The Linux Kernel

This book is for Linux enthusiasts who want to know how the Linux kernel works. It is
not an internals manual. Rather it describes the principles and mechanisms that Linux
uses; how and why the Linux kernel works the way that it does. Linux is a moving target;
this book is based upon the current, stable, 2.0.33 sources as those are what most
individuals and companies are now using.
This book is freely distributable, you may copy and redistribute it under certain conditions.
Please refer to the copyright and distribution statement.
For Gill, Esther and Stephen
Legal Notice
UNIX is a trademark of Univel.
Linux is a trademark of Linus Torvalds, and has no connection to UNIXTM or
Univel.
Copyright
c 1996,1997,1998,1999 David A Rusling
3 Foxglove Close, Wokingham, Berkshire RG41 3NF, UK
david.rusling@arm.com
This book (\The Linux Kernel") may be reproduced and distributed in whole or
in part, without fee, subject to the following conditions:
 The copyright notice above and this permission notice must be preserved
complete on all complete or partial copies.
 Any translation or derived work must be approved by the author in writing
before distribution.
 If you distribute this work in part, instructions for obtaining the complete
version of this manual must be included, and a means for obtaining a complete
version provided.
 Small portions may be reproduced as illustrations for reviews or quotes in
other works without this permission notice if proper citation is given.
Exceptions to these rules may be granted for academic purposes: Write to the
author and ask. These restrictions are here to protect us as authors, not to restrict
you as learners and educators.
All source code in this document is placed under the GNU General Public License,
available via anonymous FTP from prep.ai.mit.edu:/pub/gnu/COPYING. It is also
reproduced in appendix D.
Preface
Linux is a phenomenon of the Internet. Born out of the hobby project of a student it
has grown to become more popular than any other freely available operating system.
To many Linux is an enigma. How can something that is free be worthwhile? In
a world dominated by a handful of large software corporations, how can something
that has been written by a bunch of \hackers" (sic) hope to compete? How can
software contributed to by many dierent people in many dierent countries around
the world have a hope of being stable and eective? Yet stable and eective it is
and compete it does. Many Universities and research establishments use it for their
everyday computing needs. People are running it on their home PCs and I would
wager that most companies are using it somewhere even if they do not always realize
that they do. Linux is used to browse the web, host web sites, write theses, send
electronic mail and, as always with computers, to play games. Linux is emphatically
not a toy; it is a fully developed and professionally written operating system used by
enthusiasts all over the world.
The roots of Linux can be traced back to the origins of UnixTM . In 1969, Ken
Thompson of the Research Group at Bell Laboratories began experimenting on a
multi-user, multi-tasking operating system using an otherwise idle PDP-7. He was
soon joined by Dennis Richie and the two of them, along with other members of the
Research Group produced the early versions of UnixTM. Richie was strongly in
uenced
by an earlier project, MULTICS and the name UnixTM is itself a pun on the name
MULTICS. Early versions were written in assembly code, but the third version was
rewritten in a new programming language, C. C was designed and written by Richie
expressly as a programming language for writing operating systems. This rewrite
allowed UnixTM to move onto the more powerful PDP-11/45 and 11/70 computers
then being produced by DIGITAL. The rest, as they say, is history. UnixTM moved
out of the laboratory and into mainstream computing and soon most major computer
manufacturers were producing their own versions.
Linux was the solution to a simple need. The only software that Linus Torvalds,
Linux's author and principle maintainer was able to aord was Minix. Minix is a
simple, UnixTM like, operating system widely used as a teaching aid. Linus was less
than impressed with its features, his solution was to write his own software. He took
UnixTM as his model as that was an operating system that he was familiar with in his
day to day student life. He started with an Intel 386 based PC and started to write.
Progress was rapid and, excited by this, Linus oered his eorts to other students
via the emerging world wide computer networks, then mainly used by the academic
community. Others saw the software and started contributing. Much of this new
software was itself the solution to a problem that one of the contributors had. Before
long, Linux had become an operating system. It is important to note that Linux
iii
iv
contains no UnixTM code, it is a rewrite based on published POSIX standards. Linux
is built with and uses a lot of the GNU (GNU's Not UnixTM) software produced by
the Free Software Foundation in Cambridge, Massachusetts.
Most people use Linux as a simple tool, often just installing one of the many good
CD ROM-based distributions. A lot of Linux users use it to write applications or
to run applications written by others. Many Linux users read the HOWTOs1 avidly
and feel both the thrill of success when some part of the system has been correctly
congured and the frustration of failure when it has not. A minority are bold enough
to write device drivers and oer kernel patches to Linus Torvalds, the creator and
maintainer of the Linux kernel. Linus accepts additions and modications to the
kernel sources from anyone, anywhere. This might sound like a recipe for anarchy
but Linus exercises strict quality control and merges all new code into the kernel
himself. At any one time though, there are only a handful of people contributing
sources to the Linux kernel.
The majority of Linux users do not look at how the operating system works, how
it ts together. This is a shame because looking at Linux is a very good way to
learn more about how an operating system functions. Not only is it well written,
all the sources are freely available for you to look at. This is because although the
authors retain the copyrights to their software, they allow the sources to be freely
redistributable under the Free Software Foundation's GNU Public License. At rst
glance though, the sources can be confusing; you will see directories called kernel,
mm and net but what do they contain and how does that code work? What is needed
is a broader understanding of the overall structure and aims of Linux. This, in
short, is the aim of this book: to promote a clear understanding of how Linux, the
operating system, works. To provide a mind model that allows you to picture what
is happening within the system as you copy a le from one place to another or read
electronic mail. I well remember the excitement that I felt when I rst realized just
how an operating system actually worked. It is that excitement that I want to pass
on to the readers of this book.
My involvement with Linux started late in 1994 when I visited Jim Paradis who was
working on a port of Linux to the Alpha AXP processor based systems. I had worked
for Digital Equipment Co. Limited since 1984, mostly in networks and communications
and in 1992 I started working for the newly formed Digital Semiconductor
division. This division's goal was to enter fully into the merchant chip vendor market
and sell chips, and in particular the Alpha AXP range of microprocessors but also
Alpha AXP system boards outside of Digital. When I rst heard about Linux I
immediately saw an opportunity to have fun. Jim's enthusiasm was catching and I
started to help on the port. As I worked on this, I began more and more to appreciate
not only the operating system but also the community of engineers that produces it.
However, Alpha AXP is only one of the many hardware platforms that Linux runs
on. Most Linux kernels are running on Intel processor based systems but a growing
number of non-Intel Linux systems are becoming more commonly available. Amongst
these are Alpha AXP, ARM, MIPS, Sparc and PowerPC. I could have written this
book using any one of those platforms but my background and technical experiences
with Linux are with Linux on the Alpha AXP and, to a lesser extent on the ARM.
This is why this book sometimes uses non-Intel hardware as an example to illustrate
1A HOWTO is just what it sounds like, a document describing how to do something. Many have
been written for Linux and all are very useful.
v
some key point. It must be noted that around 95% of the Linux kernel sources are
common to all of the hardware platforms that it runs on. Likewise, around 95% of
this book is about the machine independent parts of the Linux kernel.
Reader Prole
This book does not make any assumptions about the knowledge or experience of
the reader. I believe that interest in the subject matter will encourage a process of
self education where neccessary. That said, a degree of familiarity with computers,
preferably the PC will help the reader derive real benet from the material, as will
some knowledge of the C programming language.
Organisation of this Book
This book is not intended to be used as an internals manual for Linux. Instead
it is an introduction to operating systems in general and to Linux in particular.
The chapters each follow my rule of \working from the general to the particular".
They rst give an overview of the kernel subsystem that they are describing before
launching into its gory details.
I have deliberately not described the kernel's algorithms, its methods of doing things,
in terms of routine X() calls routine Y() which increments the foo eld of the bar
data structure. You can read the code to nd these things out. Whenever I need to
understand a piece of code or describe it to someone else I often start with drawing
its data structures on the white-board. So, I have described many of the relevant
kernel data structures and their interrelationships in a fair amount of detail.
Each chapter is fairly independent, like the Linux kernel subsystem that they each
describe. Sometimes, though, there are linkages; for example you cannot describe a
process without understanding how virtual memory works.
The Hardware Basics chapter (Chapter 1) gives a brief introduction to the modern
PC. An operating system has to work closely with the hardware system that acts
as its foundations. The operating system needs certain services that can only be
provided by the hardware. In order to fully understand the Linux operating system,
you need to understand the basics of the underlying hardware.
The Software Basics chapter (Chapter 2) introduces basic software principles and
looks at assembly and C programing languages. It looks at the tools that are used
to build an operating system like Linux and it gives an overview of the aims and
functions of an operating system.
The Memory Management chapter (Chapter 3) describes the way that Linux handles
the physical and virtual memory in the system.
The Processes chapter (Chapter 4) describes what a process is and how the Linux
kernel creates, manages and deletes the processes in the system.
Processes communicate with each other and with the kernel to coordinate their activities.
Linux supports a number of Inter-Process Communication (IPC) mechanisms.
Signals and pipes are two of them but Linux also supports the System V IPC mechanisms
named after the UnixTM release in which they rst appeared. These interprocess
communications mechanisms are described in Chapter 5.
vi
The Peripheral Component Interconnect (PCI) standard is now rmly established
as the low cost, high performance data bus for PCs. The PCI chapter (Chapter 6)
describes how the Linux kernel initializes and uses PCI buses and devices in the
system.
The Interrupts and Interrupt Handling chapter (Chapter 7) looks at how the Linux
kernel handles interrupts. Whilst the kernel has generic mechanisms and interfaces
for handling interrupts, some of the interrupt handling details are hardware and
architecture specic.
One of Linux's strengths is its support for the many available hardware devices for
the modern PC. The Device Drivers chapter (Chapter 8) describes how the Linux
kernel controls the physical devices in the system.
The File system chapter (Chapter 9) describes how the Linux kernel maintains the
les in the le systems that it supports. It describes the Virtual File System (VFS)
and how the Linux kernel's real le systems are supported.
Networking and Linux are terms that are almost synonymous. In a very real sense
Linux is a product of the Internet or World Wide Web (WWW). Its developers and
users use the web to exchange information ideas, code and Linux itself is often used
to support the networking needs of organizations. Chapter 10 describes how Linux
supports the network protocols known collectively as TCP/IP.
The Kernel Mechanisms chapter (Chapter 11) looks at some of the general tasks and
mechanisms that the Linux kernel needs to supply so that other parts of the kernel
work eectively together.
The Modules chapter (Chapter 12) describes how the Linux kernel can dynamically
load functions, for example le systems, only when they are needed.
The Processors chapter (Chapter 13) gives a brief description of some of the processors
that Linux has been ported to.
The Sources chapter (Chapter 14) describes where in the Linux kernel sources you
should start looking for particular kernel functions.
Conventions used in this Book
The following is a list of the typographical conventions used in this book.
serif font identies commands or other text that is to be typed
literally by the user.
type font refers to data structures or elds
within data structures.
Throughout the text there references to pieces of code within the Linux kernel source
tree (for example the boxed margin note adjacent to this text ). These See foo() in are given
foo/bar.c
in case you wish to look at the source code itself and all of the le references are
relative to /usr/src/linux. Taking foo/bar.c as an example, the full lename
would be /usr/src/linux/foo/bar.c If you are running Linux (and you should),
then looking at the code is a worthwhile experience and you can use this book as an
aid to understanding the code and as a guide to its many data structures.
vii
Trademarks
ARM is a trademark of ARM Holdings PLC.
Caldera, OpenLinux and the \C" logo are trademarks of Caldera, Inc.
Caldera OpenDOS 1997 Caldera, Inc.
DEC is a trademark of Digital Equipment Corporation.
DIGITAL is a trademark of Digital Equipment Corporation.
Linux is a trademark of Linus Torvalds.
Motif is a trademark of The Open System Foundation, Inc.
MSDOS is a trademark of Microsoft Corporation.
Red Hat, glint and the Red Hat logo are trademarks of Red Hat Software, Inc.
UNIX is a registered trademark of X/Open.
XFree86 is a trademark of XFree86 Project, Inc.
X Window System is a trademark of the X Consortium and the Massachusetts Institute
of Technology.
The Author
I was born in 1957, a few weeks before Sputnik was launched, in the north of England.
I rst met Unix at University, where a lecturer used it as an example when teaching
the notions of kernels, scheduling and other operating systems goodies. I loved using
the newly delivered PDP-11 for my nal year project. After graduating (in 1982 with
a First Class Honours degree in Computer Science) I worked for Prime Computers
(Primos) and then after a couple of years for Digital (VMS, Ultrix). At Digital I
worked on many things but for the last 5 years there, I worked for the semiconductor
group on Alpha and StrongARM evaluation boards. In 1998 I moved to ARM where
I have a small group of engineers writing low level rmware and porting operating
systems. My children (Esther and Stephen) describe me as a geek.
People often ask me about Linux at work and at home and I am only too happy
to oblige. The more that I use Linux in both my professional and personal life the
more that I become a Linux zealot. You may note that I use the term `zealot' and
not `bigot'; I dene a Linux zealot to be an enthusiast that recognizes that there
are other operating systems but prefers not to use them. As my wife, Gill, who
uses Windows 95 once remarked \I never realized that we would have his and her
operating systems". For me, as an engineer, Linux suits my needs perfectly. It is
a superb,
exible and adaptable engineering tool that I use at work and at home.
Most freely available software easily builds on Linux and I can often simply download
pre-built executable les or install them from a CD ROM. What else could I use to
learn to program in C++, Perl or learn about Java for free?
Acknowledgements
I must thank the many people who have been kind enough to take the time to email
me with comments about this book. I have attempted to incorporated those
viii
comments in each new version that I have produced and I am more than happy to
receive comments, however please note my new e-mail address.
A number of lecturers have written to me asking if they can use some or parts of
this book in order to teach computing. My answer is an emphatic yes; this is one
use of the book that I particularly wanted. Who knows, there may be another Linus
Torvalds sat in the class.
Special thanks must go to John Rigby and Michael Bauer who gave me full, detailed
review notes of the whole book. Not an easy task. Alan Cox and Stephen Tweedie
have patiently answered my questions - thanks. I used Larry Ewing's penguins to
brighten up the chapters a bit. Finally, thank you to Greg Hankins for accepting
this book into the Linux Documentation Project and onto their web site.
Contents
Preface iii
1 Hardware Basics 1
1.1 The CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Buses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Controllers and Peripherals . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Address Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Timers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Software Basics 7
2.1 Computer Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Assembly Languages . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.2 The C Programming Language and Compiler . . . . . . . . . . 8
2.1.3 Linkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 What is an Operating System? . . . . . . . . . . . . . . . . . . . . . . 9
2.2.1 Memory management . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.2 Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.3 Device drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.4 The Filesystems . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Kernel Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Linked Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 Hash Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.3 Abstract Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . 13
3 Memory Management 15
3.1 An Abstract Model of Virtual Memory . . . . . . . . . . . . . . . . . . 16
3.1.1 Demand Paging . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.2 Swapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1.3 Shared Virtual Memory . . . . . . . . . . . . . . . . . . . . . . 19
3.1.4 Physical and Virtual Addressing Modes . . . . . . . . . . . . . 19
3.1.5 Access Control . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
ix
x CONTENTS
3.2 Caches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.3 Linux Page Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.4 Page Allocation and Deallocation . . . . . . . . . . . . . . . . . . . . . 23
3.4.1 Page Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4.2 Page Deallocation . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.5 Memory Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.6 Demand Paging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7 The Linux Page Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.8 Swapping Out and Discarding Pages . . . . . . . . . . . . . . . . . . . 28
3.8.1 Reducing the Size of the Page and Buer Caches . . . . . . . . 29
3.8.2 Swapping Out System V Shared Memory Pages . . . . . . . . . 30
3.8.3 Swapping Out and Discarding Pages . . . . . . . . . . . . . . . 30
3.9 The Swap Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.10 Swapping Pages In . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4 Processes 35
4.1 Linux Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2 Identiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.3 Scheduling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.3.1 Scheduling in Multiprocessor Systems . . . . . . . . . . . . . . 41
4.4 Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.5 Virtual Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.6 Creating a Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.7 Times and Timers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.8 Executing Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.8.1 ELF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.8.2 Script Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5 Interprocess Communication Mechanisms 51
5.1 Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.2 Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.3 Sockets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.1 System V IPC Mechanisms . . . . . . . . . . . . . . . . . . . . 55
5.3.2 Message Queues . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.3.3 Semaphores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.3.4 Shared Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6 PCI 61
6.1 PCI Address Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.2 PCI Conguration Headers . . . . . . . . . . . . . . . . . . . . . . . . 62
6.3 PCI I/O and PCI Memory Addresses . . . . . . . . . . . . . . . . . . . 64
CONTENTS xi
6.4 PCI-ISA Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.5 PCI-PCI Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.5.1 PCI-PCI Bridges: PCI I/O and PCI Memory Windows . . . . 65
6.5.2 PCI-PCI Bridges: PCI Conguration Cycles and PCI Bus
Numbering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.6 Linux PCI Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.6.1 The Linux Kernel PCI Data Structures . . . . . . . . . . . . . 67
6.6.2 The PCI Device Driver . . . . . . . . . . . . . . . . . . . . . . 68
6.6.3 PCI BIOS Functions . . . . . . . . . . . . . . . . . . . . . . . . 70
6.6.4 PCI Fixup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7 Interrupts and Interrupt Handling 75
7.1 Programmable Interrupt Controllers . . . . . . . . . . . . . . . . . . . 77
7.2 Initializing the Interrupt Handling Data Structures . . . . . . . . . . . 77
7.3 Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
8 Device Drivers 81
8.1 Polling and Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
8.2 Direct Memory Access (DMA) . . . . . . . . . . . . . . . . . . . . . . 84
8.3 Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
8.4 Interfacing Device Drivers with the Kernel . . . . . . . . . . . . . . . . 85
8.4.1 Character Devices . . . . . . . . . . . . . . . . . . . . . . . . . 86
8.4.2 Block Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
8.5 Hard Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
8.5.1 IDE Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
8.5.2 Initializing the IDE Subsystem . . . . . . . . . . . . . . . . . . 91
8.5.3 SCSI Disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
8.6 Network Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8.6.1 Initializing Network Devices . . . . . . . . . . . . . . . . . . . . 96
9 The File system 99
9.1 The Second Extended File system (EXT2) . . . . . . . . . . . . . . . . 101
9.1.1 The EXT2 Inode . . . . . . . . . . . . . . . . . . . . . . . . . . 102
9.1.2 The EXT2 Superblock . . . . . . . . . . . . . . . . . . . . . . . 103
9.1.3 The EXT2 Group Descriptor . . . . . . . . . . . . . . . . . . . 104
9.1.4 EXT2 Directories . . . . . . . . . . . . . . . . . . . . . . . . . . 104
9.1.5 Finding a File in an EXT2 File System . . . . . . . . . . . . . 105
9.1.6 Changing the Size of a File in an EXT2 File System . . . . . . 105
9.2 The Virtual File System (VFS) . . . . . . . . . . . . . . . . . . . . . . 107
9.2.1 The VFS Superblock . . . . . . . . . . . . . . . . . . . . . . . . 108
9.2.2 The VFS Inode . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
xii CONTENTS
9.2.3 Registering the File Systems . . . . . . . . . . . . . . . . . . . 110
9.2.4 Mounting a File System . . . . . . . . . . . . . . . . . . . . . . 110
9.2.5 Finding a File in the Virtual File System . . . . . . . . . . . . 111
9.2.6 Creating a File in the Virtual File System . . . . . . . . . . . . 112
9.2.7 Unmounting a File System . . . . . . . . . . . . . . . . . . . . 112
9.2.8 The VFS Inode Cache . . . . . . . . . . . . . . . . . . . . . . . 112
9.2.9 The Directory Cache . . . . . . . . . . . . . . . . . . . . . . . . 113
9.3 The Buer Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
9.3.1 The bdflush Kernel Daemon . . . . . . . . . . . . . . . . . . . 115
9.3.2 The update Process . . . . . . . . . . . . . . . . . . . . . . . . . 116
9.4 The /proc File System . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
9.5 Device Special Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10 Networks 119
10.1 An Overview of TCP/IP Networking . . . . . . . . . . . . . . . . . . . 119
10.2 The Linux TCP/IP Networking Layers . . . . . . . . . . . . . . . . . . 122
10.3 The BSD Socket Interface . . . . . . . . . . . . . . . . . . . . . . . . . 124
10.4 The INET Socket Layer . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.4.1 Creating a BSD Socket . . . . . . . . . . . . . . . . . . . . . . 127
10.4.2 Binding an Address to an INET BSD Socket . . . . . . . . . . 127
10.4.3 Making a Connection on an INET BSD Socket . . . . . . . . . 128
10.4.4 Listening on an INET BSD Socket . . . . . . . . . . . . . . . . 129
10.4.5 Accepting Connection Requests . . . . . . . . . . . . . . . . . . 130
10.5 The IP Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
10.5.1 Socket Buers . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
10.5.2 Receiving IP Packets . . . . . . . . . . . . . . . . . . . . . . . . 131
10.5.3 Sending IP Packets . . . . . . . . . . . . . . . . . . . . . . . . . 132
10.5.4 Data Fragmentation . . . . . . . . . . . . . . . . . . . . . . . . 133
10.6 The Address Resolution Protocol (ARP) . . . . . . . . . . . . . . . . . 133
10.7 IP Routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
10.7.1 The Route Cache . . . . . . . . . . . . . . . . . . . . . . . . . . 135
10.7.2 The Forwarding Information Database . . . . . . . . . . . . . . 136
11 Kernel Mechanisms 139
11.1 Bottom Half Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
11.2 Task Queues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
11.3 Timers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
11.4 Wait Queues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
11.5 Buzz Locks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
11.6 Semaphores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
CONTENTS xiii
12 Modules 145
12.1 Loading a Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
12.2 Unloading a Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
13 Processors 151
13.1 X86 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
13.2 ARM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
13.3 Alpha AXP Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
14 The Linux Kernel Sources 153
A Linux Data Structures 159
B Useful Web and FTP Sites 177
C Linux Documentation Project Manifesto 179
C.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
C.2 Getting Involved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
C.3 Current Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
C.4 FTP sites for LDP works . . . . . . . . . . . . . . . . . . . . . . . . . 180
C.5 Documentation Conventions . . . . . . . . . . . . . . . . . . . . . . . . 180
C.6 Copyright and License . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
C.7 Publishing LDP Manuals . . . . . . . . . . . . . . . . . . . . . . . . . 181
D The GNU General Public License 183
D.1 Preamble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
D.2 Terms and Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
D.3 How to Apply These Terms . . . . . . . . . . . . . . . . . . . . . . . . 188
Glossary 191
Bibliography 194
xiv CONTENTS
List of Figures
1.1 A typical PC motherboard. . . . . . . . . . . . . . . . . . . . . . . . . 2
3.1 Abstract model of Virtual to Physical address mapping . . . . . . . . 16
3.2 Alpha AXP Page Table Entry . . . . . . . . . . . . . . . . . . . . . . . 20
3.3 Three Level Page Tables . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.4 The free area data structure . . . . . . . . . . . . . . . . . . . . . . . 25
3.5 Areas of Virtual Memory . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.6 The Linux Page Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1 A Process's Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 A Process's Virtual Memory . . . . . . . . . . . . . . . . . . . . . . . . 44
4.3 Registered Binary Formats . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4 ELF Executable File Format . . . . . . . . . . . . . . . . . . . . . . . 49
5.1 Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
5.2 System V IPC Message Queues . . . . . . . . . . . . . . . . . . . . . . 56
5.3 System V IPC Semaphores . . . . . . . . . . . . . . . . . . . . . . . . 57
5.4 System V IPC Shared Memory . . . . . . . . . . . . . . . . . . . . . . 59
6.1 Example PCI Based System . . . . . . . . . . . . . . . . . . . . . . . . 62
6.2 The PCI Conguration Header . . . . . . . . . . . . . . . . . . . . . . 63
6.3 Type 0 PCI Conguration Cycle . . . . . . . . . . . . . . . . . . . . . 65
6.4 Type 1 PCI Conguration Cycle . . . . . . . . . . . . . . . . . . . . . 65
6.5 Linux Kernel PCI Data Structures . . . . . . . . . . . . . . . . . . . . 67
6.6 Conguring a PCI System: Part 1 . . . . . . . . . . . . . . . . . . . . 69
6.7 Conguring a PCI System: Part 2 . . . . . . . . . . . . . . . . . . . . 70
6.8 Conguring a PCI System: Part 3 . . . . . . . . . . . . . . . . . . . . 71
6.9 Conguring a PCI System: Part 4 . . . . . . . . . . . . . . . . . . . . 71
6.10 PCI Conguration Header: Base Address Registers . . . . . . . . . . . 72
7.1 A Logical Diagram of Interrupt Routing . . . . . . . . . . . . . . . . . 76
7.2 Linux Interrupt Handling Data Structures . . . . . . . . . . . . . . . . 79
8.1 Character Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
xv
xvi LIST OF FIGURES
8.2 Buer Cache Block Device Requests . . . . . . . . . . . . . . . . . . . 87
8.3 Linked list of disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.4 SCSI Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
9.1 Physical Layout of the EXT2 File system . . . . . . . . . . . . . . . . 101
9.2 EXT2 Inode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
9.3 EXT2 Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
9.4 A Logical Diagram of the Virtual File System . . . . . . . . . . . . . . 107
9.5 Registered File Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 110
9.6 A Mounted File System . . . . . . . . . . . . . . . . . . . . . . . . . . 112
9.7 The Buer Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10.1 TCP/IP Protocol Layers . . . . . . . . . . . . . . . . . . . . . . . . . . 121
10.2 Linux Networking Layers . . . . . . . . . . . . . . . . . . . . . . . . . 123
10.3 Linux BSD Socket Data Structures . . . . . . . . . . . . . . . . . . . . 126
10.4 The Socket Buer (sk bu) . . . . . . . . . . . . . . . . . . . . . . . . 131
10.5 The Forwarding Information Database . . . . . . . . . . . . . . . . . . 136
11.1 Bottom Half Handling Data Structures . . . . . . . . . . . . . . . . . . 139
11.2 A Task Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
11.3 System Timers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
11.4 Wait Queue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
12.1 The List of Kernel Modules . . . . . . . . . . . . . . . . . . . . . . . . 147
Chapter 1
Hardware Basics
An operating system has to work closely with the hardware system that
acts as its foundations. The operating system needs certain services that
can only be provided by the hardware. In order to fully understand
the Linux operating system, you need to understand the basics of the
underlying hardware. This chapter gives a brief introduction to that
hardware: the modern PC.
When the \Popular Electronics" magazine for January 1975 was printed with an
illustration of the Altair 8080 on its front cover, a revolution started. The Altair
8080, named after the destination of an early Star Trek episode, could be assembled
by home electronics enthusiasts for a mere $397. With its Intel 8080 processor and
256 bytes of memory but no screen or keyboard it was puny by today's standards.
Its inventor, Ed Roberts, coined the term \personal computer" to describe his new
invention, but the term PC is now used to refer to almost any computer that you
can pick up without needing help. By this denition, even some of the very powerful
Alpha AXP systems are PCs.
Enthusiastic hackers saw the Altair's potential and started to write software and
build hardware for it. To these early pioneers it represented freedom; the freedom
from huge batch processing mainframe systems run and guarded by an elite priesthood.
Overnight fortunes were made by college dropouts fascinated by this new
phenomenon, a computer that you could have at home on your kitchen table. A lot
of hardware appeared, all dierent to some degree and software hackers were happy
to write software for these new machines. Paradoxically it was IBM who rmly cast
the mould of the modern PC by announcing the IBM PC in 1981 and shipping it to
customers early in 1982. With its Intel 8088 processor, 64K of memory (expandable
to 256K), two
oppy disks and an 80 character by 25 lines Colour Graphics Adapter
1
2 Chapter 1. Hardware Basics
CPU
parallel port
COM1 COM2
power power
ISA Slots
Memory SIMM Slots
PCI Slots
Figure 1.1: A typical PC motherboard.
(CGA) it was not very powerful by today's standards but it sold well. It was followed,
in 1983, by the IBM PC-XT which had the luxury of a 10Mbyte hard drive.
It was not long before IBM PC clones were being produced by a host of companies
such as Compaq and the architecture of the PC became a de-facto standard. This
de-facto standard helped a multitude of hardware companies to compete together in
a growing market which, happily for consumers, kept prices low. Many of the system
architectural features of these early PCs have carried over into the modern PC. For
example, even the most powerful Intel Pentium Pro based system starts running in
the Intel 8086's addressing mode. When Linus Torvalds started writing what was
to become Linux, he picked the most plentiful and reasonably priced hardware, an
Intel 80386 PC.
Looking at a PC from the outside, the most obvious components are a system box,
a keyboard, a mouse and a video monitor. On the front of the system box are some
buttons, a little display showing some numbers and a
oppy drive. Most systems
these days have a CD ROM and if you feel that you have to protect your data, then
there will also be a tape drive for backups. These devices are collectively known as
the peripherals.
Although the CPU is in overall control of the system, it is not the only intelligent
device. All of the peripheral controllers, for example the IDE controller, have some
level of intelligence. Inside the PC (Figure 1.1) you will see a motherboard containing
the CPU or microprocessor, the memory and a number of slots for the ISA or PCI
peripheral controllers. Some of the controllers, for example the IDE disk controller
may be built directly onto the system board.
1.1. The CPU 3
1.1 The CPU
The CPU, or rather microprocessor, is the heart of any computer system. The microprocessor
calculates, performs logical operations and manages data
ows by reading
instructions from memory and then executing them. In the early days of computing
the functional components of the microprocessor were separate (and physically
large) units. This is when the term Central Processing Unit was coined. The modern
microprocessor combines these components onto an integrated circuit etched onto
a very small piece of silicon. The terms CPU, microprocessor and processor are all
used interchangeably in this book.
Microprocessors operate on binary data; that is data composed of ones and zeros.
These ones and zeros correspond to electrical switches being either on or o. Just
as 42 is a decimal number meaning \4 10s and 2 units", a binary number is a series
of binary digits each one representing a power of 2. In this context, a power means
the number of times that a number is multiplied by itself. 10 to the power 1 ( 101 )
is 10, 10 to the power 2 ( 102 ) is 10x10, 103 is 10x10x10 and so on. Binary 0001 is
decimal 1, binary 0010 is decimal 2, binary 0011 is 3, binary 0100 is 4 and so on. So,
42 decimal is 101010 binary or (2+8+32 or 21 +23 +25). Rather than using binary
to represent numbers in computer programs, another base, hexadecimal is usually
used. In this base, each digital represents a power of 16. As decimal numbers only
go from 0 to 9 the numbers 10 to 15 are represented as a single digit by the letters
A, B, C, D, E and F. For example, hexadecimal E is decimal 14 and hexadecimal 2A
is decimal 42 (two 16s) + 10). Using the C programming language notation (as I do
throughout this book) hexadecimal numbers are prefaced by \0x"; hexadecimal 2A
is written as 0x2A .
Microprocessors can perform arithmetic operations such as add, multiply and divide
and logical operations such as \is X greater than Y?".
The processor's execution is governed by an external clock. This clock, the system
clock, generates regular clock pulses to the processor and, at each clock pulse, the
processor does some work. For example, a processor could execute an instruction
every clock pulse. A processor's speed is described in terms of the rate of the system
clock ticks. A 100Mhz processor will receive 100,000,000 clock ticks every second. It
is misleading to describe the power of a CPU by its clock rate as dierent processors
perform dierent amounts of work per clock tick. However, all things being equal, a
faster clock speed means a more powerful processor. The instructions executed by the
processor are very simple; for example \read the contents of memory at location X
into register Y". Registers are the microprocessor's internal storage, used for storing
data and performing operations on it. The operations performed may cause the
processor to stop what it is doing and jump to another instruction somewhere else in
memory. These tiny building blocks give the modern microprocessor almost limitless
power as it can execute millions or even billions of instructions a second.
The instructions have to be fetched from memory as they are executed. Instructions
may themselves reference data within memory and that data must be fetched from
memory and saved there when appropriate.
The size, number and type of register within a microprocessor is entirely dependent
on its type. An Intel 4086 processor has a dierent register set to an Alpha AXP
processor; for a start, the Intel's are 32 bits wide and the Alpha AXP's are 64 bits
wide. In general, though, any given processor will have a number of general purpose
4 Chapter 1. Hardware Basics
registers and a smaller number of dedicated registers. Most processors have the
following special purpose, dedicated, registers:
Program Counter (PC) This register contains the address of the next instruction
to be executed. The contents of the PC are automatically incremented each
time an instruction is fetched,
Stack Pointer (SP) Processors have to have access to large amounts of external
read/write random access memory (RAM) which facilitates temporary storage
of data. The stack is a way of easily saving and restoring temporary values in
external memory. Usually, processors have special instructions which allow you
to push values onto the stack and to pop them o again later. The stack works
on a last in rst out (LIFO) basis. In other words, if you push two values, x
and y, onto a stack and then pop a value o of the stack then you will get back
the value of y.
Some processor's stacks grow upwards towards the top of memory whilst others
grow downwards towards the bottom, or base, of memory. Some processor's
support both types, for example ARM.
Processor Status (PS) Instructions may yield results; for example \is the content
of register X greater than the content of register Y?" will yield true or false as
a result. The PS register holds this and other information about the current
state of the processor. For example, most processors have at least two modes
of operation, kernel (or supervisor) and user. The PS register would hold
information identifying the current mode.
