processor_set} is not a processor
set.
@end deftypefun

@deftypefun kern_return_t task_assign (@w{task_t @var{task}}, @w{processor_set_t @var{processor_set}}, @w{boolean_t @var{assign_threads}})
The function @code{task_assign} assigns @var{task} the set
@var{processor_set}.  This assignment is for the purposes of determining
the initial assignment of newly created threads in task.  Any previous
assignment of the task is nullified.  Existing threads within the task
are also reassigned if @var{assign_threads} is @code{TRUE}.  They are
not affected if it is @code{FALSE}.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed and @code{KERN_INVALID_ARGUMENT} if @var{task} is not a task,
or @var{processor_set} is not a processor set on the same host as
@var{task}.
@end deftypefun

@deftypefun kern_return_t task_assign_default (@w{task_t @var{task}}, @w{boolean_t @var{assign_threads}})
The function @code{task_assign_default} is a variant of
@code{task_assign} that assigns the task to the default processor set on
that task's host.  This variant exists because the control port for the
default processor set is privileged and not usually available to users.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed and @code{KERN_INVALID_ARGUMENT} if @var{task} is not a task.
@end deftypefun

@deftypefun kern_return_t task_get_assignment (@w{task_t @var{task}}, @w{processor_set_name_t *@var{assigned_set}})
The function @code{task_get_assignment} returns the name of the
processor set to which the thread is currently assigned in
@var{assigned_set}.  This port can only be used to obtain information
about the processor set.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed, @code{KERN_INVALID_ADDRESS} if @var{processor_set} points to
inaccessible memory, and @code{KERN_INVALID_ARGUMENT} if @var{task} is
not a task.
@end deftypefun

@deftypefun kern_return_t thread_assign (@w{thread_t @var{thread}}, @w{processor_set_t @var{processor_set}})
The function @code{thread_assign} assigns @var{thread} the set
@var{processor_set}.  After the assignment is completed, the thread only
executes on processors assigned to the designated processor set.  If
there are no such processors, then the thread is unable to execute.  Any
previous assignment of the thread is nullified.  Unix system call
compatibility code may temporarily force threads to execute on the
master processor.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed and @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a
thread, or @var{processor_set} is not a processor set on the same host
as @var{thread}.
@end deftypefun

@deftypefun kern_return_t thread_assign_default (@w{thread_t @var{thread}})
The function @code{thread_assign_default} is a variant of
@code{thread_assign} that assigns the thread to the default processor
set on that thread's host.  This variant exists because the control port
for the default processor set is privileged and not usually available
to users.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed and @code{KERN_INVALID_ARGUMENT} if @var{thread} is not a
thread.
@end deftypefun

@deftypefun kern_return_t thread_get_assignment (@w{thread_t @var{thread}}, @w{processor_set_name_t *@var{assigned_set}})
The function @code{thread_get_assignment} returns the name of the
processor set to which the thread is currently assigned in
@var{assigned_set}.  This port can only be used to obtain information
about the processor set.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed, @code{KERN_INVALID_ADDRESS} if @var{processor_set} points to
inaccessible memory, and @code{KERN_INVALID_ARGUMENT} if @var{thread} is
not a thread.
@end deftypefun


@node Processor Set Priority
@subsection Processor Set Priority

@deftypefun kern_return_t processor_set_max_priority (@w{processor_set_t @var{processor_set}}, @w{int @var{max_priority}}, @w{boolean_t @var{change_threads}})
The function @code{processor_set_max_priority} is used to set the
maximum priority for a processor set.  The priority of a processor set
is used only for newly created threads (thread's maximum priority is set
to processor set's) and the assignment of threads to the set (thread's
maximum priority is reduced if it exceeds the set's maximum priority,
thread's priority is similarly reduced).
@code{processor_set_max_priority} changes this priority.  It also sets
the maximum priority of all threads assigned to the processor set to
this new priority if @var{change_threads} is @code{TRUE}.  If this
maximum priority is less than the priorities of any of these threads,
their priorities will also be set to this new value.

This function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{processor_set} is not a processor
set or @var{priority} is not a valid priority.
@end deftypefun


@node Processor Set Policy
@subsection Processor Set Policy

@deftypefun kern_return_t processor_set_policy_enable (@w{processor_set_t @var{processor_set}}, @w{int @var{policy}})
@deftypefunx kern_return_t processor_set_policy_disable (@w{processor_set_t @var{processor_set}}, @w{int @var{policy}}, @w{boolean_t @var{change_threads}})
Processor sets may restrict the scheduling policies to be used for
threads assigned to them.  These two calls provide the mechanism for
designating permitted and forbidden policies.  The current set of
permitted policies can be obtained from @code{processor_set_info}.
Timesharing may not be forbidden by any processor set.  This is a
compromise to reduce the complexity of the assign operation; any thread
whose policy is forbidden by the target processor set has its policy
reset to timesharing.  If the @var{change_threads} argument to
@code{processor_set_policy_disable} is true, threads currently assigned
to this processor set and using the newly disabled policy will have
their policy reset to timesharing.

@file{mach/policy.h} contains the allowed policies; it is included by
@file{mach.h}.  Not all policies (e.g. fixed priority) are supported by
all systems.

This function returns @code{KERN_SUCCESS} if the operation was completed
successfully and @code{KERN_INVALID_ARGUMENT} if @var{processor_set} is
not a processor set or @var{policy} is not a valid policy, or an attempt
was made to disable timesharing.
@end deftypefun


@node Processor Set Info
@subsection Processor Set Info

@deftypefun kern_return_t processor_set_info (@w{processor_set_name_t @var{set_name}}, @w{int @var{flavor}}, @w{host_t *@var{host}}, @w{processor_set_info_t @var{processor_set_info}}, @w{mach_msg_type_number_t *@var{processor_set_info_count}})
The function @code{processor_set_info} returns the selected information array
for a processor set, as specified by @var{flavor}.

@var{host} is set to the host on which the processor set resides.  This
is the non-privileged host port.

@var{processor_set_info} is an array of integers that is supplied by the
caller and returned filled with specified information.
@var{processor_set_info_count} is supplied as the maximum number of
integers in @var{processor_set_info}.  On return, it contains the actual
number of integers in @var{processor_set_info}.  The maximum number of
integers returned by any flavor is @code{PROCESSOR_SET_INFO_MAX}.

The type of information returned is defined by @var{flavor}, which can
be one of the following:

@table @code
@item PROCESSOR_SET_BASIC_INFO
The function returns basic information about the processor set, as
defined by @code{processor_set_basic_info_t}.  This includes the number
of tasks and threads assigned to the processor set.  The number of
integers returned is @code{PROCESSOR_SET_BASIC_INFO_COUNT}.

@item PROCESSOR_SET_SCHED_INFO
The function returns information about the scheduling policy for the
processor set as defined by @code{processor_set_sched_info_t}.  The
number of integers returned is @code{PROCESSOR_SET_SCHED_INFO_COUNT}.
@end table

Some machines may define additional (machine-dependent) flavors.

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{processor_set} is not a processor
set or @var{flavor} is not recognized.  The function returns
@code{MIG_ARRAY_TOO_LARGE} if the returned info array is too large for
@var{processor_set_info}.  In this case, @var{processor_set_info} is
filled as much as possible and @var{processor_set_info_count} is set to the
number of elements that would have been returned if there were enough
room.
@end deftypefun

@deftp {Data type} {struct processor_set_basic_info}
This structure is returned in @var{processor_set_info} by the
@code{processor_set_info} function and provides basic information about
the processor set.  You can cast a variable of type
@code{processor_set_info_t} to a pointer of this type if you provided it
as the @var{processor_set_info} parameter for the
@code{PROCESSOR_SET_BASIC_INFO} flavor of @code{processor_set_info}.  It
has the following members:

@table @code
@item int processor_count
number of processors

@item int task_count
number of tasks

@item int thread_count
number of threads

@item int load_average
scaled load average

@item int mach_factor
scaled mach factor
@end table
@end deftp

@deftp {Data type} processor_set_basic_info_t
This is a pointer to a @code{struct processor_set_basic_info}.
@end deftp

@deftp {Data type} {struct processor_set_sched_info}
This structure is returned in @var{processor_set_info} by the
@code{processor_set_info} function and provides schedule information
about the processor set.  You can cast a variable of type
@code{processor_set_info_t} to a pointer of this type if you provided it
as the @var{processor_set_info} parameter for the
@code{PROCESSOR_SET_SCHED_INFO} flavor of @code{processor_set_info}.  It
has the following members:

@table @code
@item int policies
allowed policies

@item int max_priority
max priority for new threads
@end table
@end deftp

@deftp {Data type} processor_set_sched_info_t
This is a pointer to a @code{struct processor_set_sched_info}.
@end deftp


@node Processor Interface
@section Processor Interface

@cindex processor port
@cindex port representing a processor
@deftp {Data type} processor_t 
This is a @code{mach_port_t} and used to hold the port name of a
processor port that represents the processor.  Operations on the
processor are implemented as remote procedure calls to the processor
port.
@end deftp

@menu
* Hosted Processors::             Getting a list of all processors on a host.
* Processor Control::             Starting, stopping, controlling processors.
* Processors and Sets::           Combining processors into processor sets.
* Processor Info::                Obtaining information on processors.
@end menu


@node Hosted Processors
@subsection Hosted Processors

@deftypefun kern_return_t host_processors (@w{host_priv_t @var{host_priv}}, @w{processor_array_t *@var{processor_list}}, @w{mach_msg_type_number_t *@var{processor_count}})
The function @code{host_processors} gets send rights to the processor
port for each processor existing on @var{host_priv}.  This is the
privileged port that allows its holder to control a processor.

@var{processor_list} is an array that is created as a result of this
call.  The caller may wish to @code{vm_deallocate} this array when the
data is no longer needed.  @var{processor_count} is set to the number of
processors in the @var{processor_list}.

This function returns @code{KERN_SUCCESS} if the call succeeded,
@code{KERN_INVALID_ARGUMENT} if @var{host_priv} is not a privileged host
port, and @code{KERN_INVALID_ADDRESS} if @var{processor_count} points to
inaccessible memory.
@end deftypefun


@node Processor Control
@subsection Processor Control

@deftypefun kern_return_t processor_start (@w{processor_t @var{processor}})
@deftypefunx kern_return_t processor_exit (@w{processor_t @var{processor}})
@deftypefunx kern_return_t processor_control (@w{processor_t @var{processor}}, @w{processor_info_t *@var{cmd}}, @w{mach_msg_type_number_t @var{count}})
Some multiprocessors may allow privileged software to control
processors.  The @code{processor_start}, @code{processor_exit}, and
@code{processor_control} operations implement this.  The interpretation
of the command in @var{cmd} is machine dependent.  A newly started
processor is assigned to the default processor set.  An exited processor
is removed from the processor set to which it was assigned and ceases to
be active.

@var{count} contains the length of the command @var{cmd} as a number of
ints.

Availability limited.  All of these operations are machine-dependent.
They may do nothing.  The ability to restart an exited processor is also
machine-dependent.

This function returns @code{KERN_SUCCESS} if the operation was
performed, @code{KERN_FAILURE} if the operation was not performed (a
likely reason is that it is not supported on this processor),
@code{KERN_INVALID_ARGUMENT} if @var{processor} is not a processor, and
@code{KERN_INVALID_ADDRESS} if @var{cmd} points to inaccessible memory.
@end deftypefun

@node Processors and Sets
@subsection Processors and Sets

@deftypefun kern_return_t processor_assign (@w{processor_t @var{processor}}, @w{processor_set_t @var{processor_set}}, @w{boolean_t @var{wait}})
The function @code{processor_assign} assigns @var{processor} to the
set @var{processor_set}.  After the assignment is completed, the
processor only executes threads that are assigned to that processor set.
Any previous assignment of the processor is nullified.  The master
processor cannot be reassigned.  All processors take clock interrupts at
all times.  The @var{wait} argument indicates whether the caller should
wait for the assignment to be completed or should return immediately.
Dedicated kernel threads are used to perform processor assignment, so
setting wait to @code{FALSE} allows assignment requests to be queued and
performed faster, especially if the kernel has more than one dedicated
internal thread for processor assignment.  Redirection of other device
interrupts away from processors assigned to other than the default
processor set is machine-dependent.  Intermediaries that interpose on
ports must be sure to interpose on both ports involved in this call if
they interpose on either.