1.2 Memory
All systems have a memory hierarchy with memory at dierent speeds and sizes at
dierent points in the hierarchy. The fastest memory is known as cache memory and
is what it sounds like - memory that is used to temporarily hold, or cache, contents
of the main memory. This sort of memory is very fast but expensive, therefore most
processors have a small amount of on-chip cache memory and more system based (onboard)
cache memory. Some processors have one cache to contain both instructions
and data, but others have two, one for instructions and the other for data. The
Alpha AXP processor has two internal memory caches; one for data (the D-Cache)
and one for instructions (the I-Cache). The external cache (or B-Cache) mixes the
two together. Finally there is the main memory which relative to the external cache
memory is very slow. Relative to the on-CPU cache, main memory is positively
crawling.
The cache and main memories must be kept in step (coherent). In other words, if
a word of main memory is held in one or more locations in cache, then the system
must make sure that the contents of cache and memory are the same. The job of
cache coherency is done partially by the hardware and partially by the operating
system. This is also true for a number of major system tasks where the hardware
and software must cooperate closely to achieve their aims.
1.3. Buses 5
1.3 Buses
The individual components of the system board are interconnected by multiple connection
systems known as buses. The system bus is divided into three logical functions;
the address bus, the data bus and the control bus. The address bus species
the memory locations (addresses) for the data transfers. The data bus holds the data
transfered. The data bus is bidirectional; it allows data to be read into the CPU and
written from the CPU. The control bus contains various lines used to route timing
and control signals throughout the system. Many
avours of bus exist, for example
ISA and PCI buses are popular ways of connecting peripherals to the system.
1.4 Controllers and Peripherals
Peripherals are real devices, such as graphics cards or disks controlled by controller
chips on the system board or on cards plugged into it. The IDE disks are controlled
by the IDE controller chip and the SCSI disks by the SCSI disk controller chips and
so on. These controllers are connected to the CPU and to each other by a variety
of buses. Most systems built now use PCI and ISA buses to connect together the
main system components. The controllers are processors like the CPU itself, they
can be viewed as intelligent helpers to the CPU. The CPU is in overall control of the
system.
All controllers are dierent, but they usually have registers which control them.
Software running on the CPU must be able to read and write those controlling
registers. One register might contain status describing an error. Another might be
used for control purposes; changing the mode of the controller. Each controller on
a bus can be individually addressed by the CPU, this is so that the software device
driver can write to its registers and thus control it. The IDE ribbon is a good example,
as it gives you the ability to access each drive on the bus separately. Another good
example is the PCI bus which allows each device (for example a graphics card) to be
accessed independently.
1.5 Address Spaces
The system bus connects the CPU with the main memory and is separate from the
buses connecting the CPU with the system's hardware peripherals. Collectively the
memory space that the hardware peripherals exist in is known as I/O space. I/O
space may itself be further subdivided, but we will not worry too much about that
for the moment. The CPU can access both the system space memory and the I/O
space memory, whereas the controllers themselves can only access system memory
indirectly and then only with the help of the CPU. From the point of view of the
device, say the
oppy disk controller, it will see only the address space that its
control registers are in (ISA), and not the system memory. Typically a CPU will
have separate instructions for accessing the memory and I/O space. For example,
there might be an instruction that means \read a byte from I/O address 0x3f0 into
register X". This is exactly how the CPU controls the system's hardware peripherals,
by reading and writing to their registers in I/O space. Where in I/O space the
common peripherals (IDE controller, serial port,
oppy disk controller and so on)
have their registers has been set by convention over the years as the PC architecture
6 Chapter 1. Hardware Basics
has developed. The I/O space address 0x3f0 just happens to be the address of one
of the serial port's (COM1) control registers.
There are times when controllers need to read or write large amounts of data directly
to or from system memory. For example when user data is being written to the
hard disk. In this case, Direct Memory Access (DMA) controllers are used to allow
hardware peripherals to directly access system memory but this access is under strict
control and supervision of the CPU.
1.6 Timers
All operating systems need to know the time and so the modern PC includes a special
peripheral called the Real Time Clock (RTC). This provides two things: a reliable
time of day and an accurate timing interval. The RTC has its own battery so that
it continues to run even when the PC is not powered on, this is how your PC always
\knows" the correct date and time. The interval timer allows the operating system
to accurately schedule essential work.
Chapter 2
Software Basics
A program is a set of computer instructions that perform a particular task.
That program can be written in assembler, a very low level computer
language, or in a high level, machine independent language such as the
C programming language. An operating system is a special program
which allows the user to run applications such as spreadsheets and word
processors. This chapter introduces basic programming principles and
gives an overview of the aims and functions of an operating system.
2.1 Computer Languages
2.1.1 Assembly Languages
The instructions that a CPU fetches from memory and executes are not at all understandable
to human beings. They are machine codes which tell the computer
precisely what to do. The hexadecimal number 0x89E5 is an Intel 80486 instruction
which copies the contents of the ESP register to the EBP register. One of the rst
software tools invented for the earliest computers was an assembler, a program which
takes a human readable source le and assembles it into machine code. Assembly
languages explicitly handle registers and operations on data and they are specic to
a particular microprocessor. The assembly language for an Intel X86 microprocessor
is very dierent to the assembly language for an Alpha AXP microprocessor. The
following Alpha AXP assembly code shows the sort of operations that a program can
perform:
ldr r16, (r15) ; Line 1
7
8 Chapter 2. Software Basics
ldr r17, 4(r15) ; Line 2
beq r16,r17,100 ; Line 3
str r17, (r15) ; Line 4
100: ; Line 5
The rst statement (on line 1) loads register 16 from the address held in register
15. The next instruction loads register 17 from the next location in memory. Line 3
compares the contents of register 16 with that of register 17 and, if they are equal,
branches to label 100. If the registers do not contain the same value then the program
continues to line 4 where the contents of r17 are saved into memory. If the registers
do contain the same value then no data needs to be saved. Assembly level programs
are tedious and tricky to write and prone to errors. Very little of the Linux kernel is
written in assembly language and those parts that are are written only for eciency
and they are specic to particular microprocessors.
2.1.2 The C Programming Language and Compiler
Writing large programs in assembly language is a dicult and time consuming task.
It is prone to error and the resulting program is not portable, being tied to one
particular processor family. It is far better to use a machine independent language
like C[7, The C Programming Language]. C allows you to describe programs in terms
of their logical algorithms and the data that they operate on. Special programs called
compilers read the C program and translate it into assembly language, generating
machine specic code from it. A good compiler can generate assembly instructions
that are very nearly as ecient as those written by a good assembly programmer.
Most of the Linux kernel is written in the C language. The following C fragment:
if (x != y)
x = y ;
performs exactly the same operations as the previous example assembly code. If the
contents of the variable x are not the same as the contents of variable y then the
contents of y will be copied to x. C code is organized into routines, each of which
perform a task. Routines may return any value or data type supported by C. Large
programs like the Linux kernel comprise many separate C source modules each with
its own routines and data structures. These C source code modules group together
logical functions such as lesystem handling code.
C supports many types of variables, a variable is a location in memory which can be
referenced by a symbolic name. In the above C fragment x and y refer to locations
in memory. The programmer does not care where in memory the variables are put,
it is the linker (see below) that has to worry about that. Some variables contain
dierent sorts of data, integer and
oating point and others are pointers.
Pointers are variables that contain the address, the location in memory of other
data. Consider a variable called x, it might live in memory at address 0x80010000.
You could have a pointer, called px, which points at x. px might live at address
0x80010030. The value of px would be 0x80010000: the address of the variable x.
C allows you to bundle together related variables into data structures. For example,
struct {
2.2. What is an Operating System? 9
int i ;
char b ;
} my_struct ;
is a data structure called my struct which contains two elements, an integer (32 bits
of data storage) called i and a character (8 bits of data) called b.
2.1.3 Linkers
Linkers are programs that link together several object modules and libraries to form
a single, coherent, program. Object modules are the machine code output from an
assembler or compiler and contain executable machine code and data together with
information that allows the linker to combine the modules together to form a program.
For example one module might contain all of a program's database functions
and another module its command line argument handling functions. Linkers x up
references between these object modules, where a routine or data structure referenced
in one module actually exists in another module. The Linux kernel is a single,
large program linked together from its many constituent object modules.
2.2 What is an Operating System?
Without software a computer is just a pile of electronics that gives o heat. If the
hardware is the heart of a computer then the software is its soul. An operating system
is a collection of system programs which allow the user to run application software.
The operating system abstracts the real hardware of the system and presents the
system's users and its applications with a virtual machine. In a very real sense
the software provides the character of the system. Most PCs can run one or more
operating systems and each one can have a very dierent look and feel. Linux is
made up of a number of functionally separate pieces that, together, comprise the
operating system. One obvious part of Linux is the kernel itself; but even that would
be useless without libraries or shells.
In order to start understanding what an operating system is, consider what happens
when you type an apparently simple command:
$ ls
Mail c images perl
docs tcl
$
The $ is a prompt put out by a login shell (in this case bash). This means that it
is waiting for you, the user, to type some command. Typing ls causes the keyboard
driver to recognize that characters have been typed. The keyboard driver passes
them to the shell which processes that command by looking for an executable image
of the same name. It nds that image, in /bin/ls. Kernel services are called to pull
the ls executable image into virtual memory and start executing it. The ls image
makes calls to the le subsystem of the kernel to nd out what les are available.
The lesystem might make use of cached lesystem information or use the disk
10 Chapter 2. Software Basics
device driver to read this information from the disk. It might even cause a network
driver to exchange information with a remote machine to nd out details of remote
les that this system has access to (lesystems can be remotely mounted via the
Networked File System or NFS). Whichever way the information is located, ls writes
that information out and the video driver displays it on the screen.
All of the above seems rather complicated but it shows that even most simple commands
reveal that an operating system is in fact a co-operating set of functions that
together give you, the user, a coherent view of the system.
2.2.1 Memory management
With innite resources, for example memory, many of the things that an operating
system has to do would be redundant. One of the basic tricks of any operating
system is the ability to make a small amount of physical memory behave like rather
more memory. This apparently large memory is known as virtual memory. The idea
is that the software running in the system is fooled into believing that it is running
in a lot of memory. The system divides the memory into easily handled pages and
swaps these pages onto a hard disk as the system runs. The software does not notice
because of another trick, multi-processing.
2.2.2 Processes
A process could be thought of as a program in action, each process is a separate
entity that is running a particular program. If you look at the processes on your
Linux system, you will see that there are rather a lot. For example, typing ps shows
the following processes on my system:
$ ps
PID TTY STAT TIME COMMAND
158 pRe 1 0:00 -bash
174 pRe 1 0:00 sh /usr/X11R6/bin/startx
175 pRe 1 0:00 xinit /usr/X11R6/lib/X11/xinit/xinitrc --
178 pRe 1 N 0:00 bowman
182 pRe 1 N 0:01 rxvt -geometry 120x35 -fg white -bg black
184 pRe 1 < 0:00 xclock -bg grey -geometry -1500-1500 -padding 0
185 pRe 1 < 0:00 xload -bg grey -geometry -0-0 -label xload
187 pp6 1 9:26 /bin/bash
202 pRe 1 N 0:00 rxvt -geometry 120x35 -fg white -bg black
203 ppc 2 0:00 /bin/bash
1796 pRe 1 N 0:00 rxvt -geometry 120x35 -fg white -bg black
1797 v06 1 0:00 /bin/bash
3056 pp6 3 < 0:02 emacs intro/introduction.tex
3270 pp6 3 0:00 ps
$
If my system had many CPUs then each process could (theoretically at least) run
on a dierent CPU. Unfortunately, there is only one so again the operating system
resorts to trickery by running each process in turn for a short period. This period of
time is known as a time-slice. This trick is known as multi-processing or scheduling
2.3. Kernel Data Structures 11
and it fools each process into thinking that it is the only process. Processes are
protected from one another so that if one process crashes or malfunctions then it will
not aect any others. The operating system achieves this by giving each process a
separate address space which only they have access to.
2.2.3 Device drivers
Device drivers make up the major part of the Linux kernel. Like other parts of the
operating system, they operate in a highly privileged environment and can cause
disaster if they get things wrong. Device drivers control the interaction between the
operating system and the hardware device that they are controlling. For example,
the lesystem makes use of a general block device interface when writing blocks to
an IDE disk. The driver takes care of the details and makes device specic things
happen. Device drivers are specic to the controller chip that they are driving which
is why, for example, you need the NCR810 SCSI driver if your system has an NCR810
SCSI controller.
2.2.4 The Filesystems
In Linux, as it is for UnixTM , the separate lesystems that the system may use
are not accessed by device identiers (such as a drive number or a drive name) but
instead they are combined into a single hierarchical tree structure that represents the
lesystem as a single entity. Linux adds each new lesystem into this single lesystem
tree as they are mounted onto a mount directory, for example /mnt/cdrom. One of
the most important features of Linux is its support for many dierent lesystems.
This makes it very
exible and well able to coexist with other operating systems. The
most popular lesystem for Linux is the EXT2 lesystem and this is the lesystem
supported by most of the Linux distributions.
A lesystem gives the user a sensible view of les and directories held on the hard
disks of the system regardless of the lesystem type or the characteristics of the
underlying physical device. Linux transparently supports many dierent lesystems
(for example MS-DOS and EXT2) and presents all of the mounted les and lesystems
as one integrated virtual lesystem. So, in general, users and processes do not need
to know what sort of lesystem that any le is part of, they just use them.
The block device drivers hide the dierences between the physical block device types
(for example, IDE and SCSI) and, so far as each lesystem is concerned, the physical
devices are just linear collections of blocks of data. The block sizes may vary between
devices, for example 512 bytes is common for
oppy devices whereas 1024 bytes is
common for IDE devices and, again, this is hidden from the users of the system. An
EXT2 lesystem looks the same no matter what device holds it.
2.3 Kernel Data Structures
The operating system must keep a lot of information about the current state of the
system. As things happen within the system these data structures must be changed
to re
ect the current reality. For example, a new process might be created when
a user logs onto the system. The kernel must create a data structure representing
12 Chapter 2. Software Basics
the new process and link it with the data structures representing all of the other
processes in the system.
Mostly these data structures exist in physical memory and are accessible only by
the kernel and its subsystems. Data structures contain data and pointers; addresses
of other data structures or the addresses of routines. Taken all together, the data
structures used by the Linux kernel can look very confusing. Every data structure
has a purpose and although some are used by several kernel subsystems, they are
more simple than they appear at rst sight.
Understanding the Linux kernel hinges on understanding its data structures and the
use that the various functions within the Linux kernel makes of them. This book
bases its description of the Linux kernel on its data structures. It talks about each
kernel subsystem in terms of its algorithms, its methods of getting things done, and
their usage of the kernel's data structures.
2.3.1 Linked Lists
Linux uses a number of software engineering techniques to link together its data
structures. On a lot of occasions it uses linked or chained data structures. If each
data structure describes a single instance or occurance of something, for example a
process or a network device, the kernel must be able to nd all of the instances. In a
linked list a root pointer contains the address of the rst data structure, or element,
in the list and each data structure contains a pointer to the next element in the list.
The last element's next pointer would be 0 or NULL to show that it is the end of the
list. In a doubly linked list each element contains both a pointer to the next element
in the list but also a pointer to the previous element in the list. Using doubly linked
lists makes it easier to add or remove elements from the middle of list although you
do need more memory accesses. This is a typical operating system trade o: memory
accesses versus CPU cycles.
2.3.2 Hash Tables
Linked lists are handy ways of tying data structures together but navigating linked
lists can be inecient. If you were searching for a particular element, you might
easily have to look at the whole list before you nd the one that you need. Linux
uses another technique, hashing to get around this restriction. A hash table is an
array or vector of pointers. An array, or vector, is simply a set of things coming one
after another in memory. A bookshelf could be said to be an array of books. Arrays
are accessed by an index, the index is an oset into the array. Taking the bookshelf
analogy a little further, you could describe each book by its position on the shelf;
you might ask for the 5th book.
A hash table is an array of pointers to data structures and its index is derived
from information in those data structures. If you had data structures describing
the population of a village then you could use a person's age as an index. To nd
a particular person's data you could use their age as an index into the population
hash table and then follow the pointer to the data structure containing the person's
details. Unfortunately many people in the village are likely to have the same age
and so the hash table pointer becomes a pointer to a chain or list of data structures
each describing people of the same age. However, searching these shorter chains is
2.3. Kernel Data Structures 13
still faster than searching all of the data structures.
As a hash table speeds up access to commonly used data structures, Linux often
uses hash tables to implement caches. Caches are handy information that needs to
be accessed quickly and are usually a subset of the full set of information available.
Data structures are put into a cache and kept there because the kernel often accesses
them. There is a drawback to caches in that they are more complex to use and
maintain than simple linked lists or hash tables. If the data structure can be found
in the cache (this is known as a cache hit, then all well and good. If it cannot then
all of the relevant data structures must be searched and, if the data structure exists
at all, it must be added into the cache. In adding new data structures into the cache
an old cache entry may need discarding. Linux must decide which one to discard,
the danger being that the discarded data structure may be the next one that Linux
needs.
2.3.3 Abstract Interfaces
The Linux kernel often abstracts its interfaces. An interface is a collection of routines
and data structures which operate in a particular way. For example all network
device drivers have to provide certain routines in which particular data structures
are operated on. This way there can be generic layers of code using the services
(interfaces) of lower layers of specic code. The network layer is generic and it is
supported by device specic code that conforms to a standard interface.
Often these lower layers register themselves with the upper layer at boot time. This
registration usually involves adding a data structure to a linked list. For example
each lesystem built into the kernel registers itself with the kernel at boot time
or, if you are using modules, when the lesystem is rst used. You can see which
lesystems have registered themselves by looking at the le /proc/filesystems.
The registration data structure often includes pointers to functions. These are the
addresses of software functions that perform particular tasks. Again, using lesystem
registration as an example, the data structure that each lesystem passes to the Linux
kernel as it registers includes the address of a lesystem specc routine which must
be called whenever that lesystem is mounted.
14 Chapter 2. Software Basics
Chapter 3
Memory Management
The memory management subsystem is one of the most important parts of
the operating system. Since the early days of computing, there has been
a need for more memory than exists physically in a system. Strategies
have been developed to overcome this limitation and the most successful
of these is virtual memory. Virtual memory makes the system appear to
have more memory than it actually has by sharing it between competing
processes as they need it.
Virtual memory does more than just make your computer's memory go further. The
memory management subsystem provides:
Large Address Spaces The operating system makes the system appear as if it has
a larger amount of memory than it actually has. The virtual memory can be
many times larger than the physical memory in the system,
Protection Each process in the system has its own virtual address space. These
virtual address spaces are completely separate from each other and so a process
running one application cannot aect another. Also, the hardware virtual
memory mechanisms allow areas of memory to be protected against writing.
This protects code and data from being overwritten by rogue applications.
Memory Mapping Memory mapping is used to map image and data les into a
processes address space. In memory mapping, the contents of a le are linked
directly into the virtual address space of a process.
Fair Physical Memory Allocation The memory management subsystem allows
each running process in the system a fair share of the physical memory of the
system,
15
16 Chapter 3. Memory Management
VPFN 0
VPFN 1
VPFN 2
VPFN 3
VPFN 4
VPFN 5
VPFN 6
VPFN 7
VIRTUAL MEMORY
Process X
VPFN 0
VPFN 1
VPFN 2
VPFN 3
VPFN 4
VPFN 5
VPFN 6
VPFN 7
Process Y
VIRTUAL MEMORY
Process X
Page Tables
PFN 0
PFN 1
PFN 2
PFN 3
PFN 4
PHYSICAL MEMORY
Page Tables
Process Y
Figure 3.1: Abstract model of Virtual to Physical address mapping
Shared Virtual Memory Although virtual memory allows processes to have separate
(virtual) address spaces, there are times when you need processes to share
memory. For example there could be several processes in the system running
the bash command shell. Rather than have several copies of bash, one in each
processes virtual address space, it is better to have only one copy in physical
memory and all of the processes running bash share it. Dynamic libraries are
another common example of executing code shared between several processes.
Shared memory can also be used as an Inter Process Communication (IPC)
mechanism, with two or more processes exchanging information via memory
common to all of them. Linux supports the UnixTM System V shared memory
IPC.
3.1 An Abstract Model of Virtual Memory
Before considering the methods that Linux uses to support virtual memory it is
useful to consider an abstract model that is not cluttered by too much detail.
As the processor executes a program it reads an instruction from memory and decodes
it. In decoding the instruction it may need to fetch or store the contents of a location
in memory. The processor then executes the instruction and moves onto the next
instruction in the program. In this way the processor is always accessing memory
either to fetch instructions or to fetch and store data.
In a virtual memory system all of these addresses are virtual addresses and not
physical addresses. These virtual addresses are converted into physical addresses by
the processor based on information held in a set of tables maintained by the operating
system.
To make this translation easier, virtual and physical memory are divided into handy
sized chunks called pages. These pages are all the same size, they need not be but if
they were not, the system would be very hard to administer. Linux on Alpha AXP
systems uses 8 Kbyte pages and on Intel x86 systems it uses 4 Kbyte pages. Each
3.1. An Abstract Model of Virtual Memory 17
of these pages is given a unique number; the page frame number (PFN). In this
paged model, a virtual address is composed of two parts; an oset and a virtual page
frame number. If the page size is 4 Kbytes, bits 11:0 of the virtual address contain
the oset and bits 12 and above are the virtual page frame number. Each time the
processor encounters a virtual address it must extract the oset and the virtual page
frame number. The processor must translate the virtual page frame number into
a physical one and then access the location at the correct oset into that physical
page. To do this the processor uses page tables.
Figure 3.1 shows the virtual address spaces of two processes, process X and process
Y, each with their own page tables. These page tables map each processes virtual
pages into physical pages in memory. This shows that process X's virtual page frame
number 0 is mapped into memory in physical page frame number 1 and that process
Y's virtual page frame number 1 is mapped into physical page frame number 4. Each
entry in the theoretical page table contains the following information:
 Valid
ag. This indicates if this page table entry is valid,
 The physical page frame number that this entry is describing,
 Access control information. This describes how the page may be used. Can it
be written to? Does it contain executable code?
The page table is accessed using the virtual page frame number as an oset. Virtual
page frame 5 would be the 6th element of the table (0 is the rst element).
To translate a virtual address into a physical one, the processor must rst work out
the virtual addresses page frame number and the oset within that virtual page. By
making the page size a power of 2 this can be easily done by masking and shifting.
Looking again at Figures 3.1 and assuming a page size of 0x2000 bytes (which is
decimal 8192) and an address of 0x2194 in process Y's virtual address space then
the processor would translate that address into oset 0x194 into virtual page frame
number 1.
The processor uses the virtual page frame number as an index into the processes
page table to retrieve its page table entry. If the page table entry at that oset is
valid, the processor takes the physical page frame number from this entry. If the
entry is invalid, the process has accessed a non-existent area of its virtual memory.
In this case, the processor cannot resolve the address and must pass control to the
operating system so that it can x things up.
Just how the processor noties the operating system that the correct process has
attempted to access a virtual address for which there is no valid translation is specic
to the processor. However the processor delivers it, this is known as a page fault and
the operating system is notied of the faulting virtual address and the reason for the
page fault.
Assuming that this is a valid page table entry, the processor takes that physical page
frame number and multiplies it by the page size to get the address of the base of the
page in physical memory. Finally, the processor adds in the oset to the instruction
or data that it needs.
Using the above example again, process Y's virtual page frame number 1 is mapped
to physical page frame number 4 which starts at 0x8000 (4 x 0x2000). Adding in the
0x194 byte oset gives us a nal physical address of 0x8194.
18 Chapter 3. Memory Management
By mapping virtual to physical addresses this way, the virtual memory can be
mapped into the system's physical pages in any order. For example, in Figure 3.1
process X's virtual page frame number 0 is mapped to physical page frame number
1 whereas virtual page frame number 7 is mapped to physical page frame number
0 even though it is higher in virtual memory than virtual page frame number 0.
This demonstrates an interesting byproduct of virtual memory; the pages of virtual
memory do not have to be present in physical memory in any particular order.
3.1.1 Demand Paging
As there is much less physical memory than virtual memory the operating system
must be careful that it does not use the physical memory ineciently. One way to
save physical memory is to only load virtual pages that are currently being used by
the executing program. For example, a database program may be run to query a
database. In this case not all of the database needs to be loaded into memory, just
those data records that are being examined. If the database query is a search query
then it does not make sense to load the code from the database program that deals
with adding new records. This technique of only loading virtual pages into memory
as they are accessed is known as demand paging.
When a process attempts to access a virtual address that is not currently in memory
the processor cannot nd a page table entry for the virtual page referenced. For
example, in Figure 3.1 there is no entry in process X's page table for virtual page
frame number 2 and so if process X attempts to read from an address within virtual
page frame number 2 the processor cannot translate the address into a physical
one. At this point the processor noties the operating system that a page fault has
occurred.
If the faulting virtual address is invalid this means that the process has attempted
to access a virtual address that it should not have. Maybe the application has gone
wrong in some way, for example writing to random addresses in memory. In this case
the operating system will terminate it, protecting the other processes in the system
from this rogue process.
If the faulting virtual address was valid but the page that it refers to is not currently
in memory, the operating system must bring the appropriate page into memory from
the image on disk. Disk access takes a long time, relatively speaking, and so the
process must wait quite a while until the page has been fetched. If there are other
processes that could run then the operating system will select one of them to run.
The fetched page is written into a free physical page frame and an entry for the
virtual page frame number is added to the processes page table. The process is then
restarted at the machine instruction where the memory fault occurred. This time
the virtual memory access is made, the processor can make the virtual to physical
address translation and so the process continues to run.
Linux uses demand paging to load executable images into a processes virtual memory.
Whenever a command is executed, the le containing it is opened and its contents
are mapped into the processes virtual memory. This is done by modifying the data
structures describing this processes memory map and is known as memory mapping.
However, only the rst part of the image is actually brought into physical memory.
The rest of the image is left on disk. As the image executes, it generates page faults
and Linux uses the processes memory map in order to determine which parts of the
3.1. An Abstract Model of Virtual Memory 19
image to bring into memory for execution.
3.1.2 Swapping
If a process needs to bring a virtual page into physical memory and there are no
free physical pages available, the operating system must make room for this page by
discarding another page from physical memory.
If the page to be discarded from physical memory came from an image or data le
and has not been written to then the page does not need to be saved. Instead it can
be discarded and if the process needs that page again it can be brought back into
memory from the image or data le.
However, if the page has been modied, the operating system must preserve the
contents of that page so that it can be accessed at a later time. This type of page is
known as a dirty page and when it is removed from memory it is saved in a special
sort of le called the swap le. Accesses to the swap le are very long relative to the
speed of the processor and physical memory and the operating system must juggle
the need to write pages to disk with the need to retain them in memory to be used
again.
If the algorithm used to decide which pages to discard or swap (the swap algorithm
is not ecient then a condition known as thrashing occurs. In this case, pages are
constantly being written to disk and then being read back and the operating system
is too busy to allow much real work to be performed. If, for example, physical
page frame number 1 in Figure 3.1 is being regularly accessed then it is not a good
candidate for swapping to hard disk. The set of pages that a process is currently
using is called the working set. An ecient swap scheme would make sure that all
processes have their working set in physical memory.
Linux uses a Least Recently Used (LRU) page aging technique to fairly choose pages
which might be removed from the system. This scheme involves every page in the
system having an age which changes as the page is accessed. The more that a page
is accessed, the younger it is; the less that it is accessed the older and more stale it
becomes. Old pages are good candidates for swapping.
3.1.3 Shared Virtual Memory
Virtual memory makes it easy for several processes to share memory. All memory
access are made via page tables and each process has its own separate page table.
For two processes sharing a physical page of memory, its physical page frame number
must appear in a page table entry in both of their page tables.
Figure 3.1 shows two processes that each share physical page frame number 4. For
process X this is virtual page frame number 4 whereas for process Y this is virtual
page frame number 6. This illustrates an interesting point about sharing pages: the
shared physical page does not have to exist at the same place in virtual memory for
any or all of the processes sharing it.
3.1.4 Physical and Virtual Addressing Modes
It does not make much sense for the operating system itself to run in virtual memory.
This would be a nightmare situation where the operating system must maintain page
20 Chapter 3. Memory Management
A
S
M
F
O
E
F
O
W
F
O
R
K
R
E
K
W
E
U
R
E
H
U G
W
E
63 32
31 15 14 13 12 11 10 9 8
PFN
7 6 5 4 3 2 1 0
__PAGE_ACCESSED
__PAGE_DIRTY
V
Figure 3.2: Alpha AXP Page Table Entry
tables for itself. Most multi-purpose processors support the notion of a physical
address mode as well as a virtual address mode. Physical addressing mode requires no
page tables and the processor does not attempt to perform any address translations
in this mode. The Linux kernel is linked to run in physical address space.
The Alpha AXP processor does not have a special physical addressing mode. Instead,
it divides up the memory space into several areas and designates two of them as
physically mapped addresses. This kernel address space is known as KSEG address
space and it encompasses all addresses upwards from 0xfc0000000000. In order to
execute from code linked in KSEG (by denition, kernel code) or access data there,
the code must be executing in kernel mode. The Linux kernel on Alpha is linked to
execute from address 0xfc0000310000.
3.1.5 Access Control
The page table entries also contain access control information. As the processor is
already using the page table entry to map a processes virtual address to a physical
one, it can easily use the access control information to check that the process is not
accessing memory in a way that it should not.
There are many reasons why you would want to restrict access to areas of memory.
Some memory, such as that containing executable code, is naturally read only memory;
the operating system should not allow a process to write data over its executable
code. By contrast, pages containing data can be written to but attempts to execute
that memory as instructions should fail. Most processors have at least two modes
of execution: kernel and user. You would not want kernel code executing by a user
or kernel data structures to be accessible except when the processor is running in
kernel mode.
The access control information is held in the PTE and is processor specic; gure 3.2
shows the PTE for Alpha AXP. The bit elds have the following meanings:
V Valid, if set this PTE is valid,
FOE \Fault on Execute", Whenever an attempt to execute instructions in this page
occurs, the processor reports a page fault and passes control to the operating
system,
3.2. Caches 21
FOW \Fault on Write", as above but page fault on an attempt to write to this
page,
FOR \Fault on Read", as above but page fault on an attempt to read from this
page,
ASM Address Space Match. This is used when the operating system wishes to clear
only some of the entries from the Translation Buer,
KRE Code running in kernel mode can read this page,
URE Code running in user mode can read this page,
GH Granularity hint used when mapping an entire block with a single Translation
Buer entry rather than many,
KWE Code running in kernel mode can write to this page,
UWE Code running in user mode can write to this page,
page frame number For PTEs with the V bit set, this eld contains the physical
Page Frame Number (page frame number) for this PTE. For invalid PTEs, if
this eld is not zero, it contains information about where the page is in the
swap le.
The following two bits are dened and used by Linux:
PAGE DIRTY if set, the page needs to be written out to the swap le,
PAGE ACCESSED Used by Linux to mark a page as having been accessed.
3.2 Caches
If you were to implement a system using the above theoretical model then it would
work, but not particularly eciently. Both operating system and processor designers
try hard to extract more performance from the system. Apart from making the
processors, memory and so on faster the best approach is to maintain caches of
useful information and data that make some operations faster. Linux uses a number
of memory management related caches:
Buer Cache The buer cache contains data buers that are used by the block
device drivers. These buers are of xed sizes (for example 512 bytes) and See
fs/buffer.c
contain blocks of information that have either been read from a block device
or are being written to it. A block device is one that can only be accessed by
reading and writing xed sized blocks of data. All hard disks are block devices.
The buer cache is indexed via the device identier and the desired block
number and is used to quickly nd a block of data. Block devices are only ever
accessed via the buer cache. If data can be found in the buer cache then it
does not need to be read from the physical block device, for example a hard
disk, and access to it is much faster.
Page Cache This is used to speed up access to images and data on disk. It is used See
mm/filemap.c
to cache the logical contents of a le a page at a time and is accessed via the
le and oset within the le. As pages are read into memory from disk, they
are cached in the page cache.
22 Chapter 3. Memory Management
Swap Cache Only modied (or dirty) pages are saved in the swap See swap.h, le. So long
mm/swap state.c
mm/swapfile.c
as these pages are not modied after they have been written to the swap le
then the next time the page is swapped out there is no need to write it to the
swap le as the page is already in the swap le. Instead the page can simply
be discarded. In a heavily swapping system this saves many unnecessary and
costly disk operations.
Hardware Caches One commonly implemented hardware cache is in the processor;
a cache of Page Table Entries. In this case, the processor does not always read
the page table directly but instead caches translations for pages as it needs
them. These are the Translation Look-aside Buers and contain cached copies
of the page table entries from one or more processes in the system.
When the reference to the virtual address is made, the processor will attempt to
nd a matching TLB entry. If it nds one, it can directly translate the virtual
address into a physical one and perform the correct operation on the data. If
the processor cannot nd a matching TLB entry then it must get the operating
system to help. It does this by signalling the operating system that a TLB miss
has occurred. A system specic mechanism is used to deliver that exception
to the operating system code that can x things up. The operating system
generates a new TLB entry for the address mapping. When the exception has
been cleared, the processor will make another attempt to translate the virtual
address. This time it will work because there is now a valid entry in the TLB
for that address.
The drawback of using caches, hardware or otherwise, is that in order to save eort
Linux must use more time and space maintaining these caches and, if the caches
become corrupted, the system will crash.