This function returns @code{KERN_SUCCESS} if the assignment has been
performed, @code{KERN_INVALID_ARGUMENT} if @var{processor} is not a
processor, or @var{processor_set} is not a processor set on the same
host as @var{processor}.
@end deftypefun

@deftypefun kern_return_t processor_get_assignment (@w{processor_t @var{processor}}, @w{processor_set_name_t *@var{assigned_set}})
The function @code{processor_get_assignment} obtains the current
assignment of a processor.  The name port of the processor set is
returned in @var{assigned_set}.
@end deftypefun

@node Processor Info
@subsection Processor Info

@deftypefun kern_return_t processor_info (@w{processor_t @var{processor}}, @w{int @var{flavor}}, @w{host_t *@var{host}}, @w{processor_info_t @var{processor_info}}, @w{mach_msg_type_number_t *@var{processor_info_count}})
The function @code{processor_info} returns the selected information array
for a processor, as specified by @var{flavor}.

@var{host} is set to the host on which the processor set resides.  This
is the non-privileged host port.

@var{processor_info} is an array of integers that is supplied by the
caller and returned filled with specified information.
@var{processor_info_count} is supplied as the maximum number of integers in
@var{processor_info}.  On return, it contains the actual number of
integers in @var{processor_info}.  The maximum number of integers
returned by any flavor is @code{PROCESSOR_INFO_MAX}.

The type of information returned is defined by @var{flavor}, which can
be one of the following:

@table @code
@item PROCESSOR_BASIC_INFO
The function returns basic information about the processor, as defined
by @code{processor_basic_info_t}.  This includes the slot number of the
processor.  The number of integers returned is
@code{PROCESSOR_BASIC_INFO_COUNT}.
@end table

Machines which require more configuration information beyond the slot
number are expected to define additional (machine-dependent) flavors.

The function returns @code{KERN_SUCCESS} if the call succeeded and
@code{KERN_INVALID_ARGUMENT} if @var{processor} is not a processor or
@var{flavor} is not recognized.  The function returns
@code{MIG_ARRAY_TOO_LARGE} if the returned info array is too large for
@var{processor_info}.  In this case, @var{processor_info} is filled as
much as possible and @var{processor_infoCnt} is set to the number of
elements that would have been returned if there were enough room.
@end deftypefun

@deftp {Data type} {struct processor_basic_info}
This structure is returned in @var{processor_info} by the
@code{processor_info} function and provides basic information about the
processor.  You can cast a variable of type @code{processor_info_t} to a
pointer of this type if you provided it as the @var{processor_info}
parameter for the @code{PROCESSOR_BASIC_INFO} flavor of
@code{processor_info}.  It has the following members:

@table @code
@item cpu_type_t cpu_type
cpu type

@item cpu_subtype_t cpu_subtype
cpu subtype

@item boolean_t running
is processor running?

@item int slot_num
slot number

@item boolean_t is_master
is this the master processor
@end table
@end deftp

@deftp {Data type} processor_basic_info_t
This is a pointer to a @code{struct processor_basic_info}.
@end deftp


@node Device Interface
@chapter Device Interface

The GNU Mach microkernel provides a simple device interface that allows
the user space programs to access the underlying hardware devices.  Each
device has a unique name, which is a string up to 127 characters long.
To open a device, the device master port has to be supplied.  The device
master port is only available through the bootstrap port.  Anyone who
has control over the device master port can use all hardware devices.
@c XXX FIXME bootstrap port, bootstrap

@cindex device port
@cindex port representing a device
@deftp {Data type} device_t 
This is a @code{mach_port_t} and used to hold the port name of a
device port that represents the device.  Operations on the device are
implemented as remote procedure calls to the device port.  Each device
provides a sequence of records.  The length of a record is specific to
the device.  Data can be transferred ``out-of-line'' or ``in-line''
(@pxref{Memory}).
@end deftp

All constants and functions in this chapter are defined in
@file{device/device.h}.

@menu
* Device Reply Server::           Handling device reply messages.
* Device Open::                   Opening hardware devices.
* Device Close::                  Closing hardware devices.
* Device Read::                   Reading data from the device.
* Device Write::                  Writing data to the device.
* Device Map::                    Mapping devices into virtual memory.
* Device Status::                 Querying and manipulating a device.
* Device Filter::                 Filtering packets arriving on a device.
@end menu


@node Device Reply Server
@section Device Reply Server

Beside the usual synchronous interface, an asynchronous interface is
provided.  For this, the caller has to receive and handle the reply
messages separately from the function call.

@deftypefun boolean_t device_reply_server (@w{msg_header_t *@var{in_msg}}, @w{msg_header_t *@var{out_msg}})
The function @code{device_reply_server} is produced by the
remote procedure call generator to handle a received message.  This
function does all necessary argument handling, and actually calls one of
the following functions: @code{ds_device_open_reply},
@code{ds_device_read_reply}, @code{ds_device_read_reply_inband},
@code{ds_device_write_reply} and @code{ds_device_write_reply_inband}.

The @var{in_msg} argument is the message that has been received from the
kernel.  The @var{out_msg} is a reply message, but this is not used for
this server.

The function returns @code{TRUE} to indicate that the message in
question was applicable to this interface, and that the appropriate
routine was called to interpret the message.  It returns @code{FALSE} to
indicate that the message did not apply to this interface, and that no
other action was taken.
@end deftypefun


@node Device Open
@section Device Open

@deftypefun kern_return_t device_open (@w{mach_port_t @var{master_port}}, @w{dev_mode_t @var{mode}}, @w{dev_name_t @var{name}}, @w{device_t *@var{device}})
The function @code{device_open} opens the device @var{name} and returns
a port to it in @var{device}.  The open count for the device is
incremented by one.  If the open count was 0, the open handler for the
device is invoked.

@var{master_port} must hold the master device port.  @var{name}
specifies the device to open, and is a string up to 128 characters long.
@var{mode} is the open mode.  It is a bitwise-or of the following
constants:

@table @code
@item D_READ
Request read access for the device.

@item D_WRITE
Request write access for the device.

@item D_NODELAY
Do not delay an open.
@c XXX Is this really used at all?  Maybe for tape drives?  What does it mean?
@end table

The function returns @code{D_SUCCESS} if the device was successfully
opened, @code{D_INVALID_OPERATION} if @var{master_port} is not the
master device port, @code{D_WOULD_BLOCK} is the device is busy and
@code{D_NOWAIT} was specified in mode, @code{D_ALREADY_OPEN} if the
device is already open in an incompatible mode and
@code{D_NO_SUCH_DEVICE} if @var{name} does not denote a know device.
@end deftypefun

@deftypefun kern_return_t device_open_request (@w{mach_port_t @var{master_port}}, @w{mach_port_t @var{reply_port}}, @w{dev_mode_t @var{mode}}, @w{dev_name_t @var{name}})
@deftypefunx kern_return_t ds_device_open_reply (@w{mach_port_t @var{reply_port}}, @w{kern_return_t @var{return}}, @w{device_t *@var{device}})
This is the asynchronous form of the @code{device_open} function.
@code{device_open_request} performs the open request.  The meaning for
the parameters is as in @code{device_open}.  Additionally, the caller
has to supply a reply port to which the @code{ds_device_open_reply}
message is sent by the kernel when the open has been performed.  The
return value of the open operation is stored in @var{return_code}.

As neither function receives a reply message, only message transmission
errors apply.  If no error occurs, @code{KERN_SUCCESS} is returned.
@end deftypefun


@node Device Close
@section Device Close

@deftypefun kern_return_t device_close (@w{device_t @var{device}})
The function @code{device_close} decrements the open count of the device
by one.  If the open count drops to zero, the close handler for the
device is called.  The device to close is specified by its port
@var{device}.

The function returns @code{D_SUCCESS} if the device was successfully
closed and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote a
device port.
@end deftypefun


@node Device Read
@section Device Read

@deftypefun kern_return_t device_read (@w{device_t @var{device}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{int @var{bytes_wanted}}, @w{io_buf_ptr_t *@var{data}}, @w{mach_msg_type_number_t *@var{data_count}})
The function @code{device_read} reads @var{bytes_wanted} bytes from
@var{device}, and stores them in a buffer allocated with
@code{vm_allocate}, which address is returned in @var{data}.  The caller
must deallocated it if it is no longer needed.  The number of bytes
actually returned is stored in @var{data_count}.

If @var{mode} is @code{D_NOWAIT}, the operation does not block.
Otherwise @var{mode} should be 0.  @var{recnum} is the record number to
be read, its meaning is device specific.

The function returns @code{D_SUCCESS} if some data was successfully
read, @code{D_WOULD_BLOCK} if no data is currently available and
@code{D_NOWAIT} is specified, and @code{D_NO_SUCH_DEVICE} if
@var{device} does not denote a device port.
@end deftypefun

@deftypefun kern_return_t device_read_inband (@w{device_t @var{device}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{int @var{bytes_wanted}}, @w{io_buf_ptr_inband_t *@var{data}}, @w{mach_msg_type_number_t *@var{data_count}})
The @code{device_read_inband} function works as the @code{device_read}
function, except that the data is returned ``in-line'' in the reply IPC
message (@pxref{Memory}).
@end deftypefun

@deftypefun kern_return_t device_read_request (@w{device_t @var{device}}, @w{mach_port_t @var{reply_port}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{int @var{bytes_wanted}})
@deftypefunx kern_return_t ds_device_read_reply (@w{mach_port_t @var{reply_port}}, @w{kern_return_t @var{return_code}}, @w{io_buf_ptr_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}})
This is the asynchronous form of the @code{device_read} function.
@code{device_read_request} performs the read request.  The meaning for
the parameters is as in @code{device_read}.  Additionally, the caller
has to supply a reply port to which the @code{ds_device_read_reply}
message is sent by the kernel when the read has been performed.  The
return value of the read operation is stored in @var{return_code}.

As neither function receives a reply message, only message transmission
errors apply.  If no error occurs, @code{KERN_SUCCESS} is returned.
@end deftypefun

@deftypefun kern_return_t device_read_request_inband (@w{device_t @var{device}}, @w{mach_port_t @var{reply_port}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{int @var{bytes_wanted}})
@deftypefunx kern_return_t ds_device_read_reply_inband (@w{mach_port_t @var{reply_port}}, @w{kern_return_t @var{return_code}}, @w{io_buf_ptr_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}})
The @code{device_read_request_inband} and
@code{ds_device_read_reply_inband} functions work as the
@code{device_read_request} and @code{ds_device_read_reply} functions,
except that the data is returned ``in-line'' in the reply IPC message
(@pxref{Memory}).
@end deftypefun


@node Device Write
@section Device Write

@deftypefun kern_return_t device_write (@w{device_t @var{device}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{io_buf_ptr_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}}, @w{int *@var{bytes_written}})
The function @code{device_write} writes @var{data_count} bytes from the
buffer @var{data} to @var{device}.  The number of bytes actually written
is returned in @var{bytes_written}.

If @var{mode} is @code{D_NOWAIT}, the function returns without waiting
for I/O completion.  Otherwise @var{mode} should be 0.  @var{recnum} is
the record number to be written, its meaning is device specific.

The function returns @code{D_SUCCESS} if some data was successfully
written and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote a
device port or the device is dead or not completely open.
@end deftypefun

@deftypefun kern_return_t device_write_inband (@w{device_t @var{device}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{int @var{bytes_wanted}}, @w{io_buf_ptr_inband_t *@var{data}}, @w{mach_msg_type_number_t *@var{data_count}})
The @code{device_write_inband} function works as the @code{device_write}
function, except that the data is sent ``in-line'' in the request IPC
message (@pxref{Memory}).
@end deftypefun

@deftypefun kern_return_t device_write_request (@w{device_t @var{device}}, @w{mach_port_t @var{reply_port}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{io_buf_ptr_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}})
@deftypefunx kern_return_t ds_device_write_reply (@w{mach_port_t @var{reply_port}}, @w{kern_return_t @var{return_code}}, @w{int @var{bytes_written}})
This is the asynchronous form of the @code{device_write} function.
@code{device_write_request} performs the write request.  The meaning for
the parameters is as in @code{device_write}.  Additionally, the caller
has to supply a reply port to which the @code{ds_device_write_reply}
message is sent by the kernel when the write has been performed.  The
return value of the write operation is stored in @var{return_code}.