3.3 Linux Page Tables
Linux assumes that there are three levels of page tables. Each Page Table accessed
contains the page frame number of the next level of Page Table. Figure 3.3 shows
how a virtual address can be broken into a number of elds; each eld providing an
oset into a particular Page Table. To translate a virtual address into a physical
one, the processor must take the contents of each level eld, convert it into an oset
into the physical page containing the Page Table and read the page frame number
of the next level of Page Table. This is repeated three times until the page frame
number of the physical page containing the virtual address is found. Now the nal
eld in the virtual address, the byte oset, is used to nd the data inside the page.
Each platform that Linux runs on must provide translation macros that allow the
kernel to traverse the page tables for a particular process. This way, the kernel does
not need to know the format of the page table entries or how they are arranged. This
See
include/asm/-
pgtable.h is so successful that Linux uses the same page table manipulation code for the Alpha
processor, which has three levels of page tables, and for Intel x86 processors, which
have two levels of page tables.
3.4. Page Allocation and Deallocation 23
Level 2
Page Table Page Table
Level 3
Level 1 Level 2 Level 3 Byte within page
PGD
VIRTUAL ADDRESS
Page Table
Level 1
PFN PFN PFN
Physical Page
Figure 3.3: Three Level Page Tables
3.4 Page Allocation and Deallocation
There are many demands on the physical pages in the system. For example, when
an image is loaded into memory the operating system needs to allocate pages. These
will be freed when the image has nished executing and is unloaded. Another use
for physical pages is to hold kernel specic data structures such as the page tables
themselves. The mechanisms and data structures used for page allocation and deallocation
are perhaps the most critical in maintaining the eciency of the virtual
memory subsystem.
All of the physical pages in the system are described by the mem map data structure
which is a list of mem map t 1 structures which is initialized at boot time. Each
See
include/linux/-
mem map t describes a single physical page in the system. Important elds (so far as mm.h
memory management is concerned) are:
count This is a count of the number of users of this page. The count is greater than
one when the page is shared between many processes,
age This eld describes the age of the page and is used to decide if the page is a
good candidate for discarding or swapping,
map nr This is the physical page frame number that this mem map t describes.
The free area vector is used by the page allocation code to nd and free pages.
The whole buer management scheme is supported by this mechanism and so far as
the code is concerned, the size of the page and physical paging mechanisms used by
the processor are irrelevant.
Each element of free area contains information about blocks of pages. The rst
element in the array describes single pages, the next blocks of 2 pages, the next
blocks of 4 pages and so on upwards in powers of two. The list element is used as a
queue head and has pointers to the page data structures in the mem map array. Free
1Confusingly the structure is also known as the page structure.
24 Chapter 3. Memory Management
blocks of pages are queued here. map is a pointer to a bitmap which keeps track of
allocated groups of pages of this size. Bit N of the bitmap is set if the Nth block of
pages is free.
Figure 3.4 shows the free area structure. Element 0 has one free page (page frame
number 0) and element 2 has 2 free blocks of 4 pages, the rst starting at page frame
number 4 and the second at page frame number 56.
3.4.1 Page Allocation
Linux uses the Buddy algorithm 2 to eectively allocate and deallocate blocks of
pages. The page allocation code attempts to allocate a block of one or more physical
See
get free pages()
in
mm/page alloc.c pages. Pages are allocated in blocks which are powers of 2 in size. That means that
it can allocate a block 1 page, 2 pages, 4 pages and so on. So long as there are enough
free pages in the system to grant this request (nr free pages > min free pages) the
allocation code will search the free area for a block of pages of the size requested.
Each element of the free area has a map of the allocated and free blocks of pages
for that sized block. For example, element 2 of the array has a memory map that
describes free and allocated blocks each of 4 pages long.
The allocation algorithm rst searches for blocks of pages of the size requested. It
follows the chain of free pages that is queued on the list element of the free area
data structure. If no blocks of pages of the requested size are free, blocks of the next
size (which is twice that of the size requested) are looked for. This process continues
until all of the free area has been searched or until a block of pages has been found.
If the block of pages found is larger than that requested it must be broken down until
there is a block of the right size. Because the blocks are each a power of 2 pages big
then this breaking down process is easy as you simply break the blocks in half. The
free blocks are queued on the appropriate queue and the allocated block of pages is
returned to the caller.
For example, in Figure 3.4 if a block of 2 pages was requested, the rst block of 4
pages (starting at page frame number 4) would be broken into two 2 page blocks.
The rst, starting at page frame number 4 would be returned to the caller as the
allocated pages and the second block, starting at page frame number 6 would be
queued as a free block of 2 pages onto element 1 of the free area array.
3.4.2 Page Deallocation
Allocating blocks of pages tends to fragment memory with larger blocks of free pages
being broken down into smaller ones. The page deallocation code recombines pages
See
free pages() in
mm/page alloc.c into larger blocks of free pages whenever it can. In fact the page block size is
important as it allows for easy combination of blocks into larger blocks.
Whenever a block of pages is freed, the adjacent or buddy block of the same size is
checked to see if it is free. If it is, then it is combined with the newly freed block
of pages to form a new free block of pages for the next size block of pages. Each
time two blocks of pages are recombined into a bigger block of free pages the page
deallocation code attempts to recombine that block into a yet larger one. In this way
the blocks of free pages are as large as memory usage will allow.
2Bibliography reference here
3.5. Memory Mapping 25
1
2
3
4
5
6
7
8
0 PFN
PHYSICAL MEMORY
4
0
56
mem_map_t mem_map_t
mem_map_t
free_area
3
4
5
1
0
2
Free PFN
map
map
map
map
Figure 3.4: The free area data structure
For example, in Figure 3.4, if page frame number 1 were to be freed, then that would
be combined with the already free page frame number 0 and queued onto element 1
of the free area as a free block of size 2 pages.
3.5 Memory Mapping
When an image is executed, the contents of the executable image must be brought
into the processes virtual address space. The same is also true of any shared libraries
that the executable image has been linked to use. The executable le is not actually
brought into physical memory, instead it is merely linked into the processes virtual
memory. Then, as the parts of the program are referenced by the running application,
the image is brought into memory from the executable image. This linking of an
image into a processes virtual address space is known as memory mapping.
Every processes virtual memory is represented by an mm struct data structure. This
contains information about the image that it is currently executing (for example
bash) and also has pointers to a number of vm area struct data structures. Each
vm area struct data structure describes the start and end of the area of virtual
memory, the processes access rights to that memory and a set of operations for
that memory. These operations are a set of routines that Linux must use when
manipulating this area of virtual memory. For example, one of the virtual memory
operations performs the correct actions when the process has attempted to access
this virtual memory but nds (via a page fault) that the memory is not actually in
physical memory. This operation is the nopage operation. The nopage operation is
used when Linux demand pages the pages of an executable image into memory.
When an executable image is mapped into a processes virtual address a set of
vm area struct data structures is generated. Each vm area struct data structure
represents a part of the executable image; the executable code, initialized data
26 Chapter 3. Memory Management
vm_area_struct
vm_next
vm_flags
vm_inode
vm_ops
vm_end
vm_start
Virtual Area
Processes Virtual Memory
unmap()
protect()
sync()
advise()
swapin()
open()
close()
nopage()
wppage()
swapout()
Virtual Memory
Operations
Figure 3.5: Areas of Virtual Memory
(variables), unitialized data and so on. Linux supports a number of standard virtual
memory operations and as the vm area struct data structures are created, the
correct set of virtual memory operations are associated with them.
3.6 Demand Paging
Once an executable image has been memory mapped into a processes virtual memory
it can start to execute. As only the very start of the image is physically pulled into
memory it will soon access an area of virtual memory that is not yet in physical
memory. When a process accesses a virtual address that does not have a valid page
table entry, the processor will report a page fault to Linux. The page fault describes
See
handle mm fault()
in
mm/memory.c the virtual address where the page fault occurred and the type of memory access
that caused.
Linux must nd the vm area struct that represents the area of memory that the
page fault occurred in. As searching through the vm area struct data structures is
critical to the ecient handling of page faults, these are linked together in an AVL
(Adelson-Velskii and Landis) tree structure. If there is no vm area struct data
structure for this faulting virtual address, this process has accessed an illegal virtual
address. Linux will signal the process, sending a SIGSEGV signal, and if the process
does not have a handler for that signal it will be terminated.
Linux next checks the type of page fault that occurred against the types of accesses
allowed for this area of virtual memory. If the process is accessing the memory in
an illegal way, say writing to an area that it is only allowed to read from, it is also
3.7. The Linux Page Cache 27
offset
next_hash
prev_hash
inode
mem_map_t
offset
next_hash
prev_hash
inode
mem_map_t
12
0x2000
12
0x8000
:
:
:
page_hash_table
Figure 3.6: The Linux Page Cache
signalled with a memory error.
Now that Linux has determined that the page fault is legal, it must deal with it.
Linux must dierentiate between pages that are in the swap le and those that are
See
do no page() in
part of an executable image on a disk somewhere. It does this by using the page mm/memory.c
table entry for this faulting virtual address.
If the page's page table entry is invalid but not empty, the page fault is for a page
currently being held in the swap le. For Alpha AXP page table entries, these are
entries which do not have their valid bit set but which have a non-zero value in their
PFN eld. In this case the PFN eld holds information about where in the swap
(and which swap le) the page is being held. How pages in the swap le are handled
is described later in this chapter.
Not all vm area struct data structures have a set of virtual memory operations and
even those that do may not have a nopage operation. This is because by default
Linux will x up the access by allocating a new physical page and creating a valid
page table entry for it. If there is a nopage operation for this area of virtual memory,
Linux will use it.
The generic Linux nopage operation is used for memory mapped executable images
See
filemap nopage()
in
and it uses the page cache to bring the required image page into physical memory. mm/filemap.c
However the required page is brought into physical memory, the processes page tables
are updated. It may be necessary for hardware specic actions to update those
entries, particularly if the processor uses translation look aside buers. Now that
the page fault has been handled it can be dismissed and the process is restarted at
the instruction that made the faulting virtual memory access.
3.7 The Linux Page Cache
The role of the Linux page cache is to speed up access to les on disk. Memory
mapped les are read a page at a time and these pages are stored in the page cache.
Figure 3.6 shows that the page cache consists of the page hash table, a vector of
pointers to mem map t data structures. Each le in Linux is identied by a VFS
See
include/linux/-
pagemap.h
28 Chapter 3. Memory Management
inode data structure (described in Chapter 9) and each VFS inode is unique and
fully describes one and only one le. The index into the page table is derived from
the le's VFS inode and the oset into the le.
Whenever a page is read from a memory mapped le, for example when it needs
to be brought back into memory during demand paging, the page is read through
the page cache. If the page is present in the cache, a pointer to the mem map t data
structure representing it is returned to the page fault handling code. Otherwise the
page must be brought into memory from the le system that holds the image. Linux
allocates a physical page and reads the page from the le on disk.
If it is possible, Linux will initiate a read of the next page in the le. This single
page read ahead means that if the process is accessing the pages in the le serially,
the next page will be waiting in memory for the process.
Over time the page cache grows as images are read and executed. Pages will be
removed from the cache as they are no longer needed, say as an image is no longer
being used by any process. As Linux uses memory it can start to run low on physical
pages. In this case Linux will reduce the size of the page cache.
3.8 Swapping Out and Discarding Pages
When physical memory becomes scarce the Linux memory management subsystem
must attempt to free physical pages. This task falls to the kernel swap daemon
(kswapd). The kernel swap daemon is a special type of process, a kernel thread.
Kernel threads are processes have no virtual memory, instead they run in kernel
mode in the physical address space. The kernel swap daemon is slightly misnamed
in that it does more than merely swap pages out to the system's swap les. Its role
is make sure that there are enough free pages in the system to keep the memory
management system operating eciently.
The Kernel swap daemon (kswapd) is started by the kernel init process at startup
time and sits waiting for the kernel swap timer to periodically expire. See kswapd() in Every time
mm/vmscan.c
the timer expires, the swap daemon looks to see if the number of free pages in the
system is getting too low. It uses two variables, free pages high and free pages low to
decide if it should free some pages. So long as the number of free pages in the system
remains above free pages high, the kernel swap daemon does nothing; it sleeps again
until its timer next expires. For the purposes of this check the kernel swap daemon
takes into account the number of pages currently being written out to the swap le.
It keeps a count of these in nr async pages; this is incremented each time a page is
queued waiting to be written out to the swap le and decremented when the write to
the swap device has completed. free pages low and free pages high are set at system
startup time and are related to the number of physical pages in the system. If the
number of free pages in the system has fallen below free pages high or worse still
free pages low, the kernel swap daemon will try three ways to reduce the number of
physical pages being used by the system:
Reducing the size of the buer and page caches,
Swapping out System V shared memory pages,
Swapping out and discarding pages.
3.8. Swapping Out and Discarding Pages 29
If the number of free pages in the system has fallen below free pages low, the kernel
swap daemon will try to free 6 pages before it next runs. Otherwise it will try to
free 3 pages. Each of the above methods are tried in turn until enough pages have
been freed. The kernel swap daemon remembers which method it was using the last
time that it attempted to free physical pages. Each time it runs it will start trying
to free pages using this last successful method.
After it has free sucient pages, the swap daemon sleeps again until its timer expires.
If the reason that the kernel swap daemon freed pages was that the number of free
pages in the system had fallen below free pages low, it only sleeps for half its usual
time. Once the number of free pages is more than free pages low the kernel swap
daemon goes back to sleeping longer between checks.
3.8.1 Reducing the Size of the Page and Buer Caches
The pages held in the page and buer caches are good candidates for being freed into
the free area vector. The Page Cache, which contains pages of memory mapped
les, may contain unneccessary pages that are lling up the system's memory. Likewise
the Buer Cache, which contains buers read from or being written to physical
devices, may also contain unneeded buers. When the physical pages in the system
start to run out, discarding pages from these caches is relatively easy as it requires
no writing to physical devices (unlike swapping pages out of memory). Discarding
these pages does not have too many harmful side eects other than making access
to physical devices and memory mapped les slower. However, if the discarding of
pages from these caches is done fairly, all processes will suer equally.
Every time the Kernel swap daemon tries to shrink these caches it examines a block
See
shrink mmap() in
of pages in the mem map page vector to see if any can be discarded from physical mm/filemap.c
memory. The size of the block of pages examined is higher if the kernel swap daemon
is intensively swapping; that is if the number of free pages in the system has fallen
dangerously low. The blocks of pages are examined in a cyclical manner; a dierent
block of pages is examined each time an attempt is made to shrink the memory map.
This is known as the clock algorithm as, rather like the minute hand of a clock, the
whole mem map page vector is examined a few pages at a time.
Each page being examined is checked to see if it is cached in either the page cache or
the buer cache. You should note that shared pages are not considered for discarding
at this time and that a page cannot be in both caches at the same time. If the page
is not in either cache then the next page in the mem map page vector is examined.
Pages are cached in the buer cache (or rather the buers within the pages are
cached) to make buer allocation and deallocation more ecient. The memory map
shrinking code tries to free the buers that are contained within the page being
examined. If all the buers are freed, then the pages that contain them are also be
See try to
free buffer() in
freed. If the examined page is in the Linux page cache, it is removed from the page fs/buffer.c
cache and freed.
When enough pages have been freed on this attempt then the kernel swap daemon
will wait until the next time it is periodically woken. As none of the freed pages
were part of any process's virtual memory (they were cached pages), then no page
tables need updating. If there were not enough cached pages discarded then the swap
daemon will try to swap out some shared pages.
30 Chapter 3. Memory Management
3.8.2 Swapping Out System V Shared Memory Pages
System V shared memory is an inter-process communication mechanism which allows
two or more processes to share virtual memory in order to pass information
amongst themselves. How processes share memory in this way is described in
more detail in Chapter 5. For now it is enough to say that each area of System
V shared memory is described by a shmid ds data structure. This contains
a pointer to a list of vm area struct data structures, one for each process sharing
this area of virtual memory. The vm area struct data structures describe where
in each processes virtual memory this area of System V shared memory goes. Each
vm area struct data structure for this System V shared memory is linked together
using the vm next shared and vm prev shared pointers. Each shmid ds data structure
also contains a list of page table entries each of which describes the physical
page that a shared virtual page maps to.
The kernel swap daemon also uses a clock algorithm when swapping out System V
shared memory pages. . Each time it runs it remembers which page of which shared See shm swap()
in ipc/shm.c
virtual memory area it last swapped out. It does this by keeping two indices, the
rst is an index into the set of shmid ds data structures, the second into the list of
page table entries for this area of System V shared memory. This makes sure that it
fairly victimizes the areas of System V shared memory.
As the physical page frame number for a given virtual page of System V shared
memory is contained in the page tables of all of the processes sharing this area of
virtual memory, the kernel swap daemon must modify all of these page tables to
show that the page is no longer in memory but is now held in the swap le. For
each shared page it is swapping out, the kernel swap daemon nds the page table
entry in each of the sharing processes page tables (by following a pointer from each
vm area struct data structure). If this processes page table entry for this page of
System V shared memory is valid, it converts it into an invalid but swapped out
page table entry and reduces this (shared) page's count of users by one. The format
of a swapped out System V shared page table entry contains an index into the set
of shmid ds data structures and an index into the page table entries for this area of
System V shared memory.
If the page's count is zero after the page tables of the sharing processes have all been
modied, the shared page can be written out to the swap le. The page table entry
in the list pointed at by the shmid ds data structure for this area of System V shared
memory is replaced by a swapped out page table entry. A swapped out page table
entry is invalid but contains an index into the set of open swap les and the oset
in that le where the swapped out page can be found. This information will be used
when the page has to be brought back into physical memory.
3.8.3 Swapping Out and Discarding Pages
The swap daemon looks at each process in the system in turn to see if it is a good
candidate for swapping. Good candidates are processes that can be swapped (some
See swap out()
in
mm/vmscan.c cannot) and that have one or more pages which can be swapped or discarded from
memory. Pages are swapped out of physical memory into the system's swap les only
if the data in them cannot be retrieved another way.
A lot of the contents of an executable image come from the image's le and can easily
be re-read from that le. For example, the executable instructions of an image will
3.9. The Swap Cache 31
never be modied by the image and so will never be written to the swap le. These
pages can simply be discarded; when they are again referenced by the process, they
will be brought back into memory from the executable image.
Once the process to swap has been located, the swap daemon looks through all of its
virtual memory regions looking for areas which are not shared or locked. Linux does
To do this it
follows the
vm next pointer
along the list of
vm area struct
structures
queued on the
mm struct for the
process.
not swap out all of the swappable pages of the process that it has selected; instead
it removes only a small number of pages. Pages cannot be swapped or discarded if
See
swap out vma()
in
mm/vmscan.c
they are locked in memory.
The Linux swap algorithm uses page aging. Each page has a counter (held in the
mem map t data structure) that gives the Kernel swap daemon some idea whether or
not a page is worth swapping. Pages age when they are unused and rejuvinate on
access; the swap daemon only swaps out old pages. The default action when a page
is rst allocated, is to give it an initial age of 3. Each time it is touched, it's age is
increased by 3 to a maximum of 20. Every time the Kernel swap daemon runs it ages
pages, decrementing their age by 1. These default actions can be changed and for
this reason they (and other swap related information) are stored in the swap control
data structure.
If the page is old (age = 0), the swap daemon will process it further. Dirty pages are
pages which can be swapped out. Linux uses an architecture specic bit in the PTE to
describe pages this way (see Figure 3.2). However, not all dirty pages are necessarily
written to the swap le. Every virtual memory region of a process may have its own
swap operation (pointed at by the vm ops pointer in the vm area struct) and that
method is used. Otherwise, the swap daemon will allocate a page in the swap le
and write the page out to that device.
The page's page table entry is replaced by one which is marked as invalid but which
contains information about where the page is in the swap le. This is an oset into
the swap le where the page is held and an indication of which swap le is being used.
Whatever the swap method used, the original physical page is made free by putting
it back into the free area. Clean (or rather not dirty) pages can be discarded and
put back into the free area for re-use.
If enough of the swappable processes pages have been swapped out or discarded,
the swap daemon will again sleep. The next time it wakes it will consider the next
process in the system. In this way, the swap daemon nibbles away at each processes
physical pages until the system is again in balance. This is much fairer than swapping
out whole processes.
3.9 The Swap Cache
When swapping pages out to the swap les, Linux avoids writing pages if it does not
have to. There are times when a page is both in a swap le and in physical memory.
This happens when a page that was swapped out of memory was then brought back
into memory when it was again accessed by a process. So long as the page in memory
is not written to, the copy in the swap le remains valid.
Linux uses the swap cache to track these pages. The swap cache is a list of page
table entries, one per physical page in the system. This is a page table entry for a
swapped out page and describes which swap le the page is being held in together
with its location in the swap le. If a swap cache entry is non-zero, it represents
32 Chapter 3. Memory Management
a page which is being held in a swap le that has not been modied. If the page
is subsequently modied (by being written to), its entry is removed from the swap
cache.
When Linux needs to swap a physical page out to a swap le it consults the swap
cache and, if there is a valid entry for this page, it does not need to write the page
out to the swap le. This is because the page in memory has not been modied since
it was last read from the swap le.
The entries in the swap cache are page table entries for swapped out pages. They
are marked as invalid but contain information which allow Linux to nd the right
swap le and the right page within that swap le.
3.10 Swapping Pages In
The dirty pages saved in the swap les may be needed again, for example when
an application writes to an area of virtual memory whose contents are held in a
swapped out physical page. Accessing a page of virtual memory that is not held
in physical memory causes a page fault to occur. The page fault is the processor
signalling the operating system that it cannot translate a virtual address into a
physical one. In this case this is because the page table entry describing this page
of virtual memory was marked as invalid when the page was swapped out. The
processor cannot handle the virtual to physical address translation and so hands
control back to the operating system describing as it does so the virtual address that
faulted and the reason for the fault. The format of this information and how the
processor passes control to the operating system is processor specic. The processor
See
do page fault()
in
arch/i386/-
mm/fault.c
specic page fault handling code must locate the vm area struct data structure that
describes the area of virtual memory that contains the faulting virtual address. It
does this by searching the vm area struct data structures for this process until it
nds the one containing the faulting virtual address. This is very time critical code
and a processes vm area struct data structures are so arranged as to make this
search take as little time as possible.
Having carried out the appropriate processor specic actions and found that the
faulting virtual address is for a valid area of virtual memory, the page fault processing
becomes generic and applicable to all processors that Linux runs on. The generic
See
do no page() in
mm/memory.c page fault handling code looks for the page table entry for the faulting virtual address.
If the page table entry it nds is for a swapped out page, Linux must swap the page
back into physical memory. The format of the page table entry for a swapped out
page is processor specic but all processors mark these pages as invalid and put the
information neccessary to locate the page within the swap le into the page table
entry. Linux needs this information in order to bring the page back into physical
memory.
At this point, Linux knows the faulting virtual address and has a page table
See
do swap page()
in
mm/memory.c entry containing information about where this page has been swapped to. The
vm area struct data structure may contain a pointer to a routine which will swap
any page of the area of virtual memory that it describes back into physical memory.
This is its swapin operation. If there is a swapin operation for this area of virtual
memory then Linux will use it. This is, in fact, how swapped out System V shared
See
shm swap in() in
ipc/shm.c memory pages are handled as it requires special handling because the format of a
swapped out System V shared page is a little dierent from that of an ordinairy
3.10. Swapping Pages In 33
swapped out page. There may not be a swapin operation, in which case Linux will
assume that this is an ordinairy page that does not need to be specially handled. It See swap in() in
mm/page alloc.c
allocates a free physical page and reads the swapped out page back from the swap
le. Information telling it where in the swap le (and which swap le) is taken from
the the invalid page table entry.
If the access that caused the page fault was not a write access then the page is left
in the swap cache and its page table entry is not marked as writable. If the page is
subsequently written to, another page fault will occur and, at that point, the page
is marked as dirty and its entry is removed from the swap cache. If the page is not
written to and it needs to be swapped out again, Linux can avoid the write of the
page to its swap le because the page is already in the swap le.
If the access that caused the page to be brought in from the swap le was a write
operation, this page is removed from the swap cache and its page table entry is
marked as both dirty and writable.
34 Chapter 3. Memory Management
Chapter 4
Processes
This chapter describes what a process is and how the Linux kernel creates,
manages and deletes the processes in the system.
Processes carry out tasks within the operating system. A program is a set of machine
code instructions and data stored in an executable image on disk and is, as such, a
passive entity; a process can be thought of as a computer program in action. It is a
dynamic entity, constantly changing as the machine code instructions are executed
by the processor. As well as the program's instructions and data, the process also
includes the program counter and all of the CPU's registers as well as the process
stacks containing temporary data such as routine parameters, return addresses and
saved variables. The current executing program, or process, includes all of the current
activity in the microprocessor. Linux is a multiprocessing operating system.
Processes are separate tasks each with their own rights and responsibilities. If one
process crashes it will not cause another process in the system to crash. Each individual
process runs in its own virtual address space and is not capable of interacting
with another process except through secure, kernel managed mechanisms.
During the lifetime of a process it will use many system resources. It will use the
CPUs in the system to run its instructions and the system's physical memory to hold
it and its data. It will open and use les within the lesystems and may directly
or indirectly use the physical devices in the system. Linux must keep track of the
process itself and of the system resources that it has so that it can manage it and
the other processes in the system fairly. It would not be fair to the other processes
in the system if one process monopolized most of the system's physical memory or
its CPUs.
The most precious resource in the system is the CPU, usually there is only one. Linux
35
36 Chapter 4. Processes
is a multiprocessing operating system, its objective is to have a process running on
each CPU in the system at all times, to maximize CPU utilization. If there are more
processes than CPUs (and there usually are), the rest of the processes must wait
before a CPU becomes free until they can be run. Multiprocessing is a simple idea; a
process is executed until it must wait, usually for some system resource; when it has
this resource, it may run again. In a uniprocessing system, for example DOS, the CPU
would simply sit idle and the waiting time would be wasted. In a multiprocessing
system many processes are kept in memory at the same time. Whenever a process
has to wait the operating system takes the CPU away from that process and gives
it to another, more deserving process. It is the scheduler which chooses which is
the most appropriate process to run next and Linux uses a number of scheduling
strategies to ensure fairness.
Linux supports a number of dierent executable le formats, ELF is one, Java is
another and these must be managed transparently as must the processes use of the
system's shared libraries.
4.1 Linux Processes
So that Linux can manage the processes in the system, each process is represented
by a task struct data structure (task and process are terms that Linux uses interchangeably).
The task vector is an array of pointers to every task struct data
structure in the system. This means that the maximum number of processes in the
See
include/linux/-
sched.h system is limited by the size of the task vector; by default it has 512 entries. As processes
are created, a new task struct is allocated from system memory and added
into the task vector. To make it easy to nd, the current, running, process is pointed
to by the current pointer.
As well as the normal type of process, Linux supports real time processes. These
processes have to react very quickly to external events (hence the term \real time")
and they are treated dierently from normal user processes by the scheduler. Although
the task struct data structure is quite large and complex, but its elds can
be divided into a number of functional areas:
State As a process executes it changes state according to its circumstances. Linux
processes have the following states: 1
Running The process is either running (it is the current process in the system)
or it is ready to run (it is waiting to be assigned to one of the system's
CPUs).
Waiting The process is waiting for an event or for a resource. Linux dierentiates
between two types of waiting process; interruptible and uninterruptible.
Interruptible waiting processes can be interrupted by signals whereas
uninterruptible waiting processes are waiting directly on hardware conditions
and cannot be interrupted under any circumstances.
Stopped The process has been stopped, usually by receiving a signal. A
process that is being debugged can be in a stopped state.
Zombie This is a halted process which, for some reason, still has a task struct
data structure in the task vector. It is what it sounds like, a dead process.
1REVIEW NOTE: I left out SWAPPING because it does not appear to be used.
4.1. Linux Processes 37
Scheduling Information The scheduler needs this information in order to fairly
decide which process in the system most deserves to run,
Identiers Every process in the system has a process identier. The process identi
er is not an index into the task vector, it is simply a number. Each process
also has User and group identiers, these are used to control this processes
access to the les and devices in the system,
Inter-Process Communication Linux supports the classic UnixTM IPC mechanisms
of signals, pipes and semaphores and also the System V IPC mechanisms
of shared memory, semaphores and message queues. The IPC mechanisms supported
by Linux are described in Chapter 5.
Links In a Linux system no process is independent of any other process. Every
process in the system, except the initial process has a parent process. New
processes are not created, they are copied, or rather cloned from previous processes.
Every task struct representing a process keeps pointers to its parent
process and to its siblings (those processes with the same parent process) as
well as to its own child processes. You can see the family relationship between
the running processes in a Linux system using the pstree command:
init(1)-+-crond(98)
|-emacs(387)
|-gpm(146)
|-inetd(110)
|-kerneld(18)
|-kflushd(2)
|-klogd(87)
|-kswapd(3)
|-login(160)---bash(192)---emacs(225)
|-lpd(121)
|-mingetty(161)
|-mingetty(162)
|-mingetty(163)
|-mingetty(164)
|-login(403)---bash(404)---pstree(594)
|-sendmail(134)
|-syslogd(78)
`-update(166)
Additionally all of the processes in the system are held in a doubly linked list
whose root is the init processes task struct data structure. This list allows
the Linux kernel to look at every process in the system. It needs to do this to
provide support for commands such as ps or kill.
Times and Timers The kernel keeps track of a processes creation time as well as
the CPU time that it consumes during its lifetime. Each clock tick, the kernel
updates the amount of time in jiffies that the current process has spent in
system and in user mode. Linux also supports process specic interval timers,
38 Chapter 4. Processes
processes can use system calls to set up timers to send signals to themselves
when the timers expire. These timers can be single-shot or periodic timers.
File system Processes can open and close les as they wish and the processes
task struct contains pointers to descriptors for each open le as well as pointers
to two VFS inodes. Each VFS inode uniquely describes a le or directory
within a le system and also provides a uniform interface to the underlying
le systems. How le systems are supported under Linux is described in Chapter
9. The rst is to the root of the process (its home directory) and the second
is to its current or pwd directory. pwd is derived from the UnixTM command
pwd, print working directory. These two VFS inodes have their count elds
incremented to show that one or more processes are referencing them. This is
why you cannot delete the directory that a process has as its pwd directory set
to, or for that matter one of its sub-directories.
Virtual memory Most processes have some virtual memory (kernel threads and
daemons do not) and the Linux kernel must track how that virtual memory is
mapped onto the system's physical memory.
Processor Specic Context A process could be thought of as the sum total of the
system's current state. Whenever a process is running it is using the processor's
registers, stacks and so on. This is the processes context and, when a process is
suspended, all of that CPU specic context must be saved in the task struct
for the process. When a process is restarted by the scheduler its context is
restored from here.
4.2 Identiers
Linux, like all UnixTM uses user and group identiers to check for access rights to
les and images in the system. All of the les in a Linux system have ownerships
and permissions, these permissions describe what access the system's users have to
that le or directory. Basic permissions are read, write and execute and are assigned
to three classes of user; the owner of the le, processes belonging to a particular
group and all of the processes in the system. Each class of user can have dierent
permissions, for example a le could have permissions which allow its owner to read
and write it, the le's group to read it and for all other processes in the system to
have no access at all.
REVIEW NOTE: Expand and give the bit assignments (777).
Groups are Linux's way of assigning privileges to les and directories for a group
of users rather than to a single user or to all processes in the system. You might,
for example, create a group for all of the users in a software project and arrange it
so that only they could read and write the source code for the project. A process
can belong to several groups (a maximum of 32 is the default) and these are held in
the groups vector in the task struct for each process. So long as a le has access
rights for one of the groups that a process belongs to then that process will have
appropriate group access rights to that le.
There are four pairs of process and group identiers held in a processes task struct:
uid, gid The user identier and group identier of the user that the process is
running on behalf of,
4.3. Scheduling 39
eective uid and gid There are some programs which change the uid and gid from
that of the executing process into their own (held as attributes in the VFS
inode describing the executable image). These programs are known as setuid
programs and they are useful because it is a way of restricting accesses to
services, particularly those that run on behalf of someone else, for example a
network daemon. The eective uid and gid are those from the setuid program
and the uid and gid remain as they were. The kernel checks the eective uid
and gid whenever it checks for privilege rights.
le system uid and gid These are normally the same as the eective uid and gid
and are used when checking le system access rights. They are needed for NFS
mounted lesystems where the user mode NFS server needs to access les as if
it were a particular process. In this case only the le system uid and gid are
changed (not the eective uid and gid). This avoids a situation where malicious
users could send a kill signal to the NFS server. Kill signals are delivered to
processes with a particular eective uid and gid.
saved uid and gid These are mandated by the POSIX standard and are used by
programs which change the processes uid and gid via system calls. They are
used to save the real uid and gid during the time that the original uid and gid
have been changed.
4.3 Scheduling
All processes run partially in user mode and partially in system mode. How these
modes are supported by the underlying hardware diers but generally there is a
secure mechanism for getting from user mode into system mode and back again.
User mode has far less privileges than system mode. Each time a process makes a
system call it swaps from user mode to system mode and continues executing. At
this point the kernel is executing on behalf of the process. In Linux, processes do
not preempt the current, running process, they cannot stop it from running so that
they can run. Each process decides to relinquish the CPU that it is running on when
it has to wait for some system event. For example, a process may have to wait for
a character to be read from a le. This waiting happens within the system call, in
system mode; the process used a library function to open and read the le and it,
in turn made system calls to read bytes from the open le. In this case the waiting
process will be suspended and another, more deserving process will be chosen to run.