As neither function receives a reply message, only message transmission
errors apply.  If no error occurs, @code{KERN_SUCCESS} is returned.
@end deftypefun

@deftypefun kern_return_t device_write_request_inband (@w{device_t @var{device}}, @w{mach_port_t @var{reply_port}}, @w{dev_mode_t @var{mode}}, @w{recnum_t @var{recnum}}, @w{io_buf_ptr_t @var{data}}, @w{mach_msg_type_number_t @var{data_count}})
@deftypefunx kern_return_t ds_device_write_reply_inband (@w{mach_port_t @var{reply_port}}, @w{kern_return_t @var{return_code}}, @w{int @var{bytes_written}})
The @code{device_write_request_inband} and
@code{ds_device_write_reply_inband} functions work as the
@code{device_write_request} and @code{ds_device_write_reply} functions,
except that the data is sent ``in-line'' in the request IPC message
(@pxref{Memory}).
@end deftypefun


@node Device Map
@section Device Map

@deftypefun kern_return_t device_map (@w{device_t @var{device}}, @w{vm_prot_t @var{prot}}, @w{vm_offset_t @var{offset}}, @w{vm_size_t @var{size}}, @w{mach_port_t *@var{pager}}, @w{int @var{unmap}})
The function @code{device_map} creates a new memory manager for
@var{device} and returns a port to it in @var{pager}.  The memory
manager is usable as a memory object in a @code{vm_map} call.  The call
is device dependent.

The protection for the memory object is specified by @var{prot}.  The
memory object starts at @var{offset} within the device and extends
@var{size} bytes.  @var{unmap} is currently unused.
@c XXX I suppose the caller should set it to 0.

The function returns @code{D_SUCCESS} if some data was successfully
written and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote a
device port or the device is dead or not completely open.
@end deftypefun


@node Device Status
@section Device Status

@deftypefun kern_return_t device_set_status (@w{device_t @var{device}}, @w{dev_flavor_t @var{flavor}}, @w{dev_status_t @var{status}}, @w{mach_msg_type_number_t @var{status_count}})
The function @code{device_set_status} sets the status of a device.  The
possible values for @var{flavor} and their interpretation is device
specific.

The function returns @code{D_SUCCESS} if some data was successfully
written and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote a
device port or the device is dead or not completely open.
@end deftypefun

@deftypefun kern_return_t device_get_status (@w{device_t @var{device}}, @w{dev_flavor_t @var{flavor}}, @w{dev_status_t @var{status}}, @w{mach_msg_type_number_t *@var{status_count}})
The function @code{device_get_status} gets the status of a device.  The
possible values for @var{flavor} and their interpretation is device
specific.

The function returns @code{D_SUCCESS} if some data was successfully
written and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote a
device port or the device is dead or not completely open.
@end deftypefun


@node Device Filter
@section Device Filter

@deftypefun kern_return_t device_set_filter (@w{device_t @var{device}}, @w{mach_port_t @var{receive_port}}, @w{mach_msg_type_name_t @var{receive_port_type}}, @w{int @var{priority}}, @w{filter_array_t @var{filter}}, @w{mach_msg_type_number_t @var{filter_count}})
The function @code{device_set_filter} makes it possible to filter out
selected data arriving at the device and forward it to a port.
@var{filter} is a list of filter commands, which are applied to incoming
data to determine if the data should be sent to @var{receive_port}.  The
IPC type of the send right is specified by @var{receive_port_right}, it
is either @code{MACH_MSG_TYPE_MAKE_SEND} or
@code{MACH_MSG_TYPE_MOVE_SEND}.  The @var{priority} value is used to
order multiple filters.

There can be up to @code{NET_MAX_FILTER} commands in @var{filter}.  The
actual number of commands is passed in @var{filter_count}.  For the
purpose of the filter test, an internal stack is provided.  After all
commands have been processed, the value on the top of the stack
determines if the data is forwarded or the next filter is tried.

@c XXX The following description was taken verbatim from the
@c kernel_interface.pdf document.
Each word of the command list specifies a data (push) operation (high
order NETF_NBPO bits) as well as a binary operator (low order NETF_NBPA
bits).  The value to be pushed onto the stack is chosen as follows.

@table @code
@item NETF_PUSHLIT
Use the next short word of the filter as the value.

@item NETF_PUSHZERO
Use 0 as the value.

@item NETF_PUSHWORD+N
Use short word N of the ``data'' portion of the message as the value.

@item NETF_PUSHHDR+N
Use short word N of the ``header'' portion of the message as the value.

@item NETF_PUSHIND+N
Pops the top long word from the stack and then uses short word N of the
``data'' portion of the message as the value.

@item NETF_PUSHHDRIND+N
Pops the top long word from the stack and then uses short word N of the
``header'' portion of the message as the value.

@item NETF_PUSHSTK+N
Use long word N of the stack (where the top of stack is long word 0) as
the value.

@item NETF_NOPUSH
Don't push a value.
@end table

The unsigned value so chosen is promoted to a long word before being
pushed.  Once a value is pushed (except for the case of
@code{NETF_NOPUSH}), the top two long words of the stack are popped and
a binary operator applied to them (with the old top of stack as the
second operand).  The result of the operator is pushed on the stack.
These operators are:

@table @code
@item NETF_NOP
Don't pop off any values and do no operation.

@item NETF_EQ
Perform an equal comparison.

@item NETF_LT
Perform a less than comparison.

@item NETF_LE
Perform a less than or equal comparison.

@item NETF_GT
Perform a greater than comparison.

@item NETF_GE
Perform a greater than or equal comparison.

@item NETF_AND
Perform a bitise boolean AND operation.

@item NETF_OR
Perform a bitise boolean inclusive OR operation.

@item NETF_XOR
Perform a bitise boolean exclusive OR operation.

@item NETF_NEQ
Perform a not equal comparison.

@item NETF_LSH
Perform a left shift operation.

@item NETF_RSH
Perform a right shift operation.

@item NETF_ADD
Perform an addition.

@item NETF_SUB
Perform a subtraction.

@item NETF_COR
Perform an equal comparison. If the comparison is @code{TRUE}, terminate
the filter list.  Otherwise, pop the result of the comparison off the
stack.

@item NETF_CAND
Perform an equal comparison. If the comparison is @code{FALSE},
terminate the filter list.  Otherwise, pop the result of the comparison
off the stack.

@item NETF_CNOR
Perform a not equal comparison. If the comparison is @code{FALSE},
terminate the filter list.  Otherwise, pop the result of the comparison
off the stack.

@item NETF_CNAND
Perform a not equal comparison. If the comparison is @code{TRUE},
terminate the filter list.  Otherwise, pop the result of the comparison
off the stack.  The scan of the filter list terminates when the filter
list is emptied, or a @code{NETF_C...} operation terminates the list. At
this time, if the final value of the top of the stack is @code{TRUE},
then the message is accepted for the filter.
@end table

The function returns @code{D_SUCCESS} if some data was successfully
written, @code{D_INVALID_OPERATION} if @var{receive_port} is not a valid
send right, and @code{D_NO_SUCH_DEVICE} if @var{device} does not denote
a device port or the device is dead or not completely open.
@end deftypefun


@node Kernel Debugger
@chapter Kernel Debugger

The GNU Mach kernel debugger @code{ddb} is a powerful built-in debugger
with a gdb like syntax.  It is enabled at compile time using the
@option{--enable-kdb} option.  Whenever you want to enter the debugger
while running the kernel, you can press the key combination
@key{Ctrl-Alt-D}.

@menu
* Operation::                     Basic architecture of the kernel debugger.
* Commands::                      Available commands in the kernel debugger.
* Variables::                     Access of variables from the kernel debugger.
* Expressions::                   Usage of expressions in the kernel debugger.
@end menu


@node Operation
@section Operation

The current location is called @dfn{dot}.  The dot is displayed with a
hexadecimal format at a prompt.  Examine and write commands update dot
to the address of the last line examined or the last location modified,
and set @dfn{next} to the address of the next location to be examined or
changed.  Other commands don't change dot, and set next to be the same
as dot.

The general command syntax is:

@example
@var{command}[/@var{modifier}] @var{address} [,@var{count}]
@end example

@kbd{!!} repeats the previous command, and a blank line repeats from the
address next with count 1 and no modifiers.  Specifying @var{address} sets
dot to the address.  Omitting @var{address} uses dot.  A missing @var{count}
is taken to be 1 for printing commands or infinity for stack traces.

Current @code{ddb} is enhanced to support multi-thread debugging.  A
break point can be set only for a specific thread, and the address space
or registers of non current thread can be examined or modified if
supported by machine dependent routines.  For example,

@example
break/t mach_msg_trap $task11.0
@end example

sets a break point at @code{mach_msg_trap} for the first thread of task
11 listed by a @code{show all threads} command.

In the above example, @code{$task11.0} is translated to the
corresponding thread structure's address by variable translation
mechanism described later.  If a default target thread is set in a
variable @code{$thread}, the @code{$task11.0} can be omitted.  In
general, if @code{t} is specified in a modifier of a command line, a
specified thread or a default target thread is used as a target thread
instead of the current one.  The @code{t} modifier in a command line is
not valid in evaluating expressions in a command line.  If you want to
get a value indirectly from a specific thread's address space or access
to its registers within an expression, you have to specify a default
target thread in advance, and to use @code{:t} modifier immediately
after the indirect access or the register reference like as follows:

@example
set $thread $task11.0
print $eax:t *(0x100):tuh
@end example

No sign extension and indirection @code{size(long, half word, byte)} can
be specified with @code{u}, @code{l}, @code{h} and @code{b} respectively
for the indirect access.

Note: Support of non current space/register access and user space break
point depend on the machines.  If not supported, attempts of such
operation may provide incorrect information or may cause strange
behavior.  Even if supported, the user space access is limited to the
pages resident in the main memory at that time.  If a target page is not
in the main memory, an error will be reported.

@code{ddb} has a feature like a command @code{more} for the output.  If
an output line exceeds the number set in the @code{$lines} variable, it
displays @samp{--db_more--} and waits for a response.  The valid
responses for it are:

@table @kbd
@item @key{SPC}
one more page

@item @key{RET}
one more line

@item q
abort the current command, and return to the command input mode
@end table


@node Commands
@section Commands

@table @code
@item examine(x) [/@var{modifier}] @var{addr}[,@var{count}] [ @var{thread} ]
Display the addressed locations according to the formats in the
modifier.  Multiple modifier formats display multiple locations.  If no
format is specified, the last formats specified for this command is
used.  Address space other than that of the current thread can be
specified with @code{t} option in the modifier and @var{thread}
parameter.  The format characters are

@table @code
@item b
look at by bytes(8 bits)

@item h
look at by half words(16 bits)

@item l
look at by long words(32 bits)

@item a
print the location being displayed

@item ,
skip one unit producing no output

@item A
print the location with a line number if possible

@item x
display in unsigned hex

@item z
display in signed hex

@item o
display in unsigned octal

@item d
display in signed decimal

@item u
display in unsigned decimal

@item r 
display in current radix, signed

@item c
display low 8 bits as a character.  Non-printing characters are
displayed as an octal escape code (e.g. '\000').

@item s
display the null-terminated string at the location.  Non-printing
characters are displayed as octal escapes.

@item m
display in unsigned hex with character dump at the end of each line.
The location is also displayed in hex at the beginning of each line.

@item i
display as an instruction

@item I
display as an instruction with possible alternate formats depending on
the machine:

@table @code
@item vax
don't assume that each external label is a procedure entry mask

@item i386
don't round to the next long word boundary

@item mips
print register contents
@end table
@end table

@item xf
Examine forward.  It executes an examine command with the last specified
parameters to it except that the next address displayed by it is used as
the start address.

@item xb
Examine backward.  It executes an examine command with the last
specified parameters to it except that the last start address subtracted
by the size displayed by it is used as the start address.

@item print[/axzodurc] @var{addr1} [ @var{addr2} @dots{} ]
Print @var{addr}'s according to the modifier character.  Valid formats
are: @code{a} @code{x} @code{z} @code{o} @code{d} @code{u} @code{r}
@code{c}.  If no modifier is specified, the last one specified to it is
used.  @var{addr} can be a string, and it is printed as it is.  For
example,

@example
print/x "eax = " $eax "\necx = " $ecx "\n"
@end example

will print like

@example
eax = xxxxxx
ecx = yyyyyy
@end example

@item write[/bhlt] @var{addr} [ @var{thread} ] @var{expr1} [ @var{expr2} @dots{} ]
Write the expressions at succeeding locations.  The write unit size can
be specified in the modifier with a letter b (byte), h (half word) or
l(long word) respectively.  If omitted, long word is assumed.  Target
address space can also be specified with @code{t} option in the modifier
and @var{thread} parameter.  Warning: since there is no delimiter
between expressions, strange things may happen.  It's best to enclose
each expression in parentheses.