Processes are always making system calls and so may often need to wait. Even so, if
a process executes until it waits then it still might use a disproportionate amount of
CPU time and so Linux uses pre-emptive scheduling. In this scheme, each process is
allowed to run for a small amount of time, 200ms, and, when this time has expired
another process is selected to run and the original process is made to wait for a little
while until it can run again. This small amount of time is known as a time-slice.
It is the scheduler that must select the most deserving process to run out of all of
the runnable processes in the system. A runnable process is one which is waiting
See schedule()
in
only for a CPU to run on. Linux uses a reasonably simple priority based scheduling kernel/sched.c
algorithm to choose between the current processes in the system. When it has chosen
a new process to run it saves the state of the current process, the processor specic
registers and other context being saved in the processes task struct data structure.
40 Chapter 4. Processes
It then restores the state of the new process (again this is processor specic) to run
and gives control of the system to that process. For the scheduler to fairly allocate
CPU time between the runnable processes in the system it keeps information in the
task struct for each process:
policy This is the scheduling policy that will be applied to this process. There are
two types of Linux process, normal and real time. Real time processes have a
higher priority than all of the other processes. If there is a real time process
ready to run, it will always run rst. Real time processes may have two types
of policy, round robin and rst in rst out. In round robin scheduling, each
runnable real time process is run in turn and in rst in, rst out scheduling
each runnable process is run in the order that it is in on the run queue and
that order is never changed.
priority This is the priority that the scheduler will give to this process. It is also the
amount of time (in jiffies) that this process will run for when it is allowed
to run. You can alter the priority of a process by means of system calls and
the renice command.
rt priority Linux supports real time processes and these are scheduled to have a
higher priority than all of the other non-real time processes in system. This
eld allows the scheduler to give each real time process a relative priority. The
priority of a real time processes can be altered using system calls.
counter This is the amount of time (in jiffies) that this process is allowed to run
for. It is set to priority when the process is rst run and is decremented each
clock tick.
The scheduler is run from several places within the kernel. It is run after putting the
current process onto a wait queue and it may also be run at the end of a system call,
just before a process is returned to process mode from system mode. One reason that
it might need to run is because the system timer has just set the current processes
counter to zero. Each time the scheduler is run it does the following:
See schedule()
in
kernel/sched.c
kernel work The scheduler runs the bottom half handlers and processes the scheduler
task queue. These lightweight kernel threads are described in detail in
chapter 11.
Current process The current process must be processed before another process
can be selected to run.
If the scheduling policy of the current processes is round robin then it is put
onto the back of the run queue.
If the task is INTERRUPTIBLE and it has received a signal since the last time it
was scheduled then its state becomes RUNNING.
If the current process has timed out, then its state becomes RUNNING.
If the current process is RUNNING then it will remain in that state.
Processes that were neither RUNNING nor INTERRUPTIBLE are removed from the
run queue. This means that they will not be considered for running when the
scheduler looks for the most deserving process to run.
4.3. Scheduling 41
Process selection The scheduler looks through the processes on the run queue
looking for the most deserving process to run. If there are any real time processes
(those with a real time scheduling policy) then those will get a higher
weighting than ordinary processes. The weight for a normal process is its
counter but for a real time process it is counter plus 1000. This means that if
there are any runnable real time processes in the system then these will always
be run before any normal runnable processes. The current process, which has
consumed some of its time-slice (its counter has been decremented) is at a disadvantage
if there are other processes with equal priority in the system; that
is as it should be. If several processes have the same priority, the one nearest
the front of the run queue is chosen. The current process will get put onto the
back of the run queue. In a balanced system with many processes of the same
priority, each one will run in turn. This is known as Round Robin scheduling.
However, as processes wait for resources, their run order tends to get moved
around.
Swap processes If the most deserving process to run is not the current process,
then the current process must be suspended and the new one made to run.
When a process is running it is using the registers and physical memory of the
CPU and of the system. Each time it calls a routine it passes its arguments
in registers and may stack saved values such as the address to return to in the
calling routine. So, when the scheduler is running it is running in the context
of the current process. It will be in a privileged mode, kernel mode, but it
is still the current process that is running. When that process comes to be
suspended, all of its machine state, including the program counter (PC) and
all of the processor's registers, must be saved in the processes task struct
data structure. Then, all of the machine state for the new process must be
loaded. This is a system dependent operation, no CPUs do this in quite the
same way but there is usually some hardware assistance for this act.
This swapping of process context takes place at the end of the scheduler. The
saved context for the previous process is, therefore, a snapshot of the hardware
context of the system as it was for this process at the end of the scheduler.
Equally, when the context of the new process is loaded, it too will be a snapshot
of the way things were at the end of the scheduler, including this processes
program counter and register contents.
If the previous process or the new current process uses virtual memory then
the system's page table entries may need to be updated. Again, this action
is architecture specic. Processors like the Alpha AXP, which use Translation
Look-aside Tables or cached Page Table Entries, must
ush those cached table
entries that belonged to the previous process.
4.3.1 Scheduling in Multiprocessor Systems
Systems with multiple CPUs are reasonably rare in the Linux world but a lot of
work has already gone into making Linux an SMP (Symmetric Multi-Processing)
operating system. That is, one that is capable of evenly balancing work between the
CPUs in the system. Nowhere is this balancing of work more apparent than in the
scheduler.
In a multiprocessor system, hopefully, all of the processors are busily running pro42
Chapter 4. Processes
fs_struct
count
umask
*root
*pwd
inode
inode
file
f_mode
f_pos
f_flags
f_count
f_owner
f_inode
f_op
f_version
inode
file operation
routines
fs
files
task_struct
files_struct
count
close_on_exec
open_fs
fd[0]
fd[1]
fd[255]
0x022
Figure 4.1: A Process's Files
cesses. Each will run the scheduler separately as its current process exhausts its
time-slice or has to wait for a system resource. The rst thing to notice about an
SMP system is that there is not just one idle process in the system. In a single
processor system the idle process is the rst task in the task vector, in an SMP
system there is one idle process per CPU, and you could have more than one idle
CPU. Additionally there is one current process per CPU, so SMP systems must keep
track of the current and idle processes for each processor.
In an SMP system each process's task struct contains the number of the processor
that it is currently running on (processor) and its processor number of the last
processor that it ran on (last processor). There is no reason why a process should
not run on a dierent CPU each time it is selected to run but Linux can restrict a
process to one or more processors in the system using the processor mask. If bit N
is set, then this process can run on processor N. When the scheduler is choosing a
new process to run it will not consider one that does not have the appropriate bit set
for the current processor's number in its processor mask. The scheduler also gives
a slight advantage to a process that last ran on the current processor because there
is often a performance overhead when moving a process to a dierent processor.
4.4 Files
Figure 4.1 shows that there are two data structures that describe le system specic
information for each process in the system. The rst, the fs struct contains pointers
See
include/linux/-
sched.h to this process's VFS inodes and its umask. The umask is the default mode that new
4.5. Virtual Memory 43
les will be created in, and it can be changed via system calls.
The second data structure, the files struct, contains information about all of the
les that this process is currently using. Programs read from standard input and write
to standard output. Any error messages should go to standard error. These may be
les, terminal input/output or a real device but so far as the program is concerned
they are all treated as les. Every le has its own descriptor and the files struct
contains pointers to up to 256 file data structures, each one describing a le being
used by this process. The f mode eld describes what mode the le has been created
in; read only, read and write or write only. f pos holds the position in the le
where the next read or write operation will occur. f inode points at the VFS inode
describing the le and f ops is a pointer to a vector of routine addresses; one for
each function that you might wish to perform on a le. There is, for example, a write
data function. This abstraction of the interface is very powerful and allows Linux
to support a wide variety of le types. In Linux, pipes are implemented using this
mechanism as we shall see later.
Every time a le is opened, one of the free file pointers in the files struct is used
to point to the new file structure. Linux processes expect three le descriptors to
be open when they start. These are known as standard input, standard output and
standard error and they are usually inherited from the creating parent process. All
accesses to les are via standard system calls which pass or return le descriptors.
These descriptors are indices into the process's fd vector, so standard input, standard
output and standard error have le descriptors 0, 1 and 2. Each access to the le
uses the file data structure's le operation routines to together with the VFS inode
to achieve its needs.
4.5 Virtual Memory
A process's virtual memory contains executable code and data from many sources.
First there is the program image that is loaded; for example a command like ls.
This command, like all executable images, is composed of both executable code and
data. The image le contains all of the information neccessary to load the executable
code and associated program data into the virtual memory of the process. Secondly,
processses can allocate (virtual) memory to use during their processing, say to hold
the contents of les that it is reading. This newly allocated, virtual, memory needs to
be linked into the process's existing virtual memory so that it can be used. Thirdly,
Linux processes use libraries of commonly useful code, for example le handling
routines. It does not make sense that each process has its own copy of the library,
Linux uses shared libraries that can be used by several running processes at the same
time. The code and the data from these shared libraries must be linked into this
process's virtual address space and also into the virtual address space of the other
processes sharing the library.
In any given time period a process will not have used all of the code and data
contained within its virtual memory. It could contain code that is only used during
certain situations, such as during initialization or to process a particular event. It
may only have used some of the routines from its shared libraries. It would be
wasteful to load all of this code and data into physical memory where it would lie
unused. Multiply this wastage by the number of processes in the system and the
system would run very ineciently. Instead, Linux uses a technique called demand
44 Chapter 4. Processes
count
mmap_avl
mm_struct
mmap_sem
mmap
pgd
vm_area_struct
vm_next
vm_flags
vm_inode
vm_ops
vm_end
vm_start
task_struct
mm
Processes Virtual Memory
vm_area_struct
vm_next
vm_flags
vm_inode
vm_ops
vm_end
vm_start
0x8059BB8
0x8048000
0x0000000
Data
Code
Figure 4.2: A Process's Virtual Memory
paging where the virtual memory of a process is brought into physical memory only
when a process attempts to use it. So, instead of loading the code and data into
physical memory straight away, the Linux kernel alters the process's page table,
marking the virtual areas as existing but not in memory. When the process attempts
to acccess the code or data the system hardware will generate a page fault and
hand control to the Linux kernel to x things up. Therefore, for every area of
virtual memory in the process's address space Linux needs to know where that virtual
memory comes from and how to get it into memory so that it can x up these page
faults.
The Linux kernel needs to manage all of these areas of virtual memory and the contents
of each process's virtual memory is described by a mm struct data structure
pointed at from its task struct. The process's mm struct data structure also contains
information about the loaded executable image and a pointer to the process's
page tables. It contains pointers to a list of vm area struct data structures, each
representing an area of virtual memory within this process.
This linked list is in ascending virtual memory order, gure 4.2 shows the layout in
virtual memory of a simple process together with the kernel data structures managing
it. As those areas of virtual memory are from several sources, Linux abstracts the
interface by having the vm area struct point to a set of virtual memory handling
routines (via vm ops). This way all of the process's virtual memory can be handled
in a consistent way no matter how the underlying services managing that memory
dier. For example there is a routine that will be called when the process attempts
to access the memory and it does not exist, this is how page faults are handled.
The process's set of vm area struct data structures is accessed repeatedly by the
4.6. Creating a Process 45
Linux kernel as it creates new areas of virtual memory for the process and as it xes
up references to virtual memory not in the system's physical memory. This makes
the time that it takes to nd the correct vm area struct critical to the performance
of the system. To speed up this access, Linux also arranges the vm area struct data
structures into an AVL (Adelson-Velskii and Landis) tree. This tree is arranged so
that each vm area struct (or node) has a left and a right pointer to its neighbouring
vm area struct structure. The left pointer points to node with a lower starting
virtual address and the right pointer points to a node with a higher starting virtual
address. To nd the correct node, Linux goes to the root of the tree and follows
each node's left and right pointers until it nds the right vm area struct. Of course,
nothing is for free and inserting a new vm area struct into this tree takes additional
processing time.
When a process allocates virtual memory, Linux does not actually reserve physical
memory for the process. Instead, it describes the virtual memory by creating a
new vm area struct data structure. This is linked into the process's list of virtual
memory. When the process attempts to write to a virtual address within that new
virtual memory region then the system will page fault. The processor will attempt
to decode the virtual address, but as there are no Page Table Entries for any of this
memory, it will give up and raise a page fault exception, leaving the Linux kernel to
x things up. Linux looks to see if the virtual address referenced is in the current
process's virtual address space. If it is, Linux creates the appropriate PTEs and
allocates a physical page of memory for this process. The code or data may need to
be brought into that physical page from the lesystem or from the swap disk. The
process can then be restarted at the instruction that caused the page fault and, this
time as the memory physically exists, it may continue.
4.6 Creating a Process
When the system starts up it is running in kernel mode and there is, in a sense, only
one process, the initial process. Like all processes, the initial process has a machine
state represented by stacks, registers and so on. These will be saved in the initial
process's task struct data structure when other processes in the system are created
and run. At the end of system initialization, the initial process starts up a kernel
thread (called init) and then sits in an idle loop doing nothing. Whenever there
is nothing else to do the scheduler will run this, idle, process. The idle process's
task struct is the only one that is not dynamically allocated, it is statically dened
at kernel build time and is, rather confusingly, called init task.
The init kernel thread or process has a process identier of 1 as it is the system's
rst real process. It does some initial setting up of the system (such as opening the
system console and mounting the root le system) and then executes the system initialization
program. This is one of /etc/init, /bin/init or /sbin/init depending
on your system. The init program uses /etc/inittab as a script le to create new
processes within the system. These new processes may themselves go on to create
new processes. For example the getty process may create a login process when a
user attempts to login. All of the processes in the system are descended from the
init kernel thread.
New processes are created by cloning old processes, or rather by cloning the current
process. A new task is created by a system call (fork or clone) and the cloning See do fork() in
kernel/fork.c
46 Chapter 4. Processes
happens within the kernel in kernel mode. At the end of the system call there is a
new process waiting to run once the scheduler chooses it. A new task struct data
structure is allocated from the system's physical memory with one or more physical
pages for the cloned process's stacks (user and kernel). A new process identier may
be created, one that is unique within the set of process identiers in the system.
However, it is perfectly reasonable for the cloned process to keep its parents process
identier. The new task struct is entered into the task vector and the contents of
the old (current) process's task struct are copied into the cloned task struct.
When cloning processes Linux allows the two processes to share resources rather
than have two separate copies. This applies to the process's les, signal handlers and
virtual memory. When the resources are to be shared their respective count elds
are incremented so that Linux will not deallocate these resources until both processes
have nished using them. So, for example, if the cloned process is to share virtual
memory, its task struct will contain a pointer to the mm struct of the original
process and that mm struct has its count eld incremented to show the number of
current processes sharing it.
Cloning a process's virtual memory is rather tricky. A new set of vm area struct
data structures must be generated together with their owning mm struct data structure
and the cloned process's page tables. None of the process's virtual memory is
copied at this point. That would be a rather dicult and lengthy task for some
of that virtual memory would be in physical memory, some in the executable image
that the process is currently executing and possibly some would be in the swap
le. Instead Linux uses a technique called \copy on write" which means that virtual
memory will only be copied when one of the two processes tries to write to it. Any
virtual memory that is not written to, even if it can be, will be shared between the
two processes without any harm occuring. The read only memory, for example the
executable code, will always be shared. For \copy on write" to work, the writeable
areas have their page table entries marked as read only and the vm area struct
data structures describing them are marked as \copy on write". When one of the
processes attempts to write to this virtual memory a page fault will occur. It is at
this point that Linux will make a copy of the memory and x up the two processes'
page tables and virtual memory data structures.
4.7 Times and Timers
The kernel keeps track of a process's creation time as well as the CPU time that it
consumes during its lifetime. Each clock tick, the kernel updates the amount of time
in jiffies that the current process has spent in system and in user mode.
In addition to these accounting timers, Linux supports process specic interval
timers. A process can use these timers to send itself various signals each See time that
kernel/itimer.c
they expire. Three sorts of interval timers are supported:
Real the timer ticks in real time, and when the timer has expired, the process is
sent a SIGALRM signal.
Virtual This timer only ticks when the process is running and when it expires it
sends a SIGVTALRM signal.
4.8. Executing Programs 47
Prole This timer ticks both when the process is running and when the system is
executing on behalf of the process itself. SIGPROF is signalled when it expires.
One or all of the interval timers may be running and Linux keeps all of the neccessary
information in the process's task struct data structure. System calls can be made
to set up these interval timers and to start them, stop them and read their current
values. The virtual and prole timers are handled the same way. Every clock tick
See
do it virtual()
in
kernel/sched.c
See
do it prof() in
kernel/sched.c
the current process's interval timers are decremented and, if they have expired, the
appropriate signal is sent.
Real time interval timers are a little dierent and for these Linux uses the timer
mechanism described in Chapter 11. Each process has its own timer list data
structure and, when the real interval timer is running, this is queued on the system
timer list. When the timer expires the timer bottom half handler removes it from
the queue and calls the interval timer handler. This generates the SIGALRM signal
See
it real fn() in
and restarts the interval timer, adding it back into the system timer queue. kernel/itimer.c
4.8 Executing Programs
In Linux, as in UnixTM, programs and commands are normally executed by a command
interpreter. A command interpreter is a user process like any other process
and is called a shell 2. There are many shells in Linux, some of the most popular
are sh, bash and tcsh. With the exception of a few built in commands, such as cd
and pwd, a command is an executable binary le. For each command entered, the
shell searches the directories in the process's search path, held in the PATH environment
variable, for an executable image with a matching name. If the le is found it
is loaded and executed. The shell clones itself using the fork mechanism described
above and then the new child process replaces the binary image that it was executing,
the shell, with the contents of the executable image le just found. Normally
the shell waits for the command to complete, or rather for the child process to exit.
You can cause the shell to run again by pushing the child process to the background
by typing control-Z, which causes a SIGSTOP signal to be sent to the child process,
stopping it. You then use the shell command bg to push it into a background, the
shell sends it a SIGCONT signal to restart it, where it will stay until either it ends or
it needs to do terminal input or output.
An executable le can have many formats or even be a script le. Script les have
to be recognized and the appropriate interpreter run to handle them; for example
/bin/sh interprets shell scripts. Executable object les contain executable code and
data together with enough information to allow the operating system to load them
into memory and execute them. The most commonly used object le format used
by Linux is ELF but, in theory, Linux is
exible enough to handle almost any object
le format.
As with le systems, the binary formats supported by Linux are either built into the
kernel at kernel build time or available to be loaded as modules. The kernel keeps
a list of supported binary formats (see gure 4.3) and when an attempt is made
to execute a le, each binary format is tried in turn until one works. Commonly See do execve()
in fs/exec.c
supported Linux binary formats are a.out and ELF. Executable les do not have to
2Think of a nut the kernel is the edible bit in the middle and the shell goes around it, providing
an interface.
48 Chapter 4. Processes
*load_shlib()
linux_binfmt
next
use_count
*load_binary()
*core_dump()
*load_shlib()
linux_binfmt
next
use_count
*load_binary()
*core_dump()
*load_shlib()
linux_binfmt
next
use_count
*load_binary()
*core_dump()
formats
Figure 4.3: Registered Binary Formats
be read completely into memory, a technique known as demand loading is used. As
each part of the executable image is used by a process it is brought into memory.
Unused parts of the image may be discarded from memory.
4.8.1 ELF
The ELF (Executable and Linkable Format) object le format, designed by the Unix
System Laboratories, is now rmly established as the most commonly used format
in Linux. Whilst there is a slight performance overhead when compared with other
object le formats such as ECOFF and a.out, ELF is felt to be more
exible. ELF
executable les contain executable code, sometimes refered to as text, and data.
Tables within the executable image describe how the program should be placed into
the process's virtual memory. Statically linked images are built by the linker (ld),
or link editor, into one single image containing all of the code and data needed to
run this image. The image also species the layout in memory of this image and the
address in the image of the rst code to execute.
Figure 4.4 shows the layout of a statically linked ELF executable image. It is a simple
See
include/linux/-
elf.h C program that prints \hello world" and then exits. The header describes it as an
ELF image with two physical headers (e phnum is 2) starting 52 bytes (e phoff) from
the start of the image le. The rst physical header describes the executable code in
the image. It goes at virtual address 0x8048000 and there is 65532 bytes of it. This
is because it is a statically linked image which contains all of the library code for
the printf() call to output \hello world". The entry point for the image, the rst
instruction for the program, is not at the start of the image but at virtual address
0x8048090 (e entry). The code starts immediately after the second physical header.
This physical header describes the data for the program and is to be loaded into
virtual memory at address 0x8059BB8. This data is both readable and writeable.
You will notice that the size of the data in the le is 2200 bytes (p filesz) whereas
its size in memory is 4248 bytes. This because the rst 2200 bytes contain preinitialized
data and the next 2048 bytes contain data that will be initialized by the
executing code.
When Linux loads an ELF executable image into the process's virtual address space,
it does not actually load the image. It sets up the virtual memory data structures,
See do load
elf binary() in
fs/binfmt elf.c the process's vm area struct tree and its page tables. When the program is executed
page faults will cause the program's code and data to be fetched into physical
memory. Unused portions of the program will never be loaded into memory. Once
the ELF binary format loader is satised that the image is a valid ELF executable
image it
ushes the process's current executable image from its virtual memory. As
this process is a cloned image (all processes are) this, old, image is the program that
the parent process was executing, for example the command interpreter shell such
4.8. Executing Programs 49
ELF Executable Image
68536
PT_LOAD
PF_R, PF_W
2200
4248
0x8059BB8
p_type
p_offset
p_vaddr
p_filesz
p_memsz
p_flags
Data
Code
Physical Header
Physical Header
68532
68532
0x8048000
0
PT_LOAD
PF_R, PF_X
’E’ ’L’ ’F’
0x8048090
52
32
2
p_type
p_offset
p_vaddr
p_filesz
p_memsz
p_flags
e_ident
e_entry
e_phoff
e_phentsize
e_phnum
Figure 4.4: ELF Executable File Format
as bash. This
ushing of the old executable image discards the old virtual memory
data structures and resets the process's page tables. It also clears away any signal
handlers that were set up and closes any les that are open. At the end of the
ush
the process is ready for the new executable image. No matter what format the executable
image is, the same information gets set up in the process's mm struct. There
are pointers to the start and end of the image's code and data. These values are
found as the ELF executable images physical headers are read and the sections of
the program that they describe are mapped into the process's virtual address space.
That is also when the vm area struct data structures are set up and the process's
page tables are modied. The mm struct data structure also contains pointers to the
parameters to be passed to the program and to this process's environment variables.
ELF Shared Libraries
A dynamically linked image, on the other hand, does not contain all of the code and
data required to run. Some of it is held in shared libraries that are linked into the
image at run time. The ELF shared library's tables are also used by the dynamic
linker when the shared library is linked into the image at run time. Linux uses several
dynamic linkers, ld.so.1, libc.so.1 and ld-linux.so.1, all to be found in /lib.
The libraries contain commonly used code such as language subroutines. Without
dynamic linking, all programs would need their own copy of the these libraries and
would need far more disk space and virtual memory. In dynamic linking, information
is included in the ELF image's tables for every library routine referenced. The
information indicates to the dynamic linker how to locate the library routine and
50 Chapter 4. Processes
link it into the program's address space.
REVIEW NOTE: Do I need more detail here, worked example?
4.8.2 Script Files
Script les are executables that need an interpreter to run them. There are a wide
variety of interpreters available for Linux; for example wish, perl and command shells
such as tcsh. Linux uses the standard UnuxTM convention of having the rst line of
a script le contain the name of the interpreter. So, a typical script le would start:
#!/usr/bin/wish
The script binary loader tries to nd the intepreter for the script. It does this by
See
do load script()
in fs/-
binfmt script.c attempting to open the executable le that is named in the rst line of the script.
If it can open it, it has a pointer to its VFS inode and it can go ahead and have
it interpret the script le. The name of the script le becomes argument zero (the
rst argument) and all of the other arguments move up one place (the original rst
argument becomes the new second argument and so on). Loading the interpreter
is done in the same way as Linux loads all of its executable les. Linux tries each
binary format in turn until one works. This means that you could in theory stack
several interpreters and binary formats making the Linux binary format handler a
very
exible piece of software.
Chapter 5
Interprocess Communication
Mechanisms
Processes communicate with each other and with the kernel to coordinate
their activities. Linux supports a number of Inter-Process Communication
(IPC) mechanisms. Signals and pipes are two of them but Linux also
supports the System V IPC mechanisms named after the UnixTM release
in which they rst appeared.
5.1 Signals
Signals are one of the oldest inter-process communication methods used by UnixTM
systems. They are used to signal asynchronous events to one or more processes. A
signal could be generated by a keyboard interrupt or an error condition such as the
process attempting to access a non-existent location in its virtual memory. Signals
are also used by the shells to signal job control commands to their child processes.
There are a set of dened signals that the kernel can generate or that can be generated
by other processes in the system, provided that they have the correct privileges. You
can list a system's set of signals using the kill command (kill -l), on my Intel Linux
box this gives:
1) SIGHUP 2) SIGINT 3) SIGQUIT 4) SIGILL
5) SIGTRAP 6) SIGIOT 7) SIGBUS 8) SIGFPE
9) SIGKILL 10) SIGUSR1 11) SIGSEGV 12) SIGUSR2
13) SIGPIPE 14) SIGALRM 15) SIGTERM 17) SIGCHLD
51
52 Chapter 5. Interprocess Communication Mechanisms
18) SIGCONT 19) SIGSTOP 20) SIGTSTP 21) SIGTTIN
22) SIGTTOU 23) SIGURG 24) SIGXCPU 25) SIGXFSZ
26) SIGVTALRM 27) SIGPROF 28) SIGWINCH 29) SIGIO
30) SIGPWR
The numbers are dierent for an Alpha AXP Linux box. Processes can choose to
ignore most of the signals that are generated, with two notable exceptions: neither
the SIGSTOP signal which causes a process to halt its execution nor the SIGKILL
signal which causes a process to exit can be ignored. Otherwise though, a process
can choose just how it wants to handle the various signals. Processes can block
the signals and, if they do not block them, they can either choose to handle them
themselves or allow the kernel to handle them. If the kernel handles the signals, it will
do the default actions required for this signal. For example, the default action when
a process receives the SIGFPE (
oating point exception) signal is to core dump and
then exit. Signals have no inherent relative priorities. If two signals are generated
for a process at the same time then they may be presented to the process or handled
in any order. Also there is no mechanism for handling multiple signals of the same
kind. There is no way that a process can tell if it received 1 or 42 SIGCONT signals.
Linux implements signals using information stored in the task struct for the process.
The number of supported signals is limited to the word size of the processor.
Processes with a word size of 32 bits can have 32 signals whereas 64 bit processors
like the Alpha AXP may have up to 64 signals. The currently pending signals are
kept in the signal eld with a mask of blocked signals held in blocked. With the
exception of SIGSTOP and SIGKILL, all signals can be blocked. If a blocked signal
is generated, it remains pending until it is unblocked. Linux also holds information
about how each process handles every possible signal and this is held in an array of
sigaction data structures pointed at by the task struct for each process. Amongst
other things it contains either the address of a routine that will handle the signal or
a
ag which tells Linux that the process either wishes to ignore this signal or let the
kernel handle the signal for it. The process modies the default signal handling by
making system calls and these calls alter the sigaction for the appropriate signal
as well as the blocked mask.
Not every process in the system can send signals to every other process, the kernel
can and super users can. Normal processes can only send signals to processes with the
same uid and gid or to processes in the same process group1. Signals are generated
by setting the appropriate bit in the task struct's signal eld. If the process has
not blocked the signal and is waiting but interruptible (in state Interruptible) then
it is woken up by changing its state to Running and making sure that it is in the
run queue. That way the scheduler will consider it a candidate for running when the
system next schedules. If the default handling is needed, then Linux can optimize the
handling of the signal. For example if the signal SIGWINCH (the X window changed
focus) and the default handler is being used then there is nothing to be done.
Signals are not presented to the process immediately they are generated., they must
wait until the process is running again. Every time a process exits from a system call
its signal and blocked elds are checked and, if there are any unblocked signals,
they can now be delivered. This might seem a very unreliable method but every
process in the system is making system calls, for example to write a character to the
terminal, all of the time. Processes can elect to wait for signals if they wish, they
1REVIEW NOTE: Explain process groups.
5.2. Pipes 53
are suspended in state Interruptible until a signal is presented. The Linux signal
processing code looks at the sigaction structure for each of the current unblocked
signals.
If a signal's handler is set to the default action then the kernel will handle it. The
SIGSTOP signal's default handler will change the current process's state to Stopped
and then run the scheduler to select a new process to run. The default action for the
SIGFPE signal will core dump the process and then cause it to exit. Alternatively,
the process may have speced its own signal handler. This is a routine which will
be called whenever the signal is generated and the sigaction structure holds the
address of this routine. The kernel must call the process's signal handling routine
and how this happens is processor specic but all CPUs must cope with the fact
that the current process is running in kernel mode and is just about to return to
the process that called the kernel or system routine in user mode. The problem is
solved by manipulating the stack and registers of the process. The process's program
counter is set to the address of its signal handling routine and the parameters to the
routine are added to the call frame or passed in registers. When the process resumes
operation it appears as if the signal handling routine were called normally.
Linux is POSIX compatible and so the process can specify which signals are blocked
when a particular signal handling routine is called. This means changing the blocked
mask during the call to the processes signal handler. The blocked mask must be
returned to its original value when the signal handling routine has nished. Therefore
Linux adds a call to a tidy up routine which will restore the original blocked mask
onto the call stack of the signalled process. Linux also optimizes the case where
several signal handling routines need to be called by stacking them so that each time
one handling routine exits, the next one is called until the tidy up routine is called.
5.2 Pipes
The common Linux shells all allow redirection. For example
$ ls | pr | lpr
pipes the output from the ls command listing the directory's les into the standard
input of the pr command which paginates them. Finally the standard output from
the pr command is piped into the standard input of the lpr command which prints
the results on the default printer. Pipes then are unidirectional byte streams which
connect the standard output from one process into the standard input of another
process. Neither process is aware of this redirection and behaves just as it would
normally. It is the shell which sets up these temporary pipes between the processes.
In Linux, a pipe is implemented using two file data structures which both point at
the same temporary VFS inode which itself points at a physical page within memory.
Figure 5.1 shows that each file data structure contains pointers to dierent le
operation routine vectors; one for writing to the pipe, the other for reading from
the pipe. This hides the underlying dierences from the generic system calls which
See
include/linux/-
read and write to ordinary les. As the writing process writes to the pipe, bytes are inode fs i.h
copied into the shared data page and when the reading process reads from the pipe,
54 Chapter 5. Interprocess Communication Mechanisms
Data Page
Pipe
Write
Operations
Process 1 Process 2
f_mode
f_pos
f_flags
f_count
f_owner
f_inode
f_op
f_version
f_mode
f_pos
f_flags
f_count
f_owner
f_inode
f_op
f_version
file file
inode
Pipe
Operations
Read
Figure 5.1: Pipes
bytes are copied from the shared data page. Linux must synchronize access to the
pipe. It must make sure that the reader and the writer of the pipe are in step and
to do this it uses locks, wait queues and signals.
When the writer wants to write to the pipe it uses the standard write library functions.
These all pass le descriptors that are indices into the process's set of file
data structures, each one representing an open le or, as in this case, an open pipe.
The Linux system call uses the write routine pointed at by the file data structure
describing this pipe. That write routine uses information held in the VFS inode
representing the pipe to manage the write request. If there is enough room to write
See
pipe write() in
fs/pipe.c all of the bytes into the pipe and, so long as the pipe is not locked by its reader,
Linux locks it for the writer and copies the bytes to be written from the process's
address space into the shared data page. If the pipe is locked by the reader or if
there is not enough room for the data then the current process is made to sleep on
the pipe inode's wait queue and the scheduler is called so that another process can
run. It is interruptible, so it can receive signals and it will be woken by the reader
when there is enough room for the write data or when the pipe is unlocked. When
the data has been written, the pipe's VFS inode is unlocked and any waiting readers
sleeping on the inode's wait queue will themselves be woken up.
Reading data from the pipe is a very similar process to writing to it. Processes are See pipe read()
in fs/pipe.c
allowed to do non-blocking reads (it depends on the mode in which they opened
the le or pipe) and, in this case, if there is no data to be read or if the pipe is
locked, an error will be returned. This means that the process can continue to run.
The alternative is to wait on the pipe inode's wait queue until the write process
has nished. When both processes have nished with the pipe, the pipe inode is
discarded along with the shared data page.
5.3. Sockets 55
Linux also supports named pipes, also known as FIFOs because pipes operate on a
First In, First Out principle. The rst data written into the pipe is the rst data
read from the pipe. Unlike pipes, FIFOs are not temporary objects, they are entities
in the le system and can be created using the mkfo command. Processes are free to
use a FIFO so long as they have appropriate access rights to it. The way that FIFOs
are opened is a little dierent from pipes. A pipe (its two file data structures, its
VFS inode and the shared data page) is created in one go whereas a FIFO already
exists and is opened and closed by its users. Linux must handle readers opening
the FIFO before writers open it as well as readers reading before any writers have
written to it. That aside, FIFOs are handled almost exactly the same way as pipes
and they use the same data structures and operations.
5.3 Sockets
REVIEW NOTE: Add when networking chapter written.
5.3.1 System V IPC Mechanisms
Linux supports three types of interprocess communication mechanisms that rst
appeared in UnixTM System V (1983). These are message queues, semaphores and
shared memory. These System V IPC mechanisms all share common authentication
methods. Processes may access these resources only by passing a unique reference
identier to the kernel via system calls. Access to these System V IPC objects is
checked using access permissions, much like accesses to les are checked. The access
rights to the System V IPC object is set by the creator of the object via system calls.