@item set $@var{variable} [=] @var{expr}
Set the named variable or register with the value of @var{expr}.  Valid
variable names are described below.

@item break[/tuTU] @var{addr}[,@var{count}] [ @var{thread1} @dots{} ]
Set a break point at @var{addr}.  If count is supplied, continues
(@var{count}-1) times before stopping at the break point.  If the break
point is set, a break point number is printed with @samp{#}.  This
number can be used in deleting the break point or adding conditions to
it.

@table @code
@item t
Set a break point only for a specific thread.  The thread is specified
by @var{thread} parameter, or default one is used if the parameter is
omitted.

@item u
Set a break point in user space address.  It may be combined with
@code{t} or @code{T} option to specify the non-current target user
space.  Without @code{u} option, the address is considered in the kernel
space, and wrong space address is rejected with an error message.  This
option can be used only if it is supported by machine dependent
routines.

@item T
Set a break point only for threads in a specific task.  It is like
@code{t} option except that the break point is valid for all threads
which belong to the same task as the specified target thread.

@item U
Set a break point in shared user space address.  It is like @code{u}
option, except that the break point is valid for all threads which share
the same address space even if @code{t} option is specified.  @code{t}
option is used only to specify the target shared space.  Without
@code{t} option, @code{u} and @code{U} have the same meanings.  @code{U}
is useful for setting a user space break point in non-current address
space with @code{t} option such as in an emulation library space.  This
option can be used only if it is supported by machine dependent
routines.
@end table

Warning: if a user text is shadowed by a normal user space debugger,
user space break points may not work correctly.  Setting a break point
at the low-level code paths may also cause strange behavior.

@item delete[/tuTU] @var{addr}|#@var{number} [ @var{thread1} @dots{} ]
Delete the break point.  The target break point can be specified by a
break point number with @code{#}, or by @var{addr} like specified in
@code{break} command.

@item cond #@var{number} [ @var{condition} @var{commands} ]
Set or delete a condition for the break point specified by the
@var{number}.  If the @var{condition} and @var{commands} are null, the
condition is deleted.  Otherwise the condition is set for it.  When the
break point is hit, the @var{condition} is evaluated.  The
@var{commands} will be executed if the condition is true and the break
point count set by a break point command becomes zero.  @var{commands}
is a list of commands separated by semicolons.  Each command in the list
is executed in that order, but if a @code{continue} command is executed,
the command execution stops there, and the stopped thread resumes
execution.  If the command execution reaches the end of the list, and it
enters into a command input mode.  For example,

@example
set $work0 0
break/Tu xxx_start $task7.0
cond #1  (1)  set $work0 1; set $work1 0; cont
break/T  vm_fault $task7.0
cond #2  ($work0) set $work1 ($work1+1); cont
break/Tu xxx_end $task7.0
cond #3  ($work0) print $work1 " faults\n"; set $work0 0
cont
@end example

will print page fault counts from @code{xxx_start} to @code{xxx_end} in
@code{task7}.

@item step[/p] [,@var{count}]
Single step @var{count} times.  If @code{p} option is specified, print
each instruction at each step.  Otherwise, only print the last
instruction.

Warning: depending on machine type, it may not be possible to
single-step through some low-level code paths or user space code.  On
machines with software-emulated single-stepping (e.g., pmax), stepping
through code executed by interrupt handlers will probably do the wrong
thing.

@item continue[/c]
Continue execution until a breakpoint or watchpoint.  If @code{/c},
count instructions while executing.  Some machines (e.g., pmax) also
count loads and stores.

Warning: when counting, the debugger is really silently single-stepping.
This means that single-stepping on low-level code may cause strange
behavior.

@item until
Stop at the next call or return instruction.

@item next[/p]
Stop at the matching return instruction.  If @code{p} option is
specified, print the call nesting depth and the cumulative instruction
count at each call or return.  Otherwise, only print when the matching
return is hit.

@item match[/p]
A synonym for @code{next}.

@item trace[/tu] [ @var{frame_addr}|@var{thread} ][,@var{count}]
Stack trace.  @code{u} option traces user space; if omitted, only traces
kernel space.  If @code{t} option is specified, it shows the stack trace
of the specified thread or a default target thread.  Otherwise, it shows
the stack trace of the current thread from the frame address specified
by a parameter or from the current frame.  @var{count} is the number of
frames to be traced.  If the @var{count} is omitted, all frames are
printed.

Warning: If the target thread's stack is not in the main memory at that
time, the stack trace will fail.  User space stack trace is valid only
if the machine dependent code supports it.

@item search[/bhl] @var{addr} @var{value} [@var{mask}] [,@var{count}]
Search memory for a value.  This command might fail in interesting ways
if it doesn't find the searched-for value.  This is because @code{ddb}
doesn't always recover from touching bad memory.  The optional count
argument limits the search.

@item macro @var{name} @var{commands}
Define a debugger macro as @var{name}.  @var{commands} is a list of
commands to be associated with the macro.  In the expressions of the
command list, a variable @code{$argxx} can be used to get a parameter
passed to the macro.  When a macro is called, each argument is evaluated
as an expression, and the value is assigned to each parameter,
@code{$arg1}, @code{$arg2}, @dots{} respectively.  10 @code{$arg}
variables are reserved to each level of macros, and they can be used as
local variables.  The nesting of macro can be allowed up to 5 levels.
For example,

@example
macro xinit set $work0 $arg1
macro xlist examine/m $work0,4; set $work0 *($work0)
xinit *(xxx_list)
xlist
@enddots{}
@end example

will print the contents of a list starting from @code{xxx_list} by each
@code{xlist} command.

@item dmacro @var{name}
Delete the macro named @var{name}.

@item show all threads[/ul]
Display all tasks and threads information.  This version of @code{ddb}
prints more information than previous one.  It shows UNIX process
information like @command{ps} for each task.  The UNIX process
information may not be shown if it is not supported in the machine, or
the bottom of the stack of the target task is not in the main memory at
that time.  It also shows task and thread identification numbers.  These
numbers can be used to specify a task or a thread symbolically in
various commands.  The numbers are valid only in the same debugger
session.  If the execution is resumed again, the numbers may change.
The current thread can be distinguished from others by a @code{#} after
the thread id instead of @code{:}.  Without @code{l} option, it only
shows thread id, thread structure address and the status for each
thread.  The status consists of 5 letters, R(run), W(wait), S(suspended),
O(swapped out) and N(interruptible), and if corresponding
status bit is off, @code{.}  is printed instead.  If @code{l} option is
specified, more detail information is printed for each thread.