The object's reference identier is used by each mechanism as an index into a table
of resources. It is not a straight forward index but requires some manipulation to
generate the index.
All Linux data structures representing System V IPC objects in the system include
an ipc perm structure which contains the owner and creator process's user and
See
include/linux/-
group identiers. The access mode for this object (owner, group and other) and the ipc.h
IPC object's key. The key is used as a way of locating the System V IPC object's
reference identier. Two sets of keys are supported: public and private. If the key
is public then any process in the system, subject to rights checking, can nd the
reference identier for the System V IPC object. System V IPC objects can never
be referenced with a key, only by their reference identier.
5.3.2 Message Queues
Message queues allow one or more processes to write messages, which will be read
by one or more reading processes. Linux maintains a list of message queues, the
msgque vector; each element of which points to a msqid ds data structure that fully
describes the message queue. When message queues are created a new msqid ds
data structure is allocated from system memory and inserted into the vector.
Each msqid ds data structure contains an ipc perm data structure and pointers to
See
include/linux/-
the messages entered onto this queue. In addition, Linux keeps queue modication msg.h
times such as the last time that this queue was written to and so on. The msqid ds
56 Chapter 5. Interprocess Communication Mechanisms
times
*wwait
msg_qnum
*msg_last
msqid_ds
ipc
*rwait
*msg_first
msg_qnum
msg
msg_type
*msg_spot
msg_stime
msg_ts
*msg_next
msg_ts message
msg
msg_type
*msg_spot
msg_stime
msg_ts
*msg_next
msg_ts message
Figure 5.2: System V IPC Message Queues
also contains two wait queues; one for the writers to the queue and one for the readers
of the message queue.
Each time a process attempts to write a message to the write queue its eective user
and group identiers are compared with the mode in this queue's ipc perm data
structure. If the process can write to the queue then the message may be copied
from the process's address space into a msg data structure and put at the end of this
message queue. Each message is tagged with an application specic type, agreed
between the cooperating processes. However, there may be no room for the message
as Linux restricts the number and length of messages that can be written. In this
case the process will be added to this message queue's write wait queue and the
scheduler will be called to select a new process to run. It will be woken up when one
or more messages have been read from this message queue.
Reading from the queue is a similar process. Again, the processes access rights to
the write queue are checked. A reading process may choose to either get the rst
message in the queue regardless of its type or select messages with particular types.
If no messages match this criteria the reading process will be added to the message
queue's read wait queue and the scheduler run. When a new message is written to
the queue this process will be woken up and run again.
5.3.3 Semaphores
In its simplest form a semaphore is a location in memory whose value can be tested
and set by more than one process. The test and set operation is, so far as each process
is concerned, uninterruptible or atomic; once started nothing can stop it. The result
of the test and set operation is the addition of the current value of the semaphore
and the set value, which can be positive or negative. Depending on the result of the
test and set operation one process may have to sleep until the semphore's value is
changed by another process. Semaphores can be used to implement critical regions,
areas of critical code that only one process at a time should be executing.
Say you had many cooperating processes reading records from and writing records
to a single data le. You would want that le access to be strictly coordinated. You
5.3. Sockets 57
proc_next
id_next
semid
semadj
sem_undo
sem_queue
next
prev
sleeper
undo
pid
status
sma
sops
nsops
sem_nsems
undo
sem_pending_last
sem_pending
sem_base
times
ipc
semid_ds
array of
semaphores
Figure 5.3: System V IPC Semaphores
could use a semaphore with an initial value of 1 and, around the le operating code,
put two semaphore operations, the rst to test and decrement the semaphore's value
and the second to test and increment it. The rst process to access the le would
try to decrement the semaphore's value and it would succeed, the semaphore's value
now being 0. This process can now go ahead and use the data le but if another
process wishing to use it now tries to decrement the semaphore's value it would fail
as the result would be -1. That process will be suspended until the rst process has
nished with the data le. When the rst process has nished with the data le it
will increment the semaphore's value, making it 1 again. Now the waiting process
can be woken and this time its attempt to increment the semaphore will succeed.
System V IPC semaphore objects each describe a semaphore array and Linux uses
the semid ds data structure to represent this. All of the semid ds data structures in
See
include/linux/-
the system are pointed at by the semary, a vector of pointers. There are sem nsems sem.h
in each semaphore array, each one described by a sem data structure pointed at by
sem base. All of the processes that are allowed to manipulate the semaphore array of
a System V IPC semaphore object may make system calls that perform operations on
them. The system call can specify many operations and each operation is described
by three inputs; the semaphore index, the operation value and a set of
ags. The
semaphore index is an index into the semaphore array and the operation value is a
numerical value that will be added to the current value of the semaphore. First Linux
tests whether or not all of the operations would succeed. An operation will succeed
if the operation value added to the semaphore's current value would be greater than
zero or if both the operation value and the semaphore's current value are zero. If
any of the semaphore operations would fail Linux may suspend the process but only
if the operation
ags have not requested that the system call is non-blocking. If
the process is to be suspended then Linux must save the state of the semaphore
operations to be performed and put the current process onto a wait queue. It does
this by building a sem queue data structure on the stack and lling it out. The new
sem queue data structure is put at the end of this semaphore object's wait queue
(using the sem pending and sem pending last pointers). The current process is
58 Chapter 5. Interprocess Communication Mechanisms
put on the wait queue in the sem queue data structure (sleeper) and the scheduler
called to choose another process to run.
If all of the semaphore operations would have succeeded and the current process
does not need to be suspended, Linux goes ahead and applies the operations to the
appropriate members of the semaphore array. Now Linux must check that any waiting,
suspended, processes may now apply their semaphore operations. It looks at
each member of the operations pending queue (sem pending) in turn, testing to see
if the semphore operations will succeed this time. If they will then it removes the
sem queue data structure from the operations pending list and applies the semaphore
operations to the semaphore array. It wakes up the sleeping process making it available
to be restarted the next time the scheduler runs. Linux keeps looking through
the pending list from the start until there is a pass where no semaphore operations
can be applied and so no more processes can be woken.
There is a problem with semaphores, deadlocks. These occur when one process has
altered the semaphores value as it enters a critical region but then fails to leave
the critical region because it crashed or was killed. Linux protects against this by
maintaining lists of adjustments to the semaphore arrays. The idea is that when
these adjustments are applied, the semaphores will be put back to the state that
they were in before the a process's set of semaphore operations were applied. These
adjustments are kept in sem undo data structures queued both on the semid ds
data structure and on the task struct data structure for the processes using these
semaphore arrays.
Each individual semaphore operation may request that an adjustment be maintained.
Linux will maintain at most one sem undo data structure per process for
each semaphore array. If the requesting process does not have one, then one is created
when it is needed. The new sem undo data structure is queued both onto this
process's task struct data structure and onto the semaphore array's semid ds data
structure. As operations are applied to the semphores in the semaphore array the
negation of the operation value is added to this semphore's entry in the adjustment
array of this process's sem undo data structure. So, if the operation value is 2, then
-2 is added to the adjustment entry for this semaphore.
When processes are deleted, as they exit Linux works through their set of sem undo
data structures applying the adjustments to the semaphore arrays. If a semaphore set
is deleted, the sem undo data structures are left queued on the process's task struct
but the semaphore array identier is made invalid. In this case the semaphore clean
up code simply discards the sem undo data structure.
5.3.4 Shared Memory
Shared memory allows one or more processes to communicate via memory that appears
in all of their virtual address spaces. The pages of the virtual memory is
referenced by page table entries in each of the sharing processes' page tables. It does
not have to be at the same address in all of the processes' virtual memory. As with
all System V IPC objects, access to shared memory areas is controlled via keys and
access rights checking. Once the memory is being shared, there are no checks on how
the processes are using it. They must rely on other mechanisms, for example System
V semaphores, to synchronize access to the memory.
Each newly created shared memory area is represented by a shmid ds data structure.
5.3. Sockets 59
vm_area_struct
vm_next_shared
vm_area_struct
vm_next_shared
pte
pte
pte
ipc
shm_segsz
times
shm_npages
attaches
shm_pages
shmid_ds
Figure 5.4: System V IPC Shared Memory
These are kept in the shm segs vector. The shmid ds data structure decribes how See
include/linux/-
sem.h
big the area of shared memory is, how many processes are using it and information
about how that shared memory is mapped into their address spaces. It is the creator
of the shared memory that controls the access permissions to that memory and
whether its key is public or private. If it has enough access rights it may also lock
the shared memory into physical memory.
Each process that wishes to share the memory must attach to that virtual memory
via a system call. This creates a new vm area struct data structure describing the
shared memory for this process. The process can choose where in its virtual address
space the shared memory goes or it can let Linux choose a free area large enough.
The new vm area struct structure is put into the list of vm area struct pointed
at by the shmid ds. The vm next shared and vm prev shared pointers are used to
link them together. The virtual memory is not actually created during the attach;
it happens when the rst process attempts to access it.
The rst time that a process accesses one of the pages of the shared virtual memory,
a page fault will occur. When Linux xes up that page fault it nds the
vm area struct data structure describing it. This contains pointers to handler routines
for this type of shared virtual memory. The shared memory page fault handling
code looks in the list of page table entries for this shmid ds to see if one exists for
this page of the shared virtual memory. If it does not exist, it will allocate a physical
page and create a page table entry for it. As well as going into the current process's
page tables, this entry is saved in the shmid ds. This means that when the next
process that attempts to access this memory gets a page fault, the shared memory
fault handling code will use this newly created physical page for that process too. So,
the rst process that accesses a page of the shared memory causes it to be created
and thereafter access by the other processes cause that page to be added into their
virtual address spaces.
When processes no longer wish to share the virtual memory, they detach from it.
So long as other processes are still using the memory the detach only aects the
current process. Its vm area struct is removed from the shmid ds data structure
and deallocated. The current process's page tables are updated to invalidate the area
60 Chapter 5. Interprocess Communication Mechanisms
of virtual memory that it used to share. When the last process sharing the memory
detaches from it, the pages of the shared memory current in physical memory are
freed, as is the shmid ds data structure for this shared memory.
Further complications arise when shared virtual memory is not locked into physical
memory. In this case the pages of the shared memory may be swapped out to
the system's swap disk during periods of high memory usage. How shared memory
memory is swapped into and out of physical memory is described in Chapter 3.
Chapter 6
PCI
Peripheral Component Interconnect (PCI), as its name implies is a standard
that describes how to connect the peripheral components of a system
together in a structured and controlled way. The standard[3, PCI
Local Bus Specication] describes the way that the system components
are electrically connected and the way that they should behave. This
chapter looks at how the Linux kernel initializes the system's PCI buses
and devices.
Figure 6.1 is a logical diagram of an example PCI based system. The PCI buses
and PCI-PCI bridges are the glue connecting the system components together; the
CPU is connected to PCI bus 0, the primary PCI bus as is the video device. A
special PCI device, a PCI-PCI bridge connects the primary bus to the secondary
PCI bus, PCI bus 1. In the jargon of the PCI specication, PCI bus 1 is described
as being downstream of the PCI-PCI bridge and PCI bus 0 is up-stream of the
bridge. Connected to the secondary PCI bus are the SCSI and ethernet devices for
the system. Physically the bridge, secondary PCI bus and two devices would all be
contained on the same combination PCI card. The PCI-ISA bridge in the system
supports older, legacy ISA devices and the diagram shows a super I/O controller
chip, which controls the keyboard, mouse and
oppy. 1
6.1 PCI Address Spaces
The CPU and the PCI devices need to access memory that is shared between them.
This memory is used by device drivers to control the PCI devices and to pass in-
1For example?
61
62 Chapter 6. PCI
SCSI Ethernet
Video
Upstream
Downstream
PCI-ISA
Bridge
CPU
PCI-PCI
Bridge
PCI Bus 0
ISA Bus PCI Bus 1
Super I/O Controller
Figure 6.1: Example PCI Based System
formation between them. Typically the shared memory contains control and status
registers for the device. These registers are used to control the device and to read
its status. For example, the PCI SCSI device driver would read its status register
to nd out if the SCSI device was ready to write a block of information to the SCSI
disk. Or it might write to the control register to start the device running after it has
been turned on.
The CPU's system memory could be used for this shared memory but if it were,
then every time a PCI device accessed memory, the CPU would have to stall, waiting
for the PCI device to nish. Access to memory is generally limited to one system
component at a time. This would slow the system down. It is also not a good idea to
allow the system's peripheral devices to access main memory in an uncontrolled way.
This would be very dangerous; a rogue device could make the system very unstable.
Peripheral devices have their own memory spaces. The CPU can access these spaces
but access by the devices into the system's memory is very strictly controlled using
DMA (Direct Memory Access) channels. ISA devices have access to two address
spaces, ISA I/O (Input/Output) and ISA memory. PCI has three; PCI I/O, PCI
Memory and PCI Conguration space. All of these address spaces are also accessible
by the CPU with the the PCI I/O and PCI Memory address spaces being used by the
device drivers and the PCI Conguration space being used by the PCI initialization
code within the Linux kernel.
The Alpha AXP processor does not have natural access to addresses spaces other
than the system address space. It uses support chipsets to access other address
spaces such as PCI Conguration space. It uses a sparse address mapping scheme
which steals part of the large virtual address space and maps it to the PCI address
spaces.
6.2 PCI Conguration Headers
Every PCI device in the system, including the PCI-PCI bridges has a conguration
data structure that is somewhere in the PCI conguration address space. The PCI
Conguration header allows the system to identify and control the device. Exactly
6.2. PCI Conguration Headers 63
31 16 15 0
Device Id Vendor Id
Status Command
Class Code
00h
04h
08h
24h
10h
3Ch
Base Address Registers
Line Pin
Figure 6.2: The PCI Conguration Header
where the header is in the PCI Conguration address space depends on where in
the PCI topology that device is. For example, a PCI video card plugged into one
PCI slot on the PC motherboard will have its conguration header at one location
and if it is plugged into another PCI slot then its header will appear in another
location in PCI Conguration memory. This does not matter, for wherever the PCI
devices and bridges are the system will nd and congure them using the status and
conguration registers in their conguration headers.
Typically, systems are designed so that every PCI slot has it's PCI Conguration
Header in an oset that is related to its slot on the board. So, for example, the rst
slot on the board might have its PCI Conguration at oset 0 and the second slot at
oset 256 (all headers are the same length, 256 bytes) and so on. A system specic
hardware mechanism is dened so that the PCI conguration code can attempt to
examine all possible PCI Conguration Headers for a given PCI bus and know which
devices are present and which devices are absent simply by trying to read one of the
elds in the header (usually the Vendor Identication eld) and getting some sort
of error. The [3, PCI Local Bus Specication] describes one possible error message
as returning 0xFFFFFFFF when attempting to read the Vendor Identication and
Device Identication elds for an empty PCI slot.
Figure 6.2 shows the layout of the 256 byte PCI conguration header. It contains
the following elds:
See
include/linux/-
pci.h
Vendor Identication A unique number describing the originator of the PCI device.
Digital's PCI Vendor Identication is 0x1011 and Intel's is 0x8086.
Device Identication A unique number describing the device itself. For example,
Digital's 21141 fast ethernet device has a device identication of 0x0009.
64 Chapter 6. PCI
Status This eld gives the status of the device with the meaning of the bits of this
eld set by the standard. [3, PCI Local Bus Specication].
Command By writing to this eld the system controls the device, for example
allowing the device to access PCI I/O memory,
Class Code This identies the type of device that this is. There are standard
classes for every sort of device; video, SCSI and so on. The class code for SCSI
is 0x0100.
Base Address Registers These registers are used to determine and allocate the
type, amount and location of PCI I/O and PCI memory space that the device
can use.
Interrupt Pin Four of the physical pins on the PCI card carry interrupts from
the card to the PCI bus. The standard labels these as A, B, C and D. The
Interrupt Pin eld describes which of these pins this PCI device uses. Generally
it is hardwired for a pariticular device. That is, every time the system boots,
the device uses the same interrupt pin. This information allows the interrupt
handling subsystem to manage interrupts from this device,
Interrupt Line The Interrupt Line eld of the device's PCI Conguration header
is used to pass an interrupt handle between the PCI initialisation code, the
device's driver and Linux's interrupt handling subsystem. The number written
there is meaningless to the the device driver but it allows the interrupt handler
to correctly route an interrupt from the PCI device to the correct device driver's
interrupt handling code within the Linux operating system. See Chapter 7 on
page 75 for details on how Linux handles interrupts.
6.3 PCI I/O and PCI Memory Addresses
These two address spaces are used by the devices to communicate with their device
drivers running in the Linux kernel on the CPU. For example, the DECchip 21141
fast ethernet device maps its internal registers into PCI I/O space. Its Linux device
driver then reads and writes those registers to control the device. Video drivers
typically use large amounts of PCI memory space to contain video information.
Until the PCI system has been set up and the device's access to these address spaces
has been turned on using the Command eld in the PCI Conguration header, nothing
can access them. It should be noted that only the PCI conguration code reads
and writes PCI conguration addresses; the Linux device drivers only read and write
PCI I/O and PCI memory addresses.
6.4 PCI-ISA Bridges
These bridges support legacy ISA devices by translating PCI I/O and PCI Memory
space accesses into ISA I/O and ISA Memory accesses. A lot of systems now sold
contain several ISA bus slots and several PCI bus slots. Over time the need for this
backwards compatibility will dwindle and PCI only systems will be sold. Where in
the ISA address spaces (I/O and Memory) the ISA devices of the system have their
registers was xed in the dim mists of time by the early Intel 8080 based PCs. Even a
6.5. PCI-PCI Bridges 65
31 11 10 8 7 2 1 0
Device Select Func Register 0 0
Figure 6.3: Type 0 PCI Conguration Cycle
31 11 10 8 7 2 1 0
Reserved Bus Device Func Register 0 1
24 23 16 15
Figure 6.4: Type 1 PCI Conguration Cycle
$5000 Alpha AXP based computer systems will have its ISA
oppy controller at the
same place in ISA I/O space as the rst IBM PC. The PCI specication copes with
this by reserving the lower regions of the PCI I/O and PCI Memory address spaces
for use by the ISA peripherals in the system and using a single PCI-ISA bridge to
translate any PCI memory accesses to those regions into ISA accesses.
6.5 PCI-PCI Bridges
PCI-PCI bridges are special PCI devices that glue the PCI buses of the system
together. Simple systems have a single PCI bus but there is an electrical limit on the
number of PCI devices that a single PCI bus can support. Using PCI-PCI bridges to
add more PCI buses allows the system to support many more PCI devices. This is
particularly important for a high performance server. Of course, Linux fully supports
the use of PCI-PCI bridges.
6.5.1 PCI-PCI Bridges: PCI I/O and PCI Memory Windows
PCI-PCI bridges only pass a subset of PCI I/O and PCI memory read and write
requests downstream. For example, in Figure 6.1 on page 62, the PCI-PCI bridge
will only pass read and write addresses from PCI bus 0 to PCI bus 1 if they are
for PCI I/O or PCI memory addresses owned by either the SCSI or ethernet device;
all other PCI I/O and memory addresses are ignored. This ltering stops addresses
propogating needlessly throughout the system. To do this, the PCI-PCI bridges must
be programmed with a base and limit for PCI I/O and PCI Memory space access
that they have to pass from their primary bus onto their secondary bus. Once the
PCI-PCI Bridges in a system have been congured then so long as the Linux device
drivers only access PCI I/O and PCI Memory space via these windows, the PCI-PCI
Bridges are invisible. This is an important feature that makes life easier for Linux
PCI device driver writers. However, it also makes PCI-PCI bridges somewhat tricky
for Linux to congure as we shall see later on.
6.5.2 PCI-PCI Bridges: PCI Conguration Cycles and PCI
Bus Numbering
So that the CPU's PCI initialization code can address devices that are not on the
main PCI bus, there has to be a mechanism that allows bridges to decide whether
or not to pass Conguration cycles from their primary interface to their secondary
66 Chapter 6. PCI
interface. A cycle is just an address as it appears on the PCI bus. The PCI speci-
cation denes two formats for the PCI Conguration addresses; Type 0 and Type 1;
these are shown in Figure 6.3 and Figure 6.4 respectively. Type 0 PCI Conguration
cycles do not contain a bus number and these are interpretted by all devices as being
for PCI conguration addresses on this PCI bus. Bits 31:11 of the Type 0 conguraration
cycles are treated as the device select eld. One way to design a system is to
have each bit select a dierent device. In this case bit 11 would select the PCI device
in slot 0, bit 12 would select the PCI device in slot 1 and so on. Another way is to
write the device's slot number directly into bits 31:11. Which mechanism is used in
a system depends on the system's PCI memory controller.
Type 1 PCI Conguration cycles contain a PCI bus number and this type of con-
guration cycle is ignored by all PCI devices except the PCI-PCI bridges. All of
the PCI-PCI Bridges seeing Type 1 conguration cycles may choose to pass them
to the PCI buses downstream of themselves. Whether the PCI-PCI Bridge ignores
the Type 1 conguration cycle or passes it onto the downstream PCI bus depends
on how the PCI-PCI Bridge has been congured. Every PCI-PCI bridge has a primary
bus interface number and a secondary bus interface number. The primary bus
interface being the one nearest the CPU and the secondary bus interface being the
one furthest away. Each PCI-PCI Bridge also has a subordinate bus number and
this is the maximum bus number of all the PCI buses that are bridged beyond the
secondary bus interface. Or to put it another way, the subordinate bus number is the
highest numbered PCI bus downstream of the PCI-PCI bridge. When the PCI-PCI
bridge sees a Type 1 PCI conguration cycle it does one of the following things:
 Ignore it if the bus number specied is not in between the bridge's secondary
bus number and subordinate bus number (inclusive),
 Convert it to a Type 0 conguration command if the bus number specied
matches the secondary bus number of the bridge,
 Pass it onto the secondary bus interface unchanged if the bus number speci-
ed is greater than the secondary bus number and less than or equal to the
subordinate bus number.
So, if we want to address Device 1 on bus 3 of the topology Figure 6.9 on page 71
we must generate a Type 1 Conguration command from the CPU. Bridge1 passes
this unchanged onto Bus 1. Bridge2 ignores it but Bridge3 converts it into a Type
0 Conguration command and sends it out on Bus 3 where Device 1 responds to it.
It is up to each individual operating system to allocate bus numbers during PCI
conguration but whatever the numbering scheme used the following statement must
be true for all of the PCI-PCI bridges in the system:
\All PCI buses located behind a PCI-PCI bridge must reside between the seondary
bus number and the subordinate bus number (inclusive)."
If this rule is broken then the PCI-PCI Bridges will not pass and translate Type 1
PCI conguration cycles correctly and the system will fail to nd and initialise the
PCI devices in the system. To achieve this numbering scheme, Linux congures these
special devices in a particular order. Section 6.6.2 on page 68 describes Linux's PCI
bridge and bus numbering scheme in detail together with a worked example.
6.6. Linux PCI Initialization 67
pci_root
pci_dev
bus
sibling
next
pci_dev
bus
sibling
next
pci_dev
bus
sibling
next
pci_dev
bus
sibling
next
pci_dev
bus
sibling
next
pci_bus
parent
children
next
self
devices
pci_bus
parent
children
next
self
devices
bus = 0
bus = 1
PCI-ISA Bridge Video PCI-PCI Bridge
SCSI Ethernet
Figure 6.5: Linux Kernel PCI Data Structures
6.6 Linux PCI Initialization
The PCI initialisation code in Linux is broken into three logical parts:
PCI Device Driver This pseudo-device driver searches the PCI system starting at
Bus 0 and locates all PCI devices and bridges in the system. It builds a linked
list of data structures describing the topology of the system. Additionally, it
numbers all of the bridges that it nds.
See
drivers/pci/-
pci.c and
include/linux/-
PCI BIOS This software layer provides the services described in [4, PCI BIOS pci.h
ROM specication]. Even though Alpha AXP does not have BIOS services,
there is equivalent code in the Linux kernel providing the same functions,
See
arch/*/kernel/-
bios32.c
PCI Fixup System specic xup code tidies up the system specic loose ends of
PCI initialization. See
arch/*/kernel/-
bios32.c
6.6.1 The Linux Kernel PCI Data Structures
As the Linux kernel initialises the PCI system it builds data structures mirroring
the real PCI topology of the system. Figure 6.5 shows the relationships of the data
68 Chapter 6. PCI
structures that it would build for the example PCI system in Figure 6.1 on page 62.
Each PCI device (including the PCI-PCI Bridges) is described by a pci dev data
structure. Each PCI bus is described by a pci bus data structure. The result is a
tree structure of PCI buses each of which has a number of child PCI devices attached
to it. As a PCI bus can only be reached using a PCI-PCI Bridge (except the primary
PCI bus, bus 0), each pci bus contains a pointer to the PCI device (the PCI-PCI
Bridge) that it is accessed through. That PCI device is a child of the the PCI Bus's
parent PCI bus.
Not shown in the Figure 6.5 is a pointer to all of the PCI devices in the system,
pci devices. All of the PCI devices in the system have their pci dev data structures
queued onto this queue.. This queue is used by the Linux kernel to quickly nd all
of the PCI devices in the system.
6.6.2 The PCI Device Driver
The PCI device driver is not really a device driver at all but a function of the
operating system called at system initialisation time. The PCI initialisation code
must scan all of the PCI buses in the system looking for all PCI devices in the
system (including PCI-PCI bridge devices). It uses the PCI BIOS code to nd out if
See Scan bus()
in drivers/-
pci/pci.c every possible slot in the current PCI bus that it is scanning is occupied. If the PCI
slot is occupied, it builds a pci dev data structure describing the device and links
into the list of known PCI devices (pointed at by pci devices).
The PCI initialisation code starts by scanning PCI Bus 0. It tries to read the
Vendor Identication and Device Identication elds for every possible PCI device
in every possible PCI slot. When it nds an occupied slot it builds a pci dev data
structure describing the device. All of the pci dev data structures built by the PCI
initialisation code (including all of the PCI-PCI Bridges) are linked into a singly
linked list; pci devices.
If the PCI device that was found was a PCI-PCI bridge then a pci bus data structure
is built and linked into the tree of pci bus and pci dev data structures pointed at by
pci root. The PCI initialisation code can tell if the PCI device is a PCI-PCI Bridge
because it has a class code of 0x060400. The Linux kernel then congures the PCI bus
on the other (downstream) side of the PCI-PCI Bridge that it has just found. If more
PCI-PCI Bridges are found then these are also congured. This process is known as
a depthwise algorithm; the system's PCI topology is fully mapped depthwise before
searching breadthwise. Looking at Figure 6.1 on page 62, Linux would congure PCI
Bus 1 with its Ethernet and SCSI device before it congured the video device on
PCI Bus 0.
As Linux searches for downstream PCI buses it must also congure the intervening
PCI-PCI bridges' secondary and subordinate bus numbers. This is described in detail
in Section 6.6.2 below.
Conguring PCI-PCI Bridges - Assigning PCI Bus Numbers
For PCI-PCI bridges to pass PCI I/O, PCI Memory or PCI Conguration address
space reads and writes across them, they need to know the following:
Primary Bus Number The bus number immediately upstream of the PCI-PCI
Bridge,
6.6. Linux PCI Initialization 69
DI D2
1
Bridge
DI D2
DI
DI D2
CPU
Bus 0
Bus 1
Bridge Bridge
3 2
Bus ?
Bus ?
Bus ?
Bridge
4
Secondary Bus = 1
Subordinate=0xFF
Primary Bus = 0
Figure 6.6: Conguring a PCI System: Part 1
Secondary Bus Number The bus number immediately downstream of the PCIPCI
Bridge,
Subordinate Bus Number The highest bus number of all of the buses that can
be reached downstream of the bridge.
PCI I/O and PCI Memory Windows The window base and size for PCI I/O
address space and PCI Memory address space for all addresses downstream of
the PCI-PCI Bridge.
The problem is that at the time when you wish to congure any given PCI-PCI bridge
you do not know the subordinate bus number for that bridge. You do not know if
there are further PCI-PCI bridges downstream and if you did, you do not know
what numbers will be assigned to them. The answer is to use a depthwise recursive
algorithm and scan each bus for any PCI-PCI bridges assigning them numbers as
they are found. As each PCI-PCI bridge is found and its secondary bus numbered,
assign it a temporary subordinate number of 0xFF and scan and assign numbers to
all PCI-PCI bridges downstream of it. This all seems complicated but the worked
example below makes this process clearer.
PCI-PCI Bridge Numbering: Step 1 Taking the topology in Figure 6.6, the
rst bridge the scan would nd is Bridge1. The PCI bus downstream of
Bridge1 would be numbered as 1 and Bridge1 assigned a secondary bus number
of 1 and a temporary subordinate bus number of 0xFF. This means that
all Type 1 PCI Conguration addresses specifying a PCI bus number of 1 or
higher would be passed across Bridge1 and onto PCI Bus 1. They would be
translated into Type 0 Conguration cycles if they have a bus number of 1 but
left untranslated for all other bus numbers. This is exactly what the Linux
PCI initialisation code needs to do in order to go and scan PCI Bus 1.
70 Chapter 6. PCI
DI D2
1
Bridge
DI D2
DI
DI D2
CPU
Bus 0
Bus 1
Bridge Bridge
3 2
Bus ?
Bus ?
Bridge
4
Secondary Bus = 2
Bus 2
Subordinate=0xFF
Primary Bus = 0
Secondary Bus = 1
Primary Bus = 1
Subordinate=2
Figure 6.7: Conguring a PCI System: Part 2
PCI-PCI Bridge Numbering: Step 2 Linux uses a depthwise algorithm and so
the initialisation code goes on to scan PCI Bus 1. Here it nds PCI-PCI
Bridge2. There are no further PCI-PCI bridges beyond PCI-PCI Bridge2, so
it is assigned a subordinate bus number of 2 which matches the number assigned
to its secondary interface. Figure 6.7 shows how the buses and PCI-PCI bridges
are numbered at this point.
PCI-PCI Bridge Numbering: Step 3 The PCI initialisation code returns to scanning
PCI Bus 1 and nds another PCI-PCI bridge, Bridge3. It is assigned 1
as its primary bus interface number, 3 as its secondary bus interface number
and 0xFF as its subordinate bus number. Figure 6.8 on page 71 shows how the
system is congured now. Type 1 PCI conguration cycles with a bus number
of 1, 2 or 3 wil be correctly delivered to the appropriate PCI buses.
PCI-PCI Bridge Numbering: Step 4 Linux starts scanning PCI Bus 3, downstream
of PCI-PCI Bridge3. PCI Bus 3 has another PCI-PCI bridge (Bridge4)
on it, it is assigned 3 as its primary bus number and 4 as its secondary bus number.
It is the last bridge on this branch and so it is assigned a subordinate bus
interface number of 4. The initialisation code returns to PCI-PCI Bridge3 and
assigns it a subordinate bus number of 4. Finally, the PCI initialisation code
can assign 4 as the subordinate bus number for PCI-PCI Bridge1. Figure 6.9
on page 71 shows the nal bus numbers.
6.6.3 PCI BIOS Functions
The PCI BIOS functions are a series of standard routines which are common across
all platforms. For example, they are the same for both Intel and Alpha AXP based
systems. They allow the CPU controlled access to all of the PCI address spaces.
Only Linux kernel code and device drivers may use them.
See
arch/*/kernel/-
bios32.c
6.6. Linux PCI Initialization 71
DI D2
1
Bridge
DI D2
DI
DI D2
CPU
Bus 0
Bus 1
Bridge Bridge
3 2
Bus ?
Bridge
4
Bus 3 Bus 2
Subordinate=0xFF
Primary Bus = 0
Secondary Bus = 2
Subordinate=2
Primary Bus = 1
Secondary Bus = 2
Primary Bus = 1
Subordinate=0xFF
Secondary Bus = 3
Figure 6.8: Conguring a PCI System: Part 3
DI D2
1
Bridge
DI D2
DI
DI D2
CPU
Bus 0
Bus 1
Bridge Bridge
3 2
Bus 4
Bridge
4
Bus 3 Bus 2
Subordinate=4
Primary Bus = 0
Secondary Bus = 1
Subordinate=2
Primary Bus = 1
Secondary Bus = 2
Subordinate=4
Subordinate=4
Primary Bus = 1
Secondary Bus = 3
Primary Bus = 3
Secondary Bus = 4
Figure 6.9: Conguring a PCI System: Part 4
72 Chapter 6. PCI
1 0
0
31 4 3 2
Base Address
prefetchable Type
31 1 0
Reserved
Base Address 1
Base Address for PCI I/O Space
Base Address for PCI Memory Space
2
Figure 6.10: PCI Conguration Header: Base Address Registers
6.6.4 PCI Fixup
The PCI xup code for Alpha AXP does rather more than that for Intel (which
basically does nothing). For Intel based systems the system BIOS, which ran at
See
arch/*/kernel/-
bios32.c boot time, has already fully congured the PCI system. This leaves Linux with
little to do other than map that conguration. For non-Intel based systems further
conguration needs to happen to:
 Allocate PCI I/O and PCI Memory space to each device,
 Congure the PCI I/O and PCI Memory address windows for each PCI-PCI
bridge in the system,
 Generate Interrupt Line values for the devices; these control interrupt handling
for the device.
The next subsections describe how that code works.
Finding Out How Much PCI I/O and PCI Memory Space a Device Needs
Each PCI device found is queried to nd out how much PCI I/O and PCI Memory
address space it requires. To do this, each Base Address Register has all 1's written to
it and then read. The device will return 0's in the don't-care address bits, eectively
specifying the address space required.