@item show task [ @var{addr} ]
Display the information of a task specified by @var{addr}.  If
@var{addr} is omitted, current task information is displayed.

@item show thread [ @var{addr} ]
Display the information of a thread specified by @var{addr}.  If
@var{addr} is omitted, current thread information is displayed.

@item show registers[/tu [ @var{thread} ]]
Display the register set.  Target thread can be specified with @code{t}
option and @var{thread} parameter.  If @code{u} option is specified, it
displays user registers instead of kernel or currently saved one.

Warning: The support of @code{t} and @code{u} option depends on the
machine.  If not supported, incorrect information will be displayed.

@item show map @var{addr}
Prints the @code{vm_map} at @var{addr}.

@item show object @var{addr}
Prints the @code{vm_object} at @var{addr}.

@item show page @var{addr}
Prints the @code{vm_page} structure at @var{addr}.

@item show port @var{addr}
Prints the @code{ipc_port} structure at @var{addr}.

@item show ipc_port[/t [ @var{thread} ]]
Prints all @code{ipc_port} structure's addresses the target thread has.
The target thread is a current thread or that specified by a parameter.

@item show macro [ @var{name} ]
Show the definitions of macros.  If @var{name} is specified, only the
definition of it is displayed.  Otherwise, definitions of all macros are
displayed.

@item show watches
Displays all watchpoints.

@item watch[/T] @var{addr},@var{size} [ @var{task} ]
Set a watchpoint for a region.  Execution stops when an attempt to
modify the region occurs.  The @var{size} argument defaults to 4.
Without @code{T} option, @var{addr} is assumed to be a kernel address.
If you want to set a watch point in user space, specify @code{T} and
@var{task} parameter where the address belongs to.  If the @var{task}
parameter is omitted, a task of the default target thread or a current
task is assumed.  If you specify a wrong space address, the request is
rejected with an error message.

Warning: Attempts to watch wired kernel memory may cause unrecoverable
error in some systems such as i386.  Watchpoints on user addresses work
best.
@end table


@node Variables
@section Variables

The debugger accesses registers and variables as $@var{name}.  Register
names are as in the @code{show registers} command.  Some variables are
suffixed with numbers, and may have some modifier following a colon
immediately after the variable name.  For example, register variables
can have @code{u} and @code{t} modifier to indicate user register and
that of a default target thread instead of that of the current thread
(e.g. @code{$eax:tu}).

Built-in variables currently supported are:

@table @code
@item task@var{xx}[.@var{yy}]
Task or thread structure address.  @var{xx} and @var{yy} are task and
thread identification numbers printed by a @code{show all threads}
command respectively.  This variable is read only.

@item thread
The default target thread.  The value is used when @code{t} option is
specified without explicit thread structure address parameter in command
lines or expression evaluation.

@item radix
Input and output radix

@item maxoff
Addresses are printed as @var{symbol}+@var{offset} unless offset is greater than
maxoff.

@item maxwidth
The width of the displayed line.

@item lines
The number of lines.  It is used by @code{more} feature.

@item tabstops
Tab stop width.

@item arg@var{xx}
Parameters passed to a macro.  @var{xx} can be 1 to 10.

@item work@var{xx}
Work variable.  @var{xx} can be 0 to 31.
@end table


@node Expressions
@section Expressions

Almost all expression operators in C are supported except @code{~},
@code{^}, and unary @code{&}.  Special rules in @code{ddb} are:

@table @code
@item @var{identifier}
name of a symbol.  It is translated to the address(or value) of it.
@code{.}  and @code{:} can be used in the identifier.  If supported by
an object format dependent routine,
[@var{file_name}:]@var{func}[:@var{line_number}]
[@var{file_name}:]@var{variable}, and
@var{file_name}[:@var{line_number}] can be accepted as a symbol.  The
symbol may be prefixed with @code{@var{symbol_table_name}::} like
@code{emulator::mach_msg_trap} to specify other than kernel symbols.

@item @var{number}
radix is determined by the first two letters:
@table @code
@item 0x
hex
@item 0o
octal
@item 0t
decimal
@end table

otherwise, follow current radix.

@item .
dot

@item +
next

@item ..
address of the start of the last line examined.  Unlike dot or next,
this is only changed by @code{examine} or @code{write} command.

@item ´
last address explicitly specified.

@item $@var{variable}
register name or variable.  It is translated to the value of it.  It may
be followed by a @code{:} and modifiers as described above.

@item a
multiple of right hand side.

@item *@var{expr}
indirection.  It may be followed by a @code{:} and modifiers as
described above.
@end table

@include gpl.texi

@node Documentation License
@appendix Documentation License

This manual is copyrighted and licensed under the GNU Free Documentation
license.

Parts of this manual are derived from the Mach manual packages
originally provided by Carnegie Mellon University.

@menu
* GNU Free Documentation License:: The GNU Free Documentation License.
* CMU License::                   The CMU license applies to the original Mach
                                  kernel and its documentation.
@end menu

@include fdl.texi

@node CMU License
@appendixsec CMU License

@quotation
@display
Mach Operating System
Copyright @copyright{} 1991,1990,1989 Carnegie Mellon University
All Rights Reserved.
@end display

Permission to use, copy, modify and distribute this software and its
documentation is hereby granted, provided that both the copyright
notice and this permission notice appear in all copies of the
software, derivative works or modified versions, and any portions
thereof, and that both notices appear in supporting documentation.

@sc{carnegie mellon allows free use of this software in its ``as is''
condition.  carnegie mellon disclaims any liability of any kind for
any damages whatsoever resulting from the use of this software.}

Carnegie Mellon requests users of this software to return to

@display
 Software Distribution Coordinator
 School of Computer Science
 Carnegie Mellon University
 Pittsburgh PA 15213-3890
@end display

@noindent
or @email{Software.Distribution@@CS.CMU.EDU} any improvements or
extensions that they make and grant Carnegie Mellon the rights to
redistribute these changes.
@end quotation

@node Concept Index
@unnumbered Concept Index

@printindex cp


@node Function and Data Index
@unnumbered Function and Data Index

@printindex fn


@summarycontents
@contents
@bye