There are two basic types of Base Address Register, the rst indicates within which
address space the devices registers must reside; either PCI I/O or PCI Memory space.
This is indicated by Bit 0 of the register. Figure 6.10 shows the two forms of the
Base Address Register for PCI Memory and for PCI I/O.
To nd out just how much of each address space a given Base Address Register is
requesting, you write all 1s into the register and then read it back. The device will
specify zeros in the don't care address bits, eectively specifying the address space
required. This design implies that all address spaces used are a power of two and are
naturally aligned.
6.6. Linux PCI Initialization 73
For example when you initialize the DECChip 21142 PCI Fast Ethernet device, it
tells you that it needs 0x100 bytes of space of either PCI I/O or PCI Memory. The
initialization code allocates it space. The moment that it allocates space, the 21142's
control and status registers can be seen at those addresses.
Allocating PCI I/O and PCI Memory to PCI-PCI Bridges and Devices
Like all memory the PCI I/O and PCI memory spaces are nite, and to some extent
scarce. The PCI Fixup code for non-Intel systems (and the BIOS code for Intel
systems) has to allocate each device the amount of memory that it is requesting in
an ecient manner. Both PCI I/O and PCI Memory must be allocated to a device
in a naturally aligned way. For example, if a device asks for 0xB0 of PCI I/O space
then it must be aligned on an address that is a multiple of 0xB0. In addition to this,
the PCI I/O and PCI Memory bases for any given bridge must be aligned on 4K and
on 1Mbyte boundaries respectively. Given that the address spaces for downstream
devices must lie within all of the upstream PCI-PCI Bridge's memory ranges for any
given device, it is a somewhat dicult problem to allocate space eciently.
The algorithm that Linux uses relies on each device described by the bus/device tree
built by the PCI Device Driver being allocated address space in ascending PCI I/O
memory order. Again a recursive algorithm is used to walk the pci bus and pci dev
data structures built by the PCI initialisation code. Starting at the root PCI bus
(pointed at by pci root) the BIOS xup code:
 Aligns the current global PCI I/O and Memory bases on 4K and 1 Mbyte
boundaries respectively,
 For every device on the current bus (in ascending PCI I/O memory needs),
{ allocates it space in PCI I/O and/or PCI Memory,
{ moves on the global PCI I/O and Memory bases by the appropriate
amounts,
{ enables the device's use of PCI I/O and PCI Memory,
 Allocates space recursively to all of the buses downstream of the current bus.
Note that this will change the global PCI I/O and Memory bases,
 Aligns the current global PCI I/O and Memory bases on 4K and 1 Mbyte
boundaries respectively and in doing so gure out the size and base of PCI I/O
and PCI Memory windows required by the current PCI-PCI bridge,
 Programs the PCI-PCI bridge that links to this bus with its PCI I/O and PCI
Memory bases and limits,
 Turns on bridging of PCI I/O and PCI Memory accesses in the PCI-PCI Bridge.
This means that if any PCI I/O or PCI Memory addresses seen on the Bridge's
primary PCI bus that are within its PCI I/O and PCI Memory address windows
will be bridged onto its secondary PCI bus.
Taking the PCI system in Figure 6.1 on page 62 as our example the PCI Fixup code
would set up the system in the following way:
74 Chapter 6. PCI
Align the PCI bases PCI I/O is 0x4000 and PCI Memory is 0x100000. This
allows the PCI-ISA bridges to translate all addresses below these into ISA
address cycles,
The Video Device This is asking for 0x200000 of PCI Memory and so we allocate
it that amount starting at the current PCI Memory base of 0x200000 as it has
to be naturally aligned to the size requested. The PCI Memory base is moved
to 0x400000 and the PCI I/O base remains at 0x4000.
The PCI-PCI Bridge We now cross the PCI-PCI Bridge and allocate PCI memory
there, note that we do not need to align the bases as they are already
correctly aligned:
The Ethernet Device This is asking for 0xB0 bytes of both PCI I/O and
PCI Memory space. It gets allocated PCI I/O at 0x4000 and PCI Memory
at 0x400000. The PCI Memory base is moved to 0x4000B0 and the PCI
I/O base to 0x40B0.
The SCSI Device This is asking for 0x1000 PCI Memory and so it is allocated
it at 0x401000 after it has been naturally aligned. The PCI I/O base
is still 0x40B0 and the PCI Memory base has been moved to 0x402000.
The PCI-PCI Bridge's PCI I/O and Memory Windows We now return to
the bridge and set its PCI I/O window at between 0x4000 and 0x40B0 and it's
PCI Memory window at between 0x400000 and 0x402000. This means that
the PCI-PCI Bridge will ignore the PCI Memory accesses for the video device
and pass them on if they are for the ethernet or SCSI devices.
Chapter 7
Interrupts and Interrupt
Handling
This chapter looks at how interrupts are handled by the Linux kernel.
Whilst the kernel has generic mechanisms and interfaces for handling interrupts,
most of the interrupt handling details are architecture specic.
Linux uses a lot of dierent pieces of hardware to perform many dierent tasks.
The video device drives the monitor, the IDE device drives the disks and so on.
You could drive these devices synchronously, that is you could send a request for
some operation (say writing a block of memory out to disk) and then wait for the
operation to complete. That method, although it would work, is very inecient and
the operating system would spend a lot of time \busy doing nothing" as it waited
for each operation to complete. A better, more ecient, way is to make the request
and then do other, more useful work and later be interrupted by the device when it
has nished the request. With this scheme, there may be many outstanding requests
to the devices in the system all happening at the same time.
There has to be some hardware support for the devices to interrupt whatever the
CPU is doing. Most, if not all, general purpose processors such as the Alpha AXP
use a similar method. Some of the physical pins of the CPU are wired such that
changing the voltage (for example changing it from +5v to -5v) causes the CPU to
stop what it is doing and to start executing special code to handle the interruption;
the interrupt handling code. One of these pins might be connected to an interval
timer and receive an interrupt every 1000th of a second, others may be connected to
the other devices in the system, such as the SCSI controller.
Systems often use an interrupt controller to group the device interrupts together
75
76 Chapter 7. Interrupts and Interrupt Handling
P
I
C
1
P
I
C
2
Keyboard
ide1
SCSI
5 sound
CPU
0
7
0
7
2
Real Time Clock
ide0
Serial
1
4
6 floppy
3
6
Figure 7.1: A Logical Diagram of Interrupt Routing
before passing on the signal to a single interrupt pin on the CPU. This saves interrupt
pins on the CPU and also gives
exibility when designing systems. The interrupt
controller has mask and status registers that control the interrupts. Setting the bits
in the mask register enables and disables interrupts and the status register returns
the currently active interrupts in the system.
Some of the interrupts in the system may be hard-wired, for example, the real time
clock's interval timer may be permanently connected to pin 3 on the interrupt controller.
However, what some of the pins are connected to may be determined by
what controller card is plugged into a particular ISA or PCI slot. For example, pin
4 on the interrupt controller may be connected to PCI slot number 0 which might
one day have an ethernet card in it but the next have a SCSI controller in it. The
bottom line is that each system has its own interrupt routing mechanisms and the
operating system must be
exible enough to cope.
Most modern general purpose microprocessors handle the interrupts the same way.
When a hardware interrupt occurs the CPU stops executing the instructions that it
was executing and jumps to a location in memory that either contains the interrupt
handling code or an instruction branching to the interrupt handling code. This code
usually operates in a special mode for the CPU, interrupt mode, and, normally, no
other interrupts can happen in this mode. There are exceptions though; some CPUs
rank the interrupts in priority and higher level interrupts may happen. This means
that the rst level interrupt handling code must be very carefully written and it
often has its own stack, which it uses to store the CPU's execution state (all of the
CPU's normal registers and context) before it goes o and handles the interrupt.
Some CPUs have a special set of registers that only exist in interrupt mode, and the
interrupt code can use these registers to do most of the context saving it needs to
do.
When the interrupt has been handled, the CPU's state is restored and the interrupt
is dismissed. The CPU will then continue to doing whatever it was doing before
being interrupted. It is important that the interrupt processing code is as ecient
7.1. Programmable Interrupt Controllers 77
as possible and that the operating system does not block interrupts too often or for
too long.
7.1 Programmable Interrupt Controllers
Systems designers are free to use whatever interrupt architecture they wish but IBM
PCs use the Intel 82C59A-2 CMOS Programmable Interrupt Controller [6, Intel
Peripheral Components] or its derivatives. This controller has been around since
the dawn of the PC and it is programmable with its registers being at well known
locations in the ISA address space. Even very modern support logic chip sets keep
equivalent registers in the same place in ISA memory. Non-Intel based systems such
as Alpha AXP based PCs are free from these architectural constraints and so often
use dierent interrupt controllers.
Figure 7.1 shows that there are two 8 bit controllers chained together; each having
a mask and an interrupt status register, PIC1 and PIC2. The mask registers are
at addresses 0x21 and 0xA1 and the status registers are at 0x20 and 0xA0 Writing
a one to a particular bit of the mask register enables an interrupt, writing a zero
disables it. So, writing one to bit 3 would enable interrupt 3, writing zero would
disable it. Unfortunately (and irritatingly), the interrupt mask registers are write
only, you cannot read back the value that you wrote. This means that Linux must
keep a local copy of what it has set the mask registers to. It modies these saved
masks in the interrupt enable and disable routines and writes the full masks to the
registers every time.
When an interrupt is signalled, the interrupt handling code reads the two interrupt
status registers (ISRs). It treats the ISR at 0x20 as the bottom eight bits of a sixteen
bit interrupt register and the ISR at 0xA0 as the top eight bits. So, an interrupt on
bit 1 of the ISR at 0xA0 would be treated as system interrupt 9. Bit 2 of PIC1 is
not available as this is used to chain interrupts from PIC2, any interrupt on PIC2
results in bit 2 of PIC1 being set.
7.2 Initializing the Interrupt Handling Data Structures
The kernel's interrupt handling data structures are set up by the device drivers as
they request control of the system's interrupts. To do this the device driver uses a
set of Linux kernel services that are used to request an interrupt, enable it and to
disable it. The individual device drivers call these routines to register their interrupt
See
request irq(),
enable irq() and
disable irq() in
arch/*/kernel
irq.c
handling routine addresses.
Some interrupts are xed by convention for the PC architecture and so the driver
simply requests its interrupt when it is initialized. This is what the
oppy disk device
driver does; it always requests IRQ 6. There may be occassions when a device driver
does not know which interrupt the device will use. This is not a problem for PCI
device drivers as they always know what their interrupt number is. Unfortunately
there is no easy way for ISA device drivers to nd their interrupt number. Linux
solves this problem by allowing device drivers to probe for their interrupts.
First, the device driver does something to the device that causes it to interrupt.
78 Chapter 7. Interrupts and Interrupt Handling
Then all of the unassigned interrupts in the system are enabled. This means that
the device's pending interrupt will now be delivered via the programmable interrupt
controller. Linux reads the interrupt status register and returns its contents to the
device driver. A non-zero result means that one or more interrupts occured during
the probe. The driver now turns probing o and the unassigned interrupts are all
disabled. If the ISA device driver has successfully found its IRQ number then it can
See
irq probe *() in
arch/*/kernel/-
irq.c now request control of it as normal.
PCI based systems are much more dynamic than ISA based systems. The interrupt
pin that an ISA device uses is often set using jumpers on the hardware device and
xed in the device driver. On the other hand, PCI devices have their interrupts
allocated by the PCI BIOS or the PCI subsystem as PCI is initialized when the
system boots. Each PCI device may use one of four interrupt pins, A, B, C or D.
This was xed when the device was built and most devices default to interrupt on
pin A. The PCI interrupt lines A, B, C and D for each PCI slot are routed to the
interrupt controller. So, Pin A from PCI slot 4 might be routed to pin 6 of the
interrupt controller, pin B of PCI slot 4 to pin 7 of the interrupt controller and so
on.
How the PCI interrupts are routed is entirely system specic and there must be
some set up code which understands this PCI interrupt routing topology. On Intel
based PCs this is the system BIOS code that runs at boot time but for system's
without BIOS (for example Alpha AXP based systems) the Linux kernel does this
setup. The PCI set up code writes the pin number of the interrupt controller into the
See
arch/alpha/-
kernel/bios32.c PCI conguration header for each device. It determines the interrupt pin (or IRQ)
number using its knowledge of the PCI interrupt routing topology together with the
devices PCI slot number and which PCI interrupt pin that it is using. The interrupt
pin that a device uses is xed and is kept in a eld in the PCI conguration header
for this device. It writes this information into the interrupt line eld that is reserved
for this purpose. When the device driver runs, it reads this information and uses it
to request control of the interrupt from the Linux kernel.
There may be many PCI interrupt sources in the system, for example when PCI-PCI
bridges are used. The number of interrupt sources may exceed the number of pins on
the system's programmable interrupt controllers. In this case, PCI devices may share
interrupts, one pin on the interrupt controller taking interrupts from more than one
PCI device. Linux supports this by allowing the rst requestor of an interrupt source
declare whether it may be shared. Sharing interrupts results in several irqaction
data structures being pointed at by one entry in the irq action vector vector.
When a shared interrupt happens, Linux will call all of the interrupt handlers for
that source. Any device driver that can share interrupts (which should be all PCI
device drivers) must be prepared to have its interrupt handler called when there is
no interrupt to be serviced.
7.3 Interrupt Handling
One of the principal tasks of Linux's interrupt handling subsystem is to route the
interrupts to the right pieces of interrupt handling code. This code must understand
the interrupt topology of the system. If, for example, the
oppy controller interrupts
7.3. Interrupt Handling 79
Interrupt
handling
routine
for this
device
flags
name
next
handler
irq_action
1
0
2
3
irqaction
irqaction
handler
flags
name
next
irqaction
handler
flags
name
next
Figure 7.2: Linux Interrupt Handling Data Structures
on pin 6 1 of the interrupt controller then it must recognize the interrupt as from the

oppy and route it to the
oppy device driver's interrupt handling code. Linux uses a
set of pointers to data structures containing the addresses of the routines that handle
the system's interrupts. These routines belong to the device drivers for the devices in
the system and it is the responsibility of each device driver to request the interrupt
that it wants when the driver is initialized. Figure 7.2 shows that irq action is a
vector of pointers to the irqaction data structure. Each irqaction data structure
contains information about the handler for this interrupt, including the address of the
interrupt handling routine. As the number of interrupts and how they are handled
varies between architectures and, sometimes, between systems, the Linux interrupt
handling code is architecture specic. This means that the size of the irq action
vector vector varies depending on the number of interrupt sources that there are.
When the interrupt happens, Linux must rst determine its source by reading the
interrupt status register of the system's programmable interrupt controllers. It then
translates that source into an oset into the irq action vector vector. So, for
example, an interrupt on pin 6 of the interrupt controller from the
oppy controller
would be translated into the seventh pointer in the vector of interrupt handlers. If
there is not an interrupt handler for the interrupt that occurred then the Linux kernel
will log an error, otherwise it will call into the interrupt handling routines for all of
the irqaction data structures for this interrupt source.
When the device driver's interrupt handling routine is called by the Linux kernel
it must eciently work out why it was interrupted and respond. To nd the cause
of the interrupt the device driver would read the status register of the device that
interrupted. The device may be reporting an error or that a requested operation has
completed. For example the
oppy controller may be reporting that it has completed
the positioning of the
oppy's read head over the correct sector on the
oppy disk.
Once the reason for the interrupt has been determined, the device driver may need to
do more work. If it does, the Linux kernel has mechanisms that allow it to postpone
that work until later. This avoids the CPU spending too much time in interrupt
mode. See the Device Driver chapter (Chapter 8) for more details.
REVIEW NOTE: Fast and slow interrupts, are these an Intel thing?
1Actually, the
oppy controller is one of the xed interrupts in a PC system as, by convention,
the
oppy controller is always wired to interrupt 6.
80 Chapter 7. Interrupts and Interrupt Handling
Chapter 8
Device Drivers
One of the purposes of an operating system is to hide the peculiarities of
the system's hardware devices from its users. For example the Virtual File
System presents a uniform view of the mounted lesystems irrespective
of the underlying physical devices. This chapter describes how the Linux
kernel manages the physical devices in the system.
The CPU is not the only intelligent device in the system, every physical device has
its own hardware controller. The keyboard, mouse and serial ports are controlled by
a SuperIO chip, the IDE disks by an IDE controller, SCSI disks by a SCSI controller
and so on. Each hardware controller has its own control and status registers (CSRs)
and these dier between devices. The CSRs for an Adaptec 2940 SCSI controller
are completely dierent from those of an NCR 810 SCSI controller. The CSRs are
used to start and stop the device, to initialize it and to diagnose any problems with
it. Instead of putting code to manage the hardware controllers in the system into
every application, the code is kept in the Linux kernel. The software that handles or
manages a hardware controller is known as a device driver. The Linux kernel device
drivers are, essentially, a shared library of privileged, memory resident, low level
hardware handling routines. It is Linux's device drivers that handle the peculiarities
of the devices they are managing.
One of the basic features of un?x is that it abstracts the handling of devices. All
hardware devices look like regular les; they can be opened, closed, read and written
using the same, standard, system calls that are used to manipulate les. Every device
in the system is represented by a device special le, for example the rst IDE disk in
the system is represented by /dev/hda. For block (disk) and character devices, these
device special les are created by the mknod command and they describe the device
using major and minor device numbers. Network devices are also represented by
81
82 Chapter 8. Device Drivers
device special les but they are created by Linux as it nds and initializes the network
controllers in the system. All devices controlled by the same device driver have a
common major device number. The minor device numbers are used to distinguish
between dierent devices and their controllers, for example each partition on the
primary IDE disk has a dierent minor device number. So, /dev/hda2, the second
partition of the primary IDE disk has a major number of 3 and a minor number of 2.
Linux maps the device special le passed in system calls (say to mount a le system
on a block device) to the device's device driver using the major device number and
a number of system tables, for example the character device See table, chrdevs .
fs/devices.c
Linux supports three types of hardware device: character, block and network. Character
devices are read and written directly without buering, for example the system's
serial ports /dev/cua0 and /dev/cua1. Block devices can only be written to and
read from in multiples of the block size, typically 512 or 1024 bytes. Block devices
are accessed via the buer cache and may be randomly accessed, that is to say, any
block can be read or written no matter where it is on the device. Block devices can
be accessed via their device special le but more commonly they are accessed via the
le system. Only a block device can support a mounted le system. Network devices
are accessed via the BSD socket interface and the networking subsytems described
in the Networking chapter (Chapter 10).
There are many dierent device drivers in the Linux kernel (that is one of Linux's
strengths) but they all share some common attributes:
kernel code Device drivers are part of the kernel and, like other code within the
kernel, if they go wrong they can seriously damage the system. A badly written
driver may even crash the system, possibly corrupting le systems and losing
data,
Kernel interfaces Device drivers must provide a standard interface to the Linux
kernel or to the subsystem that they are part of. For example, the terminal
driver provides a le I/O interface to the Linux kernel and a SCSI device driver
provides a SCSI device interface to the SCSI subsystem which, in turn, provides
both le I/O and buer cache interfaces to the kernel.
Kernel mechanisms and services Device drivers make use of standard kernel
services such as memory allocation, interrupt delivery and wait queues to operate,
Loadable Most of the Linux device drivers can be loaded on demand as kernel
modules when they are needed and unloaded when they are no longer being
used. This makes the kernel very adaptable and ecient with the system's
resources,
Congurable Linux device drivers can be built into the kernel. Which devices are
built is congurable when the kernel is compiled,
Dynamic As the system boots and each device driver is initialized it looks for the
hardware devices that it is controlling. It does not matter if the device being
controlled by a particular device driver does not exist. In this case the device
driver is simply redundant and causes no harm apart from occupying a little
of the system's memory.
8.1. Polling and Interrupts 83
8.1 Polling and Interrupts
Each time the device is given a command, for example \move the read head to sector
42 of the
oppy disk" the device driver has a choice as to how it nds out that the
command has completed. The device drivers can either poll the device or they can
use interrupts.
Polling the device usually means reading its status register every so often until the
device's status changes to indicate that it has completed the request. As a device
driver is part of the kernel it would be disasterous if a driver were to poll as nothing
else in the kernel would run until the device had completed the request. Instead
polling device drivers use system timers to have the kernel call a routine within the
device driver at some later time. This timer routine would check the status of the
command and this is exactly how Linux's
oppy driver works. Polling by means of
timers is at best approximate, a much more ecient method is to use interrupts.
An interrupt driven device driver is one where the hardware device being controlled
will raise a hardware interrupt whenever it needs to be serviced. For example, an
ethernet device driver would interrupt whenever it receives an ethernet packet from
the network. The Linux kernel needs to be able to deliver the interrupt from the
hardware device to the correct device driver. This is achieved by the device driver
registering its usage of the interrupt with the kernel. It registers the address of an
interrupt handling routine and the interrupt number that it wishes to own. You can
see which interrupts are being used by the device drivers, as well as how many of
each type of interrupts there have been, by looking at /proc/interrupts:
0: 727432 timer
1: 20534 keyboard
2: 0 cascade
3: 79691 + serial
4: 28258 + serial
5: 1 sound blaster
11: 20868 + aic7xxx
13: 1 math error
14: 247 + ide0
15: 170 + ide1
This requesting of interrupt resources is done at driver initialization time. Some of
the interrupts in the system are xed, this is a legacy of the IBM PC's architecture.
So, for example, the
oppy disk controller always uses interrupt 6. Other interrupts,
for example the interrupts from PCI devices are dynamically allocated at boot time.
In this case the device driver must rst discover the interrupt number (IRQ) of the
device that it is controlling before it requests ownership of that interrupt. For PCI
interrupts Linux supports standard PCI BIOS callbacks to determine information
about the devices in the system, including their IRQ numbers.
How an interrupt is delivered to the CPU itself is architecture dependent but on most
architectures the interrupt is delivered in a special mode that stops other interrupts
from happening in the system. A device driver should do as little as possible in its
interrupt handling routine so that the Linux kernel can dismiss the interrupt and
return to what it was doing before it was interrupted. Device drivers that need to
do a lot of work as a result of receiving an interrupt can use the kernel's bottom half
handlers or task queues to queue routines to be called later on.
84 Chapter 8. Device Drivers
8.2 Direct Memory Access (DMA)
Using interrupts driven device drivers to transfer data to or from hardware devices
works well when the amount of data is reasonably low. For example a 9600 baud
modem can transfer approximately one character every millisecond (1=1000'th second).
If the interrupt latency, the amount of time that it takes between the hardware
device raising the interrupt and the device driver's interrupt handling routine being
called, is low (say 2 milliseconds) then the overall system impact of the data transfer
is very low. The 9600 baud modem data transfer would only take 0.002% of the
CPU's processing time. For high speed devices, such as hard disk controllers or ethernet
devices the data transfer rate is a lot higher. A SCSI device can transfer up to
40 Mbytes of information per second.
Direct Memory Access, or DMA, was invented to solve this problem. A DMA controller
allows devices to transfer data to or from the system's memory without the
intervention of the processor. A PC's ISA DMA controller has 8 DMA channels of
which 7 are available for use by the device drivers. Each DMA channel has associated
with it a 16 bit address register and a 16 bit count register. To initiate a data transfer
the device driver sets up the DMA channel's address and count registers together
with the direction of the data transfer, read or write. It then tells the device that
it may start the DMA when it wishes. When the transfer is complete the device
interrupts the PC. Whilst the transfer is taking place the CPU is free to do other
things.
Device drivers have to be careful when using DMA. First of all the DMA controller
knows nothing of virtual memory, it only has access to the physical memory in the
system. Therefore the memory that is being DMA'd to or from must be a contiguous
block of physical memory. This means that you cannot DMA directly into the virtual
address space of a process. You can however lock the process's physical pages into
memory, preventing them from being swapped out to the swap device during a DMA
operation. Secondly, the DMA controller cannot access the whole of physical memory.
The DMA channel's address register represents the rst 16 bits of the DMA address,
the next 8 bits come from the page register. This means that DMA requests are
limited to the bottom 16 Mbytes of memory.
DMA channels are scarce resources, there are only 7 of them, and they cannot be
shared between device drivers. Just like interrupts, the device driver must be able
to work out which DMA channel it should use. Like interrupts, some devices have
a xed DMA channel. The
oppy device, for example, always uses DMA channel
2. Sometimes the DMA channel for a device can be set by jumpers; a number of
ethernet devices use this technique. The more
exible devices can be told (via their
CSRs) which DMA channels to use and, in this case, the device driver can simply
pick a free DMA channel to use.
Linux tracks the usage of the DMA channels using a vector of dma chan data structures
(one per DMA channel). The dma chan data structure contains just two elds,
a pointer to a string describing the owner of the DMA channel and a
ag indicating
if the DMA channel is allocated or not. It is this vector of dma chan data structures
that is printed when you cat /proc/dma.
8.3. Memory 85
8.3 Memory
Device drivers have to be careful when using memory. As they are part of the Linux
kernel they cannot use virtual memory. Each time a device driver runs, maybe as
an interrupt is received or as a bottom half or task queue handler is scheduled, the
current process may change. The device driver cannot rely on a particular process
running even if it is doing work on its behalf. Like the rest of the kernel, device
drivers use data structures to keep track of the device that it is controlling. These
data structures can be statically allocated, part of the device driver's code, but that
would be wasteful as it makes the kernel larger than it need be. Most device drivers
allocate kernel, non-paged, memory to hold their data.
Linux provides kernel memory allocation and deallocation routines and it is these
that the device drivers use. Kernel memory is allocated in chunks that are powers
of 2. For example 128 or 512 bytes, even if the device driver asks for less. The
number of bytes that the device driver requests is rounded up to the next block size
boundary. This makes kernel memory deallocation easier as the smaller free blocks
can be recombined into bigger blocks.
It may be that Linux needs to do quite a lot of extra work when the kernel memory
is requested. If the amount of free memory is low, physical pages may need to be discarded
or written to the swap device. Normally, Linux would suspend the requestor,
putting the process onto a wait queue until there is enough physical memory. Not
all device drivers (or indeed Linux kernel code) may want this to happen and so the
kernel memory allocation routines can be requested to fail if they cannot immediately
allocate memory. If the device driver wishes to DMA to or from the allocated
memory it can also specify that the memory is DMA'able. This way it is the Linux
kernel that needs to understand what constitutes DMA'able memory for this system,
and not the device driver.
8.4 Interfacing Device Drivers with the Kernel
The Linux kernel must be able to interact with them in standard ways. Each class
of device driver, character, block and network, provides common interfaces that the
kernel uses when requesting services from them. These common interfaces mean that
the kernel can treat often very dierent devices and their device drivers absolutely
the same. For example, SCSI and IDE disks behave very dierently but the Linux
kernel uses the same interface to both of them.
Linux is very dynamic, every time a Linux kernel boots it may encounter dierent
physical devices and thus need dierent device drivers. Linux allows you to include
device drivers at kernel build time via its conguration scripts. When these drivers
are initialized at boot time they may not discover any hardware to control. Other
drivers can be loaded as kernel modules when they are needed. To cope with this
dynamic nature of device drivers, device drivers register themselves with the kernel
as they are initialized. Linux maintains tables of registered device drivers as part of
its interfaces with them. These tables include pointers to routines and information
that support the interface with that class of devices.
86 Chapter 8. Device Drivers
name
fops
chrdevs
lseek
read
write
readdir
select
ioclt
mmap
open
release
fsyn
fasync
check_media_change
revalidate
file operations
Figure 8.1: Character Devices
8.4.1 Character Devices
Character devices, the simplest of Linux's devices, are accessed as les, applications
use standard system calls to open them, read from them, write to them and close
them exactly as if the device were a le. This is true even if the device is a modem
being used by the PPP daemon to connect a Linux system onto a network. As a
character device is initialized its device driver registers itself with the Linux kernel
by adding an entry into the chrdevs vector of device struct data structures. The
device's major device identier (for example 4 for the tty device) is used as an
index into this vector. The major device identier for a device is xed. Each entry
See
include/linux/-
major.h in the chrdevs vector, a device struct data structure contains two elements; a
pointer to the name of the registered device driver and a pointer to a block of le
operations. This block of le operations is itself the addresses of routines within the
device character device driver each of which handles specic le operations such as
open, read, write and close. The contents of /proc/devices for character devices is
taken from the chrdevs vector.
When a character special le representing a character device (for example /dev/cua0)
is opened the kernel must set things up so that the correct character device driver's
le operation routines will be called. Just like an ordinairy le or directory, each
device special le is represented by a VFS inode . The VFS inode for a character
special le, indeed for all device special les, contains both the major and minor
identiers for the device. This VFS inode was created by the underlying lesystem,
for example EXT2, from information in the real lesystem when the device special
le's name was looked up.
See
ext2 read inode()
in
fs/ext2/inode.c Each VFS inode has associated with it a set of le operations and these are dierent
depending on the lesystem object that the inode represents. Whenever a VFS
inode representing a character special le is created, its le operations are set to the
default character device operations . This has only one le operation, the def chr fops open le
operation. When the character special le is opened by an application the generic
open le operation uses the device's major identier as an index into the chrdevs
See
chrdev open() in
fs/devices.c vector to retrieve the le operations block for this particular device. It also sets up the
file data structure describing this character special le, making its le operations
pointer point to those of the device driver. Thereafter all of the applications le
operations will be mapped to calls to the character devices set of le operations.
8.4. Interfacing Device Drivers with the Kernel 87
request_fn()
current_request
:
:
sem
bh
next
tail
rq_status
rq_dev
mcd
request
sem
bh
next
tail
rq_status
rq_dev
mcd
request
0x0301
39
1024
b_dev
b_blocknr
b_state
b_count
b_size
b_prev
b_next
b_data
buffer_head
blk_dev_struct
blk_dev
Figure 8.2: Buer Cache Block Device Requests
8.4.2 Block Devices
Block devices also support being accessed like les. The mechanisms used to provide
the correct set of le operations for the opened block special le are very much the
same as for character devices. Linux maintains the set of registered block devices as
the blkdevs vector. It, like the chrdevs vector, is indexed using the device's major See
fs/devices.c
device number. Its entries are also device struct data structures. Unlike character
devices, there are classes of block devices. SCSI devices are one such class and IDE
devices are another. It is the class that registers itself with the Linux kernel and
provides le operations to the kernel. The device drivers for a class of block device
provide class specic interfaces to the class. So, for example, a SCSI device driver
has to provide interfaces to the SCSI subsystem which the SCSI subsystem uses to
provide le operations for this device to the kernel.
Every block device driver must provide an interface to the buer cache as well as
the normal le operations interface. Each block device driver lls in its entry in the
blk dev vector of blk dev struct data structures . The index into this vector is,
See
drivers/block/-
ll rw blk.c
See
include/linux/-
blkdev.h
again, the device's major number. The blk dev struct data structure consists of
the address of a request routine and a pointer to a list of request data structures,
each one representing a request from the buer cache for the driver to read or write
a block of data.
Each time the buer cache wishes to read or write a block of data to or from a registered
device it adds a request data structure onto its blk dev struct. Figure 8.2
shows that each request has a pointer to one or more buffer head data structures,
each one a request to read or write a block of data. The buffer head structures are
locked (by the buer cache) and there may be a process waiting on the block operation
to this buer to complete. Each request structure is allocated from a static
list, the all requests list. If the request is being added to an empty request list, the
driver's request function is called to start processing the request queue. Otherwise
the driver will simply process every request on the request list.
Once the device driver has completed a request it must remove each of the buffer head
structures from the request structure, mark them as up to date and unlock them.
88 Chapter 8. Device Drivers
This unlocking of the buffer head will wake up any process that has been sleeping
waiting for the block operation to complete. An example of this would be where
a le name is being resolved and the EXT2 lesystem must read the block of data
that contains the next EXT2 directory entry from the block device that holds the
lesystem. The process sleeps on the buffer head that will contain the directory
entry until the device driver wakes it up. The request data structure is marked as
free so that it can be used in another block request.
8.5 Hard Disks
Disk drives provide a more permanent method for storing data, keeping it on spinning
disk platters. To write data, a tiny head magnetizes minute particles on the platter's
surface. The data is read by a head, which can detect whether a particular minute
particle is magnetized.
A disk drive consists of one or more platters, each made of nely polished glass or
ceramic composites and coated with a ne layer of iron oxide. The platters are
attached to a central spindle and spin at a constant speed that can vary between
3000 and 10,000 RPM depending on the model. Compare this to a
oppy disk which
only spins at 360 RPM. The disk's read/write heads are responsible for reading and
writing data and there is a pair for each platter, one head for each surface. The
read/write heads do not physically touch the surface of the platters, instead they

oat on a very thin (10 millionths of an inch) cushion of air. The read/write heads
are moved across the surface of the platters by an actuator. All of the read/write
heads are attached together, they all move across the surfaces of the platters together.
Each surface of the platter is divided into narrow, concentric circles called tracks.
Track 0 is the outermost track and the highest numbered track is the track closest
to the central spindle. A cylinder is the set of all tracks with the same number. So
all of the 5th tracks from each side of every platter in the disk is known as cylinder
5. As the number of cylinders is the same as the number of tracks, you often see
disk geometries described in terms of cylinders. Each track is divided into sectors.
A sector is the smallest unit of data that can be written to or read from a hard disk
and it is also the disk's block size. A common sector size is 512 bytes and the sector
size was set when the disk was formatted, usually when the disk is manufactured.
A disk is usually described by its geometry, the number of cylinders, heads and
sectors. For example, at boot time Linux describes one of my IDE disks as:
hdb: Conner Peripherals 540MB - CFS540A, 516MB w/64kB Cache, CHS=1050/16/63
This means that it has 1050 cylinders (tracks), 16 heads (8 platters) and 63 sectors
per track. With a sector, or block, size of 512 bytes this gives the disk a storage
capacity of 529200 bytes. This does not match the disk's stated capacity of 516
Mbytes as some of the sectors are used for disk partitioning information. Some disks
automatically nd bad sectors and re-index the disk to work around them.
Hard disks can be further subdivided into partitions. A partition is a large group
of sectors allocated for a particular purpose. Partitioning a disk allows the disk to
be used by several operating system or for several purposes. A lot of Linux systems
have a single disk with three partitions; one containing a DOS lesystem, another
an EXT2 lesystem and a third for the swap partition. The partitions of a hard disk
8.5. Hard Disks 89
major
major_name
minor_shift
max_p
max_nr
init()
part
sizes
nr_real
real_devices
next
gendisk
8
"sd"
major
major_name
minor_shift
max_p
max_nr
init()
part
sizes
nr_real
real_devices
next
gendisk
start_sect
nr_sects
start_sect
nr_sects
:
:
:
hd_struct[]
"ide0"
3
gendisk_head
max_p
Figure 8.3: Linked list of disks
are described by a partition table; each entry describing where the partition starts
and ends in terms of heads, sectors and cylinder numbers. For DOS formatted disks,
those formatted by fdisk, there are four primary disk partitions. Not all four entries
in the partition table have to be used. There are three types of partition supported
by fdisk, primary, extended and logical. Extended partitions are not real partitions
at all, they contain any number of logical parititions. Extended and logical partitions
were invented as a way around the limit of four primary partitions. The following is
the output from fdisk for a disk containing two primary partitions:
Disk /dev/sda: 64 heads, 32 sectors, 510 cylinders
Units = cylinders of 2048 * 512 bytes
Device Boot Begin Start End Blocks Id System
/dev/sda1 1 1 478 489456 83 Linux native
/dev/sda2 479 479 510 32768 82 Linux swap
Expert command (m for help): p
Disk /dev/sda: 64 heads, 32 sectors, 510 cylinders
Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID
1 00 1 1 0 63 32 477 32 978912 83
2 00 0 1 478 63 32 509 978944 65536 82
3 00 0 0 0 0 0 0 0 0 00
4 00 0 0 0 0 0 0 0 0 00
This shows that the rst partition starts at cylinder or track 0, head 1 and sector 1
and extends to include cylinder 477, sector 32 and head 63. As there are 32 sectors in
a track and 64 read/write heads, this partition is a whole number of cylinders in size.
fdisk alligns partitions on cylinder boundaries by default. It starts at the outermost
cylinder (0) and extends inwards, towards the spindle, for 478 cylinders. The second
partition, the swap partition, starts at the next cylinder (478) and extends to the
innermost cylinder of the disk.
During initialization Linux maps the topology of the hard disks in the system. It
nds out how many hard disks there are and of what type. Additionally, Linux
90 Chapter 8. Device Drivers
discovers how the individual disks have been partitioned. This is all represented by a
list of gendisk data structures pointed at by the gendisk head list pointer. As each
disk subsystem, for example IDE, is initialized it generates gendisk data structures
representing the disks that it nds. It does this at the same time as it registers its le
operations and adds its entry into the blk dev data structure. Each gendisk data
structure has a unique major device number and these match the major numbers
of the block special devices. For example, the SCSI disk subsystem creates a single
gendisk entry (``sd'') with a major number of 8, the major number of all SCSI
disk devices. Figure 8.3 shows two gendisk entries, the rst one for the SCSI disk
subsystem and the second for an IDE disk controller. This is ide0, the primary IDE
controller.
Although the disk subsystems build the gendisk entries during their initialization
they are only used by Linux during partition checking. Instead, each disk subsystem
maintains its own data structures which allow it to map device special major and
minor device numbers to partitions within physical disks. Whenever a block device
is read from or written to, either via the buer cache or le operations, the kernel
directs the operation to the appropriate device using the major device number found
in its block special device le (for example /dev/sda2). It is the individual device
driver or subsystem that maps the minor device number to the real physical device.
8.5.1 IDE Disks
The most common disks used in Linux systems today are Integrated Disk Electronic
or IDE disks. IDE is a disk interface rather than an I/O bus like SCSI. Each IDE
controller can support up to two disks, one the master disk and the other the slave
disk. The master and slave functions are usually set by jumpers on the disk. The
rst IDE controller in the system is known as the primary IDE controller, the next
the secondary controller and so on. IDE can manage about 3.3 Mbytes per second
of data transfer to or from the disk and the maximum IDE disk size is 538Mbytes.
Extended IDE, or EIDE, has raised the disk size to a maximum of 8.6 Gbytes and the
data transfer rate up to 16.6 Mbytes per second. IDE and EIDE disks are cheaper
than SCSI disks and most modern PCs contain one or more on board IDE controllers.
Linux names IDE disks in the order in which it nds their controllers. The master
disk on the primary controller is /dev/hda and the slave disk is /dev/hdb. /dev/hdc
is the master disk on the secondary IDE controller. The IDE subsystem registers IDE
controllers and not disks with the Linux kernel. The major identier for the primary
IDE controller is 3 and is 22 for the secondary IDE controller. This means that if a
system has two IDE controllers there will be entries for the IDE subsystem at indices
at 3 and 22 in the blk dev and blkdevs vectors. The block special les for IDE disks
re
ect this numbering, disks /dev/hda and /dev/hdb, both connected to the primary
IDE controller, have a major identier of 3. Any le or buer cache operations for
the IDE subsystem operations on these block special les will be directed to the IDE
subsystem as the kernel uses the major identier as an index. When the request is
made, it is up to the IDE subsystem to work out which IDE disk the request is for.
To do this the IDE subsystem uses the minor device number from the device special
identier, this contains information that allows it to direct the request to the correct
partition of the correct disk. The device identier for /dev/hdb, the slave IDE drive
on the primary IDE controller is (3,64). The device identier for the rst partition
of that disk (/dev/hdb1) is (3,65).
8.5. Hard Disks 91
8.5.2 Initializing the IDE Subsystem
IDE disks have been around for much of the IBM PC's history. Throughout this
time the interface to these devices has changed. This makes the initialization of the
IDE subsystem more complex than it might at rst appear.
The maximum number of IDE controllers that Linux can support is 4. Each controller
is represented by an ide hwif t data structure in the ide hwifs vector. Each
ide hwif t data structure contains two ide drive t data structures, one per possible
supported master and slave IDE drive. During the initializing of the IDE
subsystem, Linux rst looks to see if there is information about the disks present
in the system's CMOS memory. This is battery backed memory that does not lose
its contents when the PC is powered o. This CMOS memory is actually in the
system's real time clock device which always runs no matter if your PC is on or o.
The CMOS memory locations are set up by the system's BIOS and tell Linux what
IDE controllers and drives have been found. Linux retrieves the found disk's geometry
from BIOS and uses the information to set up the ide hwif t data structure
for this drive. More modern PCs use PCI chipsets such as Intel's 82430 VX chipset
which includes a PCI EIDE controller. The IDE subsystem uses PCI BIOS callbacks
to locate the PCI (E)IDE controllers in the system. It then calls PCI specic
interrogation routines for those chipsets that are present.
Once each IDE interface or controller has been discovered, its ide hwif t is set up
to re
ect the controllers and attached disks. During operation the IDE driver writes
commands to IDE command registers that exist in the I/O memory space. The
default I/O address for the primary IDE controller's control and status registers is
0x1F0 - 0x1F7. These addresses were set by convention in the early days of the IBM
PC. The IDE driver registers each controller with the Linux block buer cache and
VFS, adding it to the blk dev and blkdevs vectors respectively. The IDE drive will
also request control of the appropriate interrupt. Again these interrupts are set by
convention to be 14 for the primary IDE controller and 15 for the secondary IDE
controller. However, they like all IDE details, can be overridden by command line
options to the kernel. The IDE driver also adds a gendisk entry into the list of
gendisk's discovered during boot for each IDE controller found. This list will later
be used to discover the partition tables of all of the hard disks found at boot time.
The partition checking code understands that IDE controllers may each control two
IDE disks.
8.5.3 SCSI Disks
The SCSI (Small Computer System Interface) bus is an ecient peer-to-peer data
bus that supports up to eight devices per bus, including one or more hosts. Each
device has to have a unique identier and this is usually set by jumpers on the disks.
Data can be transfered synchronously or asynchronously between any two devices
on the bus and with 32 bit wide data transfers up to 40 Mbytes per second are
possible. The SCSI bus transfers both data and state information between devices,
and a single transaction between an initiator and a target can involve up to eight
distinct phases. You can tell the current phase of a SCSI bus from ve signals from
the bus. The eight phases are:
BUS FREE No device has control of the bus and there are no transactions currently
happening,
92 Chapter 8. Device Drivers
ARBITRATION A SCSI device has attempted to get control of the SCSI bus, it
does this by asserting its SCSI identifer onto the address pins. The highest
number SCSI identier wins.
SELECTION When a device has succeeded in getting control of the SCSI bus
through arbitration it must now signal the target of this SCSI request that it
wants to send a command to it. It does this by asserting the SCSI identier of
the target on the address pins.
RESELECTION SCSI devices may disconnect during the processing of a request.
The target may then reselect the initiator. Not all SCSI devices support this
phase.
COMMAND 6,10 or 12 bytes of command can be transfered from the initiator to
the target,
DATA IN, DATA OUT During these phases data is transfered between the initiator
and the target,
STATUS This phase is entered after completion of all commands and allows the
target to send a status byte indicating success or failure to the initiator,
MESSAGE IN, MESSAGE OUT Additional information is transfered between
the initiator and the target.
The Linux SCSI subsystem is made up of two basic elements, each of which is represented
by data structures:
host A SCSI host is a physical piece of hardware, a SCSI controller. The NCR810
PCI SCSI controller is an example of a SCSI host. If a Linux system has more
than one SCSI controller of the same type, each instance will be represented by
a separate SCSI host. This means that a SCSI device driver may control more
than one instance of its controller. SCSI hosts are almost always the initiator
of SCSI commands.
Device The most common set of SCSI device is a SCSI disk but the SCSI standard
supports several more types; tape, CD-ROM and also a generic SCSI device.
SCSI devices are almost always the targets of SCSI commands. These devices
must be treated dierently, for example with removable media such as CDROMs
or tapes, Linux needs to detect if the media was removed. The dierent
disk types have dierent major device numbers, allowing Linux to direct block
device requests to the appropriate SCSI type.
Initializing the SCSI Subsystem
Initializing the SCSI subsystem is quite complex, re
ecting the dynamic nature of
SCSI buses and their devices. Linux initializes the SCSI subsystem at boot time; it
nds the SCSI controllers (known as SCSI hosts) in the system and then probes each
of their SCSI buses nding all of their devices. It then initializes those devices and
makes them available to the rest of the Linux kernel via the normal le and buer
cache block device operations. This initialization is done in four phases:
First, Linux nds out which of the SCSI host adapters, or controllers, that were
built into the kernel at kernel build time have hardware to control. Each built in
8.5. Hard Disks 93
Device
Driver
Routines
name
Scsi_Host_Template
next
"Buslogic"
next
id
type
host
Scsi_Device
next
id
type
host
Scsi_Device
next
this_id
max_id
Scsi_Host
hostt
scsi_devices
scsi_hostlist
scsi_hosts
Figure 8.4: SCSI Data Structures
SCSI host has a Scsi Host Template entry in the builtin scsi hosts vector The
Scsi Host Template data structure contains pointers to routines that carry out SCSI
host specic actions such as detecting what SCSI devices are attached to this SCSI
host. These routines are called by the SCSI subsystem as it congures itself and they
are part of the SCSI device driver supporting this host type. Each detected SCSI host,
those for which there are real SCSI devices attached, has its Scsi Host Template
data structure added to the scsi hosts list of active SCSI hosts. Each instance
of a detected host type is represented by a Scsi Host data structure held in the
scsi hostlist list. For example a system with two NCR810 PCI SCSI controllers
would have two Scsi Host entries in the list, one per controller. Each Scsi Host
points at the Scsi Host Template representing its device driver.
Now that every SCSI host has been discovered, the SCSI subsystem must nd out
what SCSI devices are attached to each host's bus. SCSI devices are numbered between
0 and 7 inclusively, each device's number or SCSI identier being unique on
the SCSI bus to which it is attached. SCSI identiers are usually set by jumpers on
the device. The SCSI initialization code nds each SCSI device on a SCSI bus by
sending it a TEST UNIT READY command. When a device responds, its identi-
cation is read by sending it an ENQUIRY command. This gives Linux the vendor's
name and the device's model and revision names. SCSI commands are represented
by a Scsi Cmnd data structure and these are passed to the device driver for this
SCSI host by calling the device driver routines within its Scsi Host Template data
structure. Every SCSI device that is found is represented by a Scsi Device data
94 Chapter 8. Device Drivers
structure, each of which points to its parent Scsi Host. All of the Scsi Device data
structures are added to the scsi devices list. Figure 8.4 shows how the main data
structures relate to one another.
There are four SCSI device types: disk, tape, CD and generic. Each of these SCSI
types are individually registered with the kernel as dierent major block device types.
However they will only register themselves if one or more of a given SCSI device
type has been found. Each SCSI type, for example SCSI disk, maintains its own
tables of devices. It uses these tables to direct kernel block operations (le or buer
cache) to the correct device driver or SCSI host. Each SCSI type is represented by
a Scsi Device Template data structure. This contains information about this type
of SCSI device and the addresses of routines to perform various tasks. The SCSI
subsystem uses these templates to call the SCSI type routines for each type of SCSI
device. In other words, if the SCSI subsystem wishes to attach a SCSI disk device it
will call the SCSI disk type attach routine. The Scsi Type Template data structures
are added to the scsi devicelist list if one or more SCSI devices of that type have
been detected.
The nal phase of the SCSI subsystem initialization is to call the nish functions for
each registered Scsi Device Template. For the SCSI disk type this spins up all of
the SCSI disks that were found and then records their disk geometry. It also adds the
gendisk data structure representing all SCSI disks to the linked list of disks shown
in Figure 8.3.
Delivering Block Device Requests
Once Linux has initialized the SCSI subsystem, the SCSI devices may be used. Each
active SCSI device type registers itself with the kernel so that Linux can direct
block device requests to it. There can be buer cache requests via blk dev or le
operations via blkdevs. Taking a SCSI disk driver that has one or more EXT2
lesystem partitions as an example, how do kernel buer requests get directed to the
right SCSI disk when one of its EXT2 partitions is mounted?
Each request to read or write a block of data to or from a SCSI disk partition results
in a new request structure being added to the SCSI disks current request list in
the blk dev vector. If the request list is being processed, the buer cache need not
do anything else; otherwise it must nudge the SCSI disk subsystem to go and process
its request queue. Each SCSI disk in the system is represented by a Scsi Disk data
structure. These are kept in the rscsi disks vector that is indexed using part of the
SCSI disk partition's minor device number. For exmaple, /dev/sdb1 has a major
number of 8 and a minor number of 17; this generates an index of 1. Each Scsi Disk
data structure contains a pointer to the Scsi Device data structure representing
this device. That in turn points at the Scsi Host data structure which \owns" it.
The request data structures from the buer cache are translated into Scsi Cmd
structures describing the SCSI command that needs to be sent to the SCSI device
and this is queued onto the Scsi Host structure representing this device. These will
be processed by the individual SCSI device driver once the appropriate data blocks
have been read or written.
8.6. Network Devices 95
8.6 Network Devices
A network device is, so far as Linux's network subsystem is concerned, an entity that
sends and receives packets of data. This is normally a physical device such as an
ethernet card. Some network devices though are software only such as the loopback
device which is used for sending data to yourself. Each network device is represented
by a device data structure. Network device drivers register the devices that they
See
include/linux/-
control with Linux during network initialization at kernel boot time. The device netdevice.h
data structure contains information about the device and the addresses of functions
that allow the various supported network protocols to use the device's services. These
functions are mostly concerned with transmitting data using the network device. The
device uses standard networking support mechanisms to pass received data up to the
appropriate protocol layer. All network data (packets) transmitted and received
are represented by sk buff data structures, these are
exible data structures that
allow network protocol headers to be easily added and removed. How the network
protocol layers use the network devices, how they pass data back and forth using
sk buff data structures is described in detail in the Networks chapter (Chapter 10).
This chapter concentrates on the device data structure and on how network devices
are discovered and initialized.
The device data structure contains information about the network device:
Name Unlike block and character devices which have their device special les created
using the mknod command, network device special les appear spontaniously
as the system's network devices are discovered and initialized. Their
names are standard, each name representing the type of device that it is. Multiple
devices of the same type are numbered upwards from 0. Thus the ethernet
devices are known as /dev/eth0,/dev/eth1,/dev/eth2 and so on. Some common
network devices are:
/dev/ethN Ethernet devices
/dev/slN SLIP devices
/dev/pppN PPP devices
/dev/lo Loopback devices
Bus Information This is information that the device driver needs in order to control
the device. The irq number is the interrupt that this device is using. The
base address is the address of any of the device's control and status registers
in I/O memory. The DMA channel is the DMA channel number that this network
device is using. All of this information is set at boot time as the device
is initialized.
Interface Flags These describe the characteristics and abilities of the network device:
96 Chapter 8. Device Drivers
IFF UP Interface is up and running,
IFF BROADCAST Broadcast address in device is valid
IFF DEBUG Device debugging turned on
IFF LOOPBACK This is a loopback device
IFF POINTTOPOINT This is point to point link (SLIP and PPP)
IFF NOTRAILERS No network trailers
IFF RUNNING Resources allocated
IFF NOARP Does not support ARP protocol
IFF PROMISC Device in promiscuous receive mode, it will receive
all packets no matter who they are addressed to
IFF ALLMULTI Receive all IP multicast frames
IFF MULTICAST Can receive IP multicast frames
Protocol Information Each device describes how it may be used by the network
protocool layers:
mtu The size of the largest packet that this network can transmit not including
any link layer headers that it needs to add. This maximum is used by the
protocol layers, for example IP, to select suitable packet sizes to send.
Family The family indicates the protocol family that the device can support.
The family for all Linux network devices is AF INET, the Internet address
family.
Type The hardware interface type describes the media that this network device
is attached to. There are many dierent types of media that Linux
network devices support. These include Ethernet, X.25, Token Ring, Slip,
PPP and Apple Localtalk.
Addresses The device data structure holds a number of addresses that are
relevent to this network device, including its IP addresses.
Packet Queue This is the queue of sk buff packets queued waiting to be transmitted
on this network device,
Support Functions Each device provides a standard set of routines that protocol
layers call as part of their interface to this device's link layer. These include
setup and frame transmit routines as well as routines to add standard frame
headers and collect statistics. These statistics can be seen using the ifcong
command.
8.6.1 Initializing Network Devices
Network device drivers can, like other Linux device drivers, be built into the Linux
kernel. Each potential network device is represented by a device data structure
within the network device list pointed at by dev base list pointer. The network layers
call one of a number of network device service routines whose addresses are held in
the device data structure if they need device specic work performing. Initially
though, each device data structure holds only the address of an initialization or
probe routine.
There are two problems to be solved for network device drivers. Firstly, not all of
the network device drivers built into the Linux kernel will have devices to control.
Secondly, the ethernet devices in the system are always called /dev/eth0, /dev/eth1
8.6. Network Devices 97
and so on, no matter what their underlying device drivers are. The problem of
\missing" network devices is easily solved. As the initialization routine for each
network device is called, it returns a status indicating whether or not it located an
instance of the controller that it is driving. If the driver could not nd any devices, its
entry in the device list pointed at by dev base is removed. If the driver could nd
a device it lls out the rest of the device data structure with information about the
device and the addresses of the support functions within the network device driver.
The second problem, that of dynamically assigning ethernet devices to the standard
/dev/ethN device special les is solved more elegantly. There are eight standard
entries in the devices list; one for eth0, eth1 and so on to eth7. The initialization
routine is the same for all of them, it tries each ethernet device driver built into the
kernel in turn until one nds a device. When the driver nds its ethernet device it
lls out the ethN device data structure, which it now owns. It is also at this time
that the network device driver initializes the physical hardware that it is controlling
and works out which IRQ it is using, which DMA channel (if any) and so on. A
driver may nd several instances of the network device that it is controlling and, in
this case, it will take over several of the /dev/ethN device data structures. Once
all eight standard /dev/ethN have been allocated, no more ethernet devices will be
probed for.
98 Chapter 8. Device Drivers
Chapter 9
The File system
This chapter describes how the Linux kernel maintains the les in the le
systems that it supports. It describes the Virtual File System (VFS) and
explains how the Linux kernel's real le systems are supported.
One of the most important features of Linux is its support for many dierent le
systems. This makes it very
exible and well able to coexist with many other operating
systems. At the time of writing, Linux supports 15 le systems; ext, ext2,
xia, minix, umsdos, msdos, vfat, proc, smb, ncp, iso9660, sysv, hpfs, affs and
ufs, and no doubt, over time more will be added.
In Linux, as it is for UnixTM, the separate le systems the system may use are not
accessed by device identiers (such as a drive number or a drive name) but instead
they are combined into a single hierarchical tree structure that represents the le
system as one whole single entity. Linux adds each new le system into this single
le system tree as it is mounted. All le systems, of whatever type, are mounted onto
a directory and the les of the mounted le system cover up the existing contents
of that directory. This directory is known as the mount directory or mount point.
When the le system is unmounted, the mount directory's own les are once again
revealed.
When disks are initialized (using fdisk, say) they have a partition structure imposed
on them that divides the physical disk into a number of logical partitions. Each
partition may hold a single le system, for example an EXT2 le system. File systems
organize les into logical hierarchical structures with directories, soft links and so on
held in blocks on physical devices. Devices that can contain le systems are known
as block devices. The IDE disk partition /dev/hda1, the rst partition of the rst
IDE disk drive in the system, is a block device. The Linux le systems regard these
99
100 Chapter 9. The File system
block devices as simply linear collections of blocks, they do not know or care about
the underlying physical disk's geometry. It is the task of each block device driver to
map a request to read a particular block of its device into terms meaningful to its
device; the particular track, sector and cylinder of its hard disk where the block is
kept. A le system has to look, feel and operate in the same way no matter what
device is holding it. Moreover, using Linux's le systems, it does not matter (at least
to the system user) that these dierent le systems are on dierent physical media
controlled by dierent hardware controllers. The le system might not even be on
the local system, it could just as well be a disk remotely mounted over a network
link. Consider the following example where a Linux system has its root le system
on a SCSI disk:
A E boot etc lib opt tmp usr
C F cdrom fd proc root var sbin
D bin dev home mnt lost+found
Neither the users nor the programs that operate on the les themselves need know
that /C is in fact a mounted VFAT le system that is on the rst IDE disk in the
system. In the example (which is actually my home Linux system), /E is the master
IDE disk on the second IDE controller. It does not matter either that the rst IDE
controller is a PCI controller and that the second is an ISA controller which also
controls the IDE CDROM. I can dial into the network where I work using a modem
and the PPP network protocol using a modem and in this case I can remotely mount
my Alpha AXP Linux system's le systems on /mnt/remote.
The les in a le system are collections of data; the le holding the sources to this
chapter is an ASCII le called filesystems.tex. A le system not only holds the
data that is contained within the les of the le system but also the structure of
the le system. It holds all of the information that Linux users and processes see as
les, directories soft links, le protection information and so on. Moreover it must
hold that information safely and securely, the basic integrity of the operating system
depends on its le systems. Nobody would use an operating system that randomly
lost data and les1.
Minix, the rst le system that Linux had is rather restrictive and lacking in performance.
Its lenames cannot be longer than 14 characters (which is still better
than 8.3 lenames) and the maximum le size is 64MBytes. 64Mbytes might at
rst glance seem large enough but large le sizes are necessary to hold even modest
databases. The rst le system designed specically for Linux, the Extended File
system, or EXT, was introduced in April 1992 and cured a lot of the problems but it
was still felt to lack performance. So, in 1993, the Second Extended File system,
or EXT2, was added. It is this le system that is described in detail later on in this
chapter.
An important development took place when the EXT le system was added into
Linux. The real le systems were separated from the operating system and system
services by an interface layer known as the Virtual File system, or VFS. VFS allows
Linux to support many, often very dierent, le systems, each presenting a common
software interface to the VFS. All of the details of the Linux le systems are translated
by software so that all le systems appear identical to the rest of the Linux kernel
1Well, not knowingly, although I have been bitten by operating systems with more lawyers than
Linux has developers
9.1. The Second Extended File system (EXT2) 101
Data
Blocks
Block
Group N-1
Block
Group 0
Block
Group N
Block
Bitmap
Group
Descriptors
Super
Block
Inode
Table
Inode
Bitmap
Figure 9.1: Physical Layout of the EXT2 File system
and to programs running in the system. Linux's Virtual File system layer allows you
to transparently mount the many dierent le systems at the same time.
The Linux Virtual File system is implemented so that access to its les is as fast and
ecient as possible. It must also make sure that the les and their data are kept
correctly. These two requirements can be at odds with each other. The Linux VFS
caches information in memory from each le system as it is mounted and used. A
lot of care must be taken to update the le system correctly as data within these
caches is modied as les and directories are created, written to and deleted. If you
could see the le system's data structures within the running kernel, you would be
able to see data blocks being read and written by the le system. Data structures,
describing the les and directories being accessed would be created and destroyed
and all the time the device drivers would be working away, fetching and saving data.
The most important of these caches is the Buer Cache, which is integrated into
the way that the individual le systems access their underlying block devices. As
blocks are accessed they are put into the Buer Cache and kept on various queues
depending on their states. The Buer Cache not only caches data buers, it also
helps manage the asynchronous interface with the block device drivers.
9.1 The Second Extended File system (EXT2)
The Second Extended File system was devised (by Remy Card) as an extensible and
powerful le system for Linux. It is also the most successful le system so far in the
Linux community and is the basis for all of the currently shipping Linux distributions.
The EXT2 le system, like a lot of the le systems, is built on the premise that the See
fs/ext2/*
data held in les is kept in data blocks. These data blocks are all of the same length
and, although that length can vary between dierent EXT2 le systems the block
size of a particular EXT2 le system is set when it is created (using mke2fs). Every
le's size is rounded up to an integral number of blocks. If the block size is 1024
bytes, then a le of 1025 bytes will occupy two 1024 byte blocks. Unfortunately this
means that on average you waste half a block per le. Usually in computing you
trade o CPU usage for memory and disk space utilisation. In this case Linux, along
with most operating systems, trades o a relatively inecient disk usage in order to
reduce the workload on the CPU. Not all of the blocks in the le system hold data,
some must be used to contain the information that describes the structure of the le
system. EXT2 denes the le system topology by describing each le in the system
with an inode data structure. An inode describes which blocks the data within a
102 Chapter 9. The File system
Data
Data
Data
Data
Data
Data
Data
Data
Direct Blocks
Double Indirect
Mode
Size
Timestamps
Indirect blocks
Triple Indirect
Owner info
ext2_inode
Figure 9.2: EXT2 Inode
le occupies as well as the access rights of the le, the le's modication times and
the type of the le. Every le in the EXT2 le system is described by a single inode
and each inode has a single unique number identifying it. The inodes for the le
system are all kept together in inode tables. EXT2 directories are simply special
les (themselves described by inodes) which contain pointers to the inodes of their
directory entries.
Figure 9.1 shows the layout of the EXT2 le system as occupying a series of blocks in
a block structured device. So far as each le system is concerned, block devices are
just a series of blocks that can be read and written. A le system does not need to
concern itself with where on the physical media a block should be put, that is the job
of the device's driver. Whenever a le system needs to read information or data from
the block device containing it, it requests that its supporting device driver reads an
integral number of blocks. The EXT2 le system divides the logical partition that
it occupies into Block Groups. Each group duplicates information critical to the
integrity of the le system as well as holding real les and directories as blocks of
information and data. This duplication is neccessary should a disaster occur and the
le system need recovering. The subsections describe in more detail the contents of
each Block Group.
9.1.1 The EXT2 Inode
In the EXT2 le system, the inode is the basic building block; every le and directory
in the le system is described by one and only one inode. The EXT2 inodes for
each Block Group are kept in the inode table together with a bitmap that allows
the system to keep track of allocated and unallocated inodes. Figure 9.2 shows the
format of an EXT2 inode, amongst other information, it contains the following elds: See
include/linux/-
ext2 fs i.h
9.1. The Second Extended File system (EXT2) 103
mode This holds two pieces of information; what this inode describes and the permissions
that users have to it. For EXT2, an inode can describe one of le,
directory, symbolic link, block device, character device or FIFO.
Owner Information The user and group identiers of the owners of this le or
directory. This allows the le system to correctly allow the right sort of accesses,
Size The size of the le in bytes,
Timestamps The time that the inode was created and the last time that it was
modied,
Datablocks Pointers to the blocks that contain the data that this inode is describing.
The rst twelve are pointers to the physical blocks containing the data
described by this inode and the last three pointers contain more and more levels
of indirection. For example, the double indirect blocks pointer points at a
block of pointers to blocks of pointers to data blocks. This means that les less
than or equal to twelve data blocks in length are more quickly accessed than
larger les.
You should note that EXT2 inodes can describe special device les. These are not
real les but handles that programs can use to access devices. All of the device les
in /dev are there to allow programs to access Linux's devices. For example the mount
program takes as an argument the device le that it wishes to mount.
9.1.2 The EXT2 Superblock
The Superblock contains a description of the basic size and shape of this le system.
The information within it allows the le system manager to use and maintain the le
system. Usually only the Superblock in Block Group 0 is read when the le system
is mounted but each Block Group contains a duplicate copy in case of le system
corruption. Amongst other information it holds the:
See
include/linux/-
ext2 fs sb.h
Magic Number This allows the mounting software to check that this is indeed the
Superblock for an EXT2 le system. For the current version of EXT2 this is
0xEF53.
Revision Level The major and minor revision levels allow the mounting code to determine
whether or not this le system supports features that are only available
in particular revisions of the le system. There are also feature compatibility
elds which help the mounting code to determine which new features can safely
be used on this le system,
Mount Count and Maximum Mount Count Together these allow the system
to determine if the le system should be fully checked. The mount count
is incremented each time the le system is mounted and when it equals the
maximum mount count the warning message \maximal mount count reached,
running e2fsck is recommended" is displayed,
Block Group Number The Block Group number that holds this copy of the Superblock,
Block Size The size of the block for this le system in bytes, for example 1024
bytes,
104 Chapter 9. The File system
Blocks per Group The number of blocks in a group. Like the block size this is
xed when the le system is created,
Free Blocks The number of free blocks in the le system,
Free Inodes The number of free Inodes in the le system,
First Inode This is the inode number of the rst inode in the le system. The
rst inode in an EXT2 root le system would be the directory entry for the '/'
directory.
9.1.3 The EXT2 Group Descriptor
Each Block Group has a data structure describing it. Like the Superblock, all the
group descriptors for all of the Block Groups are duplicated in each Block Group
in case of le system corruption. Each Group Descriptor contains the following
See
ext2 group desc
in include/-
linux/ext2 fs.h information:
Blocks Bitmap The block number of the block allocation bitmap for this Block
Group. This is used during block allocation and deallocation,
Inode Bitmap The block number of the inode allocation bitmap for this Block
Group. This is used during inode allocation and deallocation,
Inode Table The block number of the starting block for the inode table for this
Block Group. Each inode is represented by the EXT2 inode data structure
described below.
Free blocks count, Free Inodes count, Used directory count
The group descriptors are placed on after another and together they make the group
descriptor table. Each Blocks Group contains the entire table of group descriptors
after its copy of the Superblock. Only the rst copy (in Block Group 0) is actually
used by the EXT2 le system. The other copies are there, like the copies of the
Superblock, in case the main copy is corrupted.
9.1.4 EXT2 Directories
In the EXT2 le system, directories are special les that are used to create and hold
access paths to the les in the le system. Figure 9.3 shows the layout of a directory
entry in memory. A directory le is a list of directory entries, each one containing
See
ext2 dir entry
in include/-
linux/ext2 fs.h the following information:
inode The inode for this directory entry. This is an index into the array of inodes
held in the Inode Table of the Block Group. In gure 9.3, the directory entry
for the le called file has a reference to inode number i1,
name length The length of this directory entry in bytes,
name The name of this directory entry.
The rst two entries for every directory are always the standard \." and \.." entries
meaning \this directory" and \the parent directory" respectively.
9.1. The Second Extended File system (EXT2) 105
i1 15 file i2 5 40 14 very_long_name
0 15 55
inode table
Figure 9.3: EXT2 Directory
9.1.5 Finding a File in an EXT2 File System
A Linux lename has the same format as all UnixTM lenames have. It is a series of
directory names separated by forward slashes (\/") and ending in the le's name.
One example lename would be /home/rusling/.cshrc where /home and /rusling
are directory names and the le's name is .cshrc. Like all other UnixTM systems,
Linux does not care about the format of the lename itself; it can be any length and
consist of any of the printable characters. To nd the inode representing this le
within an EXT2 le system the system must parse the lename a directory at a time
until we get to the le itself.
The rst inode we need is the inode for the root of the le system and we nd its
number in the le system's superblock. To read an EXT2 inode we must look for it
in the inode table of the appropriate Block Group. If, for example, the root inode
number is 42, then we need the 42nd inode from the inode table of Block Group 0.
The root inode is for an EXT2 directory, in other words the mode of the root inode
describes it as a directory and it's data blocks contain EXT2 directory entries.
home is just one of the many directory entries and this directory entry gives us the
number of the inode describing the /home directory. We have to read this directory
(by rst reading its inode and then reading the directory entries from the data
blocks described by its inode) to nd the rusling entry which gives us the number
of the inode describing the /home/rusling directory. Finally we read the directory
entries pointed at by the inode describing the /home/rusling directory to nd the
inode number of the .cshrc le and from this we get the data blocks containing the
information in the le.
9.1.6 Changing the Size of a File in an EXT2 File System
One common problem with a le system is its tendency to fragment. The blocks that
hold the le's data get spread all over the le system and this makes sequentially
106 Chapter 9. The File system
accessing the data blocks of a le more and more inecient the further apart the
data blocks are. The EXT2 le system tries to overcome this by allocating the new
blocks for a le physically close to its current data blocks or at least in the same
Block Group as its current data blocks. Only when this fails does it allocate data
blocks in another Block Group.
Whenever a process attempts to write data into a le the Linux le system checks
to see if the data has gone o the end of the le's last allocated block. If it has, then
it must allocate a new data block for this le. Until the allocation is complete, the
process cannot run; it must wait for the le system to allocate a new data block and
write the rest of the data to it before it can continue. The rst thing that the EXT2
block allocation routines do is to lock the EXT2 Superblock for this le system.
Allocating and deallocating changes elds within the superblock, and the Linux le
system cannot allow more than one process to do this at the same time. If another
process needs to allocate more data blocks, it will have to wait until this process has
nished. Processes waiting for the superblock are suspended, unable to run, until
control of the superblock is relinquished by its current user. Access to the superblock
is granted on a rst come, rst served basis and once a process has control of the
superblock, it keeps control until it has nished. Having locked the superblock, the
process checks that there are enough free blocks left in this le system. If there are
not enough free blocks, then this attempt to allocate more will fail and the process
will relinquish control of this le system's superblock.
If there are enough free blocks in the le system, the process tries to allocate one.
If the EXT2 le system has been built to preallocate data blocks then we may
See
ext2 new block()
in fs/ext2/-
balloc.c be able to take one of those. The preallocated blocks do not actually exist, they
are just reserved within the allocated block bitmap. The VFS inode representing
the le that we are trying to allocate a new data block for has two EXT2 specic
elds, prealloc block and prealloc count, which are the block number of the rst
preallocated data block and how many of them there are, respectively. If there were
no preallocated blocks or block preallocation is not enabled, the EXT2 le system
must allocate a new block. The EXT2 le system rst looks to see if the data block
after the last data block in the le is free. Logically, this is the most ecient block
to allocate as it makes sequential accesses much quicker. If this block is not free,
then the search widens and it looks for a data block within 64 blocks of the of the
ideal block. This block, although not ideal is at least fairly close and within the same
Block Group as the other data blocks belonging to this le.
If even that block is not free, the process starts looking in all of the other Block
Groups in turn until it nds some free blocks. The block allocation code looks for a
cluster of eight free data blocks somewhere in one of the Block Groups. If it cannot
nd eight together, it will settle for less. If block preallocation is wanted and enabled
it will update prealloc block and prealloc count accordingly.
Wherever it nds the free block, the block allocation code updates the Block Group's
block bitmap and allocates a data buer in the buer cache. That data buer is
uniquely identied by the le system's supporting device identier and the block
number of the allocated block. The data in the buer is zero'd and the buer is
marked as \dirty" to show that it's contents have not been written to the physical
disk. Finally, the superblock itself is marked as \dirty" to show that it has been
changed and it is unlocked. If there were any processes waiting for the superblock,
the rst one in the queue is allowed to run again and will gain exclusive control of
9.2. The Virtual File System (VFS) 107
Disk
Drivers
Buffer
Cache
Cache
Inode
Cache
Directory
VFS
MINIX EXT2
Figure 9.4: A Logical Diagram of the Virtual File System
the superblock for its le operations. The process's data is written to the new data
block and, if that data block is lled, the entire process is repeated and another data
block allocated.
9.2 The Virtual File System (VFS)
Figure 9.4 shows the relationship between the Linux kernel's Virtual File System
and it's real le systems. The virtual le system must manage all of the dierent le
systems that are mounted at any given time. To do this it maintains data structures
that describe the whole (virtual) le system and the real, mounted, le systems.
Rather confusingly, the VFS describes the system's les in terms of superblocks See fs/*
and inodes in much the same way as the EXT2 le system uses superblocks and
inodes. Like the EXT2 inodes, the VFS inodes describe les and directories within
the system; the contents and topology of the Virtual File System. From now on, to
avoid confusion, I will write about VFS inodes and VFS superblocks to distinquish
them from EXT2 inodes and superblocks.
As each le system is initialised, it registers itself with the VFS. This happens as
the operating system initialises itself at system boot time. The real le systems
are either built into the kernel itself or are built as loadable modules. File System
modules are loaded as the system needs them, so, for example, if the VFAT le system
is implemented as a kernel module, then it is only loaded when a VFAT le system
is mounted. When a block device based le system is mounted, and this includes
the root le system, the VFS must read its superblock. Each le system type's
superblock read routine must work out the le system's topology and map that
information onto a VFS superblock data structure. The VFS keeps a list of the
108 Chapter 9. The File system
mounted le systems in the system together with their VFS superblocks. Each VFS
superblock contains information and pointers to routines that perform particular
functions. So, for example, the superblock representing a mounted EXT2 le system
contains a pointer to the EXT2 specic inode reading routine. This EXT2 inode
read routine, like all of the le system specic inode read routines, lls out the elds
in a VFS inode. Each VFS superblock contains a pointer to the rst VFS inode on
the le system. For the root le system, this is the inode that represents the ``/''
directory. This mapping of information is very ecient for the EXT2 le system but
moderately less so for other le systems.
As the system's processes access directories and les, system routines are called that
traverse the VFS inodes in the system. For example, typing ls for a directory or cat See
fs/inode.c
for a le cause the the Virtual File System to search through the VFS inodes that
represent the le system. As every le and directory on the system is represented
by a VFS inode, then a number of inodes will be being repeatedly accessed. These
inodes are kept in the inode cache which makes access to them quicker. If an inode
is not in the inode cache, then a le system specic routine must be called in order
to read the appropriate inode. The action of reading the inode causes it to be put
into the inode cache and further accesses to the inode keep it in the cache. The less
used VFS inodes get removed from the cache.
All of the Linux le systems use a common buer cache to cache data buers from the
underlying devices to help speed up access by all of the le systems to the physical
devices holding the le systems. This buer cache is independent of See the le systems
fs/buffer.c
and is integrated into the mechanisms that the Linux kernel uses to allocate and
read and write data buers. It has the distinct advantage of making the Linux
le systems independent from the underlying media and from the device drivers that
support them. All block structured devices register themselves with the Linux kernel
and present a uniform, block based, usually asynchronous interface. Even relatively
complex block devices such as SCSI devices do this. As the real le systems read
data from the underlying physical disks, this results in requests to the block device
drivers to read physical blocks from the device that they control. Integrated into this
block device interface is the buer cache. As blocks are read by the le systems they
are saved in the global buer cache shared by all of the le systems and the Linux
kernel. Buers within it are identied by their block number and a unique identier
for the device that read it. So, if the same data is needed often, it will be retrieved
from the buer cache rather than read from the disk, which would take somewhat
longer. Some devices support read ahead where data blocks are speculatively read
just in case they are needed.
The VFS also keeps a cache of directory lookups so that the inodes for frequently
See used directories can be quickly found. As an experiment, try listing a directory that
fs/dcache.c
you have not listed recently. The rst time you list it, you may notice a slight pause
but the second time you list its contents the result is immediate. The directory cache
does not store the inodes for the directories itself; these should be in the inode cache,
the directory cache simply stores the mapping between the full directory names and
their inode numbers.
9.2.1 The VFS Superblock
Every mounted le system is represented by a VFS superblock; amongst other information,
the VFS superblock contains the:
See
include/linux/-
fs.h
9.2. The Virtual File System (VFS) 109
Device This is the device identier for the block device that this le system is
contained in. For example, /dev/hda1, the rst IDE hard disk in the system
has a device identier of 0x301,
Inode pointers The mounted inode pointer points at the rst inode in this le system.
The covered inode pointer points at the inode representing the directory
that this le system is mounted on. The root le system's VFS superblock
does not have a covered pointer,
Blocksize The block size in bytes of this le system, for example 1024 bytes,
Superblock operations A pointer to a set of superblock routines for this le system.
Amongst other things, these routines are used by the VFS to read and
write inodes and superblocks.
File System type A pointer to the mounted le system's file system type data
structure,
File System specic A pointer to information needed by this le system,
9.2.2 The VFS Inode
Like the EXT2 le system, every le, directory and so on in the VFS is represented
by one and only one VFS inode. The information in each VFS inode is built from
See
include/linux/-
information in the underlying le system by le system specic routines. VFS inodes fs.h
exist only in the kernel's memory and are kept in the VFS inode cache as long as
they are useful to the system. Amongst other information, VFS inodes contain the
following elds:
device This is the device identifer of the device holding the le or whatever that
this VFS inode represents,
inode number This is the number of the inode and is unique within this le system.
The combination of device and inode number is unique within the Virtual File
System,
mode Like EXT2 this eld describes what this VFS inode represents as well as
access rights to it,
user ids The owner identiers,
times The creation, modication and write times,
block size The size of a block for this le in bytes, for example 1024 bytes,
inode operations A pointer to a block of routine addresses. These routines are
specic to the le system and they perform operations for this inode, for example,
truncate the le that is represented by this inode.
count The number of system components currently using this VFS inode. A count
of zero means that the inode is free to be discarded or reused,
lock This eld is used to lock the VFS inode, for example, when it is being read
from the le system,
110 Chapter 9. The File system
file_systems
requires_dev
file_system_type
*read_super()
name
next
"ext2"
requires_dev
file_system_type
*read_super()
name
next
requires_dev
file_system_type
*read_super()
name "iso9660"
next
"proc"
Figure 9.5: Registered File Systems
dirty Indicates whether this VFS inode has been written to, if so the underlying le
system will need modifying,
le system specic information
9.2.3 Registering the File Systems
When you build the Linux kernel you are asked if you want each of the supported
le systems. When the kernel is built, the le system startup code contains calls to
the initialisation routines of all of the built in le systems. Linux le systems may
See sys setup()
in fs/-
filesystems.c also be built as modules and, in this case, they may be demand loaded as they are
needed or loaded by hand using insmod. Whenever a le system module is loaded
it registers itself with the kernel and unregisters itself when it is unloaded. Each
le system's initialisation routine registers itself with the Virtual File System and is
represented by a file system type data structure which contains the name of the le
system and a pointer to its VFS superblock read routine. Figure 9.5 shows that the
file system type data structures are put into a list pointed at by the file systems
pointer. Each file system type data structure contains the See following information:
file system type
in include/-
linux/fs.h
Superblock read routine This routine is called by the VFS when an instance of
the le system is mounted,
File System name The name of this le system, for example ext2,
Device needed Does this le system need a device to support? Not all le system
need a device to hold them. The /proc le system, for example, does not
require a block device,
You can see which le systems are registered by looking in at /proc/filesystems.
For example:
ext2
nodev proc
iso9660
9.2.4 Mounting a File System
When the superuser attempts to mount a le system, the Linux kernel must rst
validate the arguments passed in the system call. Although mount does some basic
checking, it does not know which le systems this kernel has been built to support
or that the proposed mount point actually exists. Consider the following mount
command:
9.2. The Virtual File System (VFS) 111
$ mount -t iso9660 -o ro /dev/cdrom /mnt/cdrom
This mount command will pass the kernel three pieces of information; the name of
the le system, the physical block device that contains the le system and, thirdly,
where in the existing le system topology the new le system is to be mounted.
The rst thing that the Virtual File System must do is to nd the le system.
To do this it searches through the list of known le systems by looking at each
See do mount()
in
file system type data structure in the list pointed at by file systems. If it nds fs/super.c
See
get fs type() in
fs/super.c
a matching name it now knows that this le system type is supported by this kernel
and it has the address of the le system specic routine for reading this le system's
superblock. If it cannot nd a matching le system name then all is not lost if the
kernel is built to demand load kernel modules (see Chapter 12). In this case the
kernel will request that the kernel daemon loads the appropriate le system module
before continuing as before.
Next if the physical device passed by mount is not already mounted, it must nd the
VFS inode of the directory that is to be the new le system's mount point. This
VFS inode may be in the inode cache or it might have to be read from the block
device supporting the le system of the mount point. Once the inode has been found
it is checked to see that it is a directory and that there is not already some other
le system mounted there. The same directory cannot be used as a mount point for
more than one le system.
At this point the VFS mount code must allocate a VFS superblock and pass it the
mount information to the superblock read routine for this le system. All of the
system's VFS superblocks are kept in the super blocks vector of super block data
structures and one must be allocated for this mount. The superblock read routine
must ll out the VFS superblock elds based on information that it reads from the
physical device. For the EXT2 le system this mapping or translation of information
is quite easy, it simply reads the EXT2 superblock and lls out the VFS superblock
from there. For other le systems, such as the MS DOS le system, it is not quite such
an easy task. Whatever the le system, lling out the VFS superblock means that
the le system must read whatever describes it from the block device that supports
it. If the block device cannot be read from or if it does not contain this type of le
system then the mount command will fail.
Each mounted le system is described by a vfsmount data structure; see gure 9.6.
These are queued on a list pointed at by vfsmntlist. Another pointer, vfsmnttail
See
add vfsmnt() in
points at the last entry in the list and the mru vfsmnt pointer points at the most fs/super.c
recently used le system. Each vfsmount structure contains the device number of the
block device holding the le system, the directory where this le system is mounted
and a pointer to the VFS superblock allocated when this le system was mounted. In
turn the VFS superblock points at the file system type data structure for this sort
of le system and to the root inode for this le system. This inode is kept resident
in the VFS inode cache all of the time that this le system is loaded.
9.2.5 Finding a File in the Virtual File System
To nd the VFS inode of a le in the Virtual File System, VFS must resolve the name
a directory at a time, looking up the VFS inode representing each of the intermediate
directories in the name. Each directory lookup involves calling the le system specic
lookup whose address is held in the VFS inode representing the parent directory. This
112 Chapter 9. The File system
requires_dev
file_system_type
*read_super()
name
next
i_dev
i_ino
vfsmntlist
"ext2"
0x0301
0x0301
0x0301
/dev/hda1
/
vfsmount
mnt_dev
mnt_devname
mnt_sb
mnt_flags
mnt_dirname
next
1024
42
s_dev
s_blocksize
s_covered
s_flags
s_type
s_mounted
super_block
VFS
inode
VFS
Figure 9.6: A Mounted File System
works because we always have the VFS inode of the root of each le system available
and pointed at by the VFS superblock for that system. Each time an inode is looked
up by the real le system it checks the directory cache for the directory. If there is
no entry in the directory cache, the real le system gets the VFS inode either from
the underlying le system or from the inode cache.
9.2.6 Creating a File in the Virtual File System
9.2.7 Unmounting a File System
The workshop manual for my MG usually describes assembly as the reverse of disassembly
and the reverse is more or less true for unmounting a le system. A le
See do umount()
in
fs/super.c system cannot be unmounted if something in the system is using one of its les. So,
for example, you cannot umount /mnt/cdrom if a process is using that directory or
any of its children. If anything is using the le system to be unmounted there may be
VFS inodes from it in the VFS inode cache, and the code checks for this by looking
through the list of inodes looking for inodes owned by the device that this le system
occupies. If the VFS superblock for the mounted le system is dirty, that is it has
been modied, then it must be written back to the le system on disk. Once it has
been written to disk, the memory occupied by the VFS superblock is returned to the
kernel's free pool of memory. Finally the vfsmount data structure for this mount is
unlinked from vfsmntlist and freed.
See
remove vfsmnt()
in
fs/super.c
9.2.8 The VFS Inode Cache
As the mounted le systems are navigated, their VFS inodes are being continually
read and, in some cases, written. The Virtual File System maintains an inode cache
to speed up accesses to all of the mounted le systems. Every time a VFS inode is
read from the inode cache the system saves an access to a See physical device.
fs/inode.c
The VFS inode cache is implmented as a hash table whose entries are pointers to
9.2. The Virtual File System (VFS) 113
lists of VFS inodes that have the same hash value. The hash value of an inode is
calculated from its inode number and from the device identier for the underlying
physical device containing the le system. Whenever the Virtual File System needs
to access an inode, it rst looks in the VFS inode cache. To nd an inode in the
cache, the system rst calculates its hash value and then uses it as an index into the
inode hash table. This gives it a pointer to a list of inodes with the same hash value.
It then reads each inode in turn until it nds one with both the same inode number
and the same device identier as the one that it is searching for.
If it can nd the inode in the cache, its count is incremented to show that it has
another user and the le system access continues. Otherwise a free VFS inode must
be found so that the le system can read the inode from memory. VFS has a number
of choices about how to get a free inode. If the system may allocate more VFS inodes
then this is what it does; it allocates kernel pages and breaks them up into new, free,
inodes and puts them into the inode list. All of the system's VFS inodes are in a
list pointed at by first inode as well as in the inode hash table. If the system
already has all of the inodes that it is allowed to have, it must nd an inode that is
a good candidate to be reused. Good candidates are inodes with a usage count of
zero; this indicates that the system is not currently using them. Really important
VFS inodes, for example the root inodes of le systems always have a usage count
greater than zero and so are never candidates for reuse. Once a candidate for reuse
has been located it is cleaned up. The VFS inode might be dirty and in this case it
needs to be written back to the le system or it might be locked and in this case the
system must wait for it to be unlocked before continuing. The candidate VFS inode
must be cleaned up before it can be reused.
However the new VFS inode is found, a le system specic routine must be called
to ll it out from information read from the underlying real le system. Whilst it is
being lled out, the new VFS inode has a usage count of one and is locked so that
nothing else accesses it until it contains valid information.
To get the VFS inode that is actually needed, the le system may need to access
several other inodes. This happens when you read a directory; only the inode for
the nal directory is needed but the inodes for the intermediate directories must also
be read. As the VFS inode cache is used and lled up, the less used inodes will be
discarded and the more used inodes will remain in the cache.
9.2.9 The Directory Cache
To speed up accesses to commonly used directories, the VFS maintains a cache of
directory entries. As directories are looked up by the real le systems their details See
fs/dcache.c
are added into the directory cache. The next time the same directory is looked up,
for example to list it or open a le within it, then it will be found in the directory
cache. Only short directory entries (up to 15 characters long) are cached but this
is reasonable as the shorter directory names are the most commonly used ones. For
example, /usr/X11R6/bin is very commonly accessed when the X server is running.
The directory cache consists of a hash table, each entry of which points at a list
of directory cache entries that have the same hash value. The hash function uses
the device number of the device holding the le system and the directory's name to
calculate the oset, or index, into the hash table. It allows cached directory entries
to be quickly found. It is no use having a cache when lookups within the cache take
114 Chapter 9. The File system
b_dev
b_blocknr
b_state
b_count
b_size
b_prev
b_next
b_data
buffer_head
b_dev
b_blocknr
b_state
b_count
b_size
b_prev
b_next
b_data
hash_table buffer_head
b_dev
b_blocknr
b_state
b_count
b_size
b_prev
b_next
b_data
buffer_head
0x0301
0x0301
0x0801
42 17
39
1024 2048
1024
Figure 9.7: The Buer Cache
too long to nd entries, or even not to nd them.
In an eort to keep the caches valid and up to date the VFS keeps lists of Least
Recently Used (LRU) directory cache entries. When a directory entry is rst put
into the cache, which is when it is rst looked up, it is added onto the end of the rst
level LRU list. In a full cache this will displace an existing entry from the front of the
LRU list. As the directory entry is accessed again it is promoted to the back of the
second LRU cache list. Again, this may displace a cached level two directory entry
at the front of the level two LRU cache list. This displacing of entries at the front
of the level one and level two LRU lists is ne. The only reason that entries are at
the front of the lists is that they have not been recently accessed. If they had, they
would be nearer the back of the lists. The entries in the second level LRU cache list
are safer than entries in the level one LRU cache list. This is the intention as these
entries have not only been looked up but also they have been repeatedly referenced.
REVIEW NOTE: Do we need a diagram for this?
9.3 The Buer Cache
As the mounted le systems are used they generate a lot of requests to the block
devices to read and write data blocks. All block data read and write requests are
given to the device drivers in the form of buffer head data structures via standard
kernel routine calls. These give all of the information that the block device drivers
need; the device identier uniquely identies the device and the block number tells
the driver which block to read. All block devices are viewed as linear collections
of blocks of the same size. To speed up access to the physical block devices, Linux
maintains a cache of block buers. All of the block buers in the system are kept
somewhere in this buer cache, even the new, unused buers. This cache is shared
between all of the physical block devices; at any one time there are many block
buers in the cache, belonging to any one of the system's block devices and often in
many dierent states. If valid data is available from the buer cache this saves the
9.3. The Buer Cache 115
system an access to a physical device. Any block buer that has been used to read
data from a block device or to write data to it goes into the buer cache. Over time
it may be removed from the cache to make way for a more deserving buer or it may
remain in the cache as it is frequently accessed.
Block buers within the cache are uniquely idented by the owning device identier
and the block number of the buer. The buer cache is composed of two functional
parts. The rst part is the lists of free block buers. There is one list per supported
buer size and the system's free block buers are queued onto these lists when they
are rst created or when they have been discarded. The currently supported buer
sizes are 512, 1024, 2048, 4096 and 8192 bytes. The second functional part is the
cache itself. This is a hash table which is a vector of pointers to chains of buers
that have the same hash index. The hash index is generated from the owning device
identier and the block number of the data block. Figure 9.7 shows the hash table
together with a few entries. Block buers are either in one of the free lists or they
are in the buer cache. When they are in the buer cache they are also queued onto
Least Recently Used (LRU) lists. There is an LRU list for each buer type and these
are used by the system to perform work on buers of a type, for example, writing
buers with new data in them out to disk. The buer's type re
ects its state and
Linux currently supports the following types:
clean Unused, new buers,
locked Buers that are locked, waiting to be written,
dirty Dirty buers. These contain new, valid data, and will be written but so far
have not been scheduled to write,
shared Shared buers,
unshared Buers that were once shared but which are now not shared,
Whenever a le system needs to read a buer from its underlying physical device, it
trys to get a block from the buer cache. If it cannot get a buer from the buer
cache, then it will get a clean one from the appropriate sized free list and this new
buer will go into the buer cache. If the buer that it needed is in the buer cache,
then it may or may not be up to date. If it is not up to date or if it is a new block
buer, the le system must request that the device driver read the appropriate block
of data from the disk.
Like all caches, the buer cache must be maintained so that it runs eciently and
fairly allocates cache entries between the block devices using the buer cache. Linux
uses the bdflush kernel daemon to perform a lot of housekeeping duties on the
cache but some happen automatically as a result of the cache being used.
9.3.1 The bdflush Kernel Daemon See bdflush()
in
fs/buffer.c The bdflush kernel daemon is a simple kernel daemon that provides a dynamic
response to the system having too many dirty buers; buers that contain data that
must be written out to disk at some time. It is started as a kernel thread at system
startup time and, rather confusingly, it calls itself \k
ushd" and that is the name
that you will see if you use the ps command to show the processes in the system.
Mostly this daemon sleeps waiting for the number of dirty buers in the system to
116 Chapter 9. The File system
grow too large. As buers are allocated and discarded the number of dirty buers in
the system is checked. If there are too many as a percentage of the total number of
buers in the system then bdflush is woken up. The default threshold is 60% but,
if the system is desperate for buers, bdflush will be woken up anyway. This value
can be seen and changed using the update command:
# update -d
bdflush version 1.4
0: 60 Max fraction of LRU list to examine for dirty blocks
1: 500 Max number of dirty blocks to write each time bdflush activated
2: 64 Num of clean buffers to be loaded onto free list by refill_freelist
3: 256 Dirty block threshold for activating bdflush in refill_freelist
4: 15 Percentage of cache to scan for free clusters
5: 3000 Time for data buffers to age before flushing
6: 500 Time for non-data (dir, bitmap, etc) buffers to age before flushing
7: 1884 Time buffer cache load average constant
8: 2 LAV ratio (used to determine threshold for buffer fratricide).
All of the dirty buers are linked into the BUF DIRTY LRU list whenever they are
made dirty by having data written to them and bdflush tries to write a reasonable
number of them out to their owning disks. Again this number can be seen and
controlled by the update command and the default is 500 (see above).
9.3.2 The update Process
The update command is more than just a command; it is also a daemon. When run
as superuser (during system initialisation) it will periodically
ush all of the older
dirty buers out to disk. It does this by calling a system service routine that does
See
sys bdflush() in
fs/buffer.c more or less the same thing as bdflush. Whenever a dirty buer is nished with,
it is tagged with the system time that it should be written out to its owning disk.
Every time that update runs it looks at all of the dirty buers in the system looking
for ones with an expired
ush time. Every expired buer is written out to disk.
9.4 The /proc File System
The /proc le system really shows the power of the Linux Virtual File System. It
does not really exist (yet another of Linux's conjuring tricks), neither the /proc
directory nor its subdirectories and its les actually exist. So how can you cat
/proc/devices? The /proc le system, like a real le system, registers itself with the
Virtual File System. However, when the VFS makes calls to it requesting inodes as
its les and directories are opened, the /proc le system creates those les and directories
from information within the kernel. For example, the kernel's /proc/devices
le is generated from the kernel's data structures describing its devices.
The /proc le system presents a user readable window into the kernel's inner workings.
Several Linux subsystems, such as Linux kernel modules described in chapter
12, create entries in the the /proc le system.
9.5. Device Special Files 117
9.5 Device Special Files
Linux, like all versions of UnixTM presents its hardware devices as special les. So,
for example, /dev/null is the null device. A device le does not use any data space in
the le system, it is only an access point to the device driver. The EXT2 le system
and the Linux VFS both implement device les as special types of inode. There are
two types of device le; character and block special les. Within the kernel itself,
the device drivers implement le semantices: you can open them, close them and
so on. Character devices allow I/O operations in character mode and block devices
require that all I/O is via the buer cache. When an I/O request is made to a device
le, it is forwarded to the appropriate device driver within the system. Often this
is not a real device driver but a pseudo-device driver for some subsystem such as
the SCSI device driver layer. Device les are referenced by a major number, which
identies the device type, and a minor type, which identies the unit, or instance
of that major type. For example, the IDE disks on the rst IDE controller in the
system have a major number of 3 and the rst partition of an IDE disk would have
a minor number of 1. So, ls -l of /dev/hda1 gives:
see
/include/linux/
major.h for all of
Linux's major
device numb$ brw-rw---- 1 root disk 3, 1 Nov 24 15:09 /dev/hda1 ers.
Within the kernel, every device is uniquely described by a kdev t data type, this is
two bytes long, the rst byte containing the minor device number and the second
byte holding the major device number. The IDE device above is held within the
See
include/linux/-
kernel as 0x0301. An EXT2 inode that represents a block or character device keeps kdev t.h
the device's major and minor numbers in its rst direct block pointer. When it is
read by the VFS, the VFS inode data structure representing it has its i rdev eld
set to the correct device identier.
118 Chapter 9. The File system
Chapter 10
Networks
Networking and Linux are terms that are almost synonymous. In a very
real sense Linux is a product of the Internet orWorld WideWeb (WWW).
Its developers and users use the web to exchange information ideas, code,
and Linux itself is often used to support the networking needs of organizations.
This chapter describes how Linux supports the network protocols
known collectively as TCP/IP.
The TCP/IP protocols were designed to support communications between computers
connected to the ARPANET, an American research network funded by the US government.
The ARPANET pioneered networking concepts such as packet switching
and protocol layering where one protocol uses the services of another. ARPANET
was retired in 1988 but its successors (NSF1 NET and the Internet) have grown even
larger. What is now known as the World Wide Web grew from the ARPANET and
is itself supported by the TCP/IP protocols. UnixTM was extensively used on the
ARPANET and the rst released networking version of UnixTM was 4.3 BSD. Linux's
networking implementation is modeled on 4.3 BSD in that it supports BSD sockets
(with some extensions) and the full range of TCP/IP networking. This programming
interface was chosen because of its popularity and to help applications be portable
between Linux and other UnixTM platforms.
10.1 An Overview of TCP/IP Networking
This section gives an overview of the main principles of TCP/IP networking. It is not
meant to be an exhaustive description, for that I suggest that you read [10, Comer].
1National Science Foundation
119
120 Chapter 10. Networks
In an IP network every machine is assigned an IP address, this is a 32 bit number
that uniquely identies the machine. The WWW is a very large, and growing, IP
network and every machine that is connected to it has to have a unique IP address
assigned to it. IP addresses are represented by four numbers separated by dots, for
example, 16.42.0.9. This IP address is actually in two parts, the network address
and the host address. The sizes of these parts may vary (there are several classes of IP
addresses) but using 16.42.0.9 as an example, the network address would be 16.42
and the host address 0.9. The host address is further subdivided into a subnetwork
and a host address. Again, using 16.42.0.9 as an example, the subnetwork address
would be 16.42.0 and the host address 16.42.0.9. This subdivision of the IP address
allows organizations to subdivide their networks. For example, 16.42 could be the
network address of the ACME Computer Company; 16.42.0 would be subnet 0
and 16.42.1 would be subnet 1. These subnets might be in separate buildings,
perhaps connected by leased telephone lines or even microwave links. IP addresses
are assigned by the network administrator and having IP subnetworks is a good way
of distributing the administration of the network. IP subnet administrators are free
to allocate IP addresses within their IP subnetworks.
Generally though, IP addresses are somewhat hard to remember. Names are much
easier. linux.acme.com is much easier to remember than 16.42.0.9 but there must
be some mechanism to convert the network names into an IP address. These names
can be statically specied in the /etc/hosts le or Linux can ask a Distributed
Name Server (DNS server) to resolve the name for it. In this case the local host
must know the IP address of one or more DNS servers and these are specied in
/etc/resolv.conf.
Whenever you connect to another machine, say when reading a web page, its IP
address is used to exchange data with that machine. This data is contained in IP
packets each of which have an IP header containing the IP addresses of the source
and destination machine's IP addresses, a checksum and other useful information.
The checksum is derived from the data in the IP packet and allows the receiver of
IP packets to tell if the IP packet was corrupted during transmission, perhaps by a
noisy telephone line. The data transmitted by an application may have been broken
down into smaller packets which are easier to handle. The size of the IP data packets
varies depending on the connection media; ethernet packets are generally bigger than
PPP packets. The destination host must reassemble the data packets before giving
the data to the receiving application. You can see this fragmentation and reassembly
of data graphically if you access a web page containing a lot of graphical images via
a moderately slow serial link.
Hosts connected to the same IP subnet can send IP packets directly to each other, all
other IP packets will be sent to a special host, a gateway. Gateways (or routers) are
connected to more than one IP subnet and they will resend IP packets received on
one subnet, but destined for another onwards. For example, if subnets 16.42.1.0
and 16.42.0.0 are connected together by a gateway then any packets sent from
subnet 0 to subnet 1 would have to be directed to the gateway so that it could route
them. The local host builds up routing tables which allow it to route IP packets to
the correct machine. For every IP destination there is an entry in the routing tables
which tells Linux which host to send IP packets to in order that they reach their
destination. These routing tables are dynamic and change over time as applications
use the network and as the network topology changes.
10.1. An Overview of TCP/IP Networking 121
Destination
ethernet
address
Source
ethernet
address
Destination
IP address
Source
IP address
Destination
TCP address
ETHERNET FRAME
IP PACKET
Data Checksum
Length Protocol Checksum Data
TCP PACKET
SEQ ACK Data
address
Source TCP
Protocol
Figure 10.1: TCP/IP Protocol Layers
The IP protocol is a transport layer that is used by other protocols to carry their data.
The Transmission Control Protocol (TCP) is a reliable end to end protocol that uses
IP to transmit and receive its own packets. Just as IP packets have their own header,
TCP has its own header. TCP is a connection based protocol where two networking
applications are connected by a single, virtual connection even though there may be
many subnetworks, gateways and routers between them. TCP reliably transmits and
receives data between the two applications and guarantees that there will be no lost or
duplicated data. When TCP transmits its packet using IP, the data contained within
the IP packet is the TCP packet itself. The IP layer on each communicating host
is responsible for transmitting and receiving IP packets. User Datagram Protocol
(UDP) also uses the IP layer to transport its packets, unlike TCP, UDP is not a
reliable protocol but oers a datagram service. This use of IP by other protocols
means that when IP packets are received the receiving IP layer must know which
upper protocol layer to give the data contained in this IP packet to. To facilitate this
every IP packet header has a byte containing a protocol identier. When TCP asks
the IP layer to transmit an IP packet , that IP packet's header states that it contains
a TCP packet. The receiving IP layer uses that protocol identier to decide which
layer to pass the received data up to, in this case the TCP layer. When applications
communicate via TCP/IP they must specify not only the target's IP address but also
the port address of the application. A port address uniquely identies an application
and standard network applications use standard port addresses; for example, web
servers use port 80. These registered port addresses can be seen in /etc/services.
This layering of protocols does not stop with TCP, UDP and IP. The IP protocol
layer itself uses many dierent physical media to transport IP packets to other IP
hosts. These media may themselves add their own protocol headers. One such
example is the ethernet layer, but PPP and SLIP are others. An ethernet network
allows many hosts to be simultaneously connected to a single physical cable. Every
transmitted ethernet frame can be seen by all connected hosts and so every ethernet
device has a unique address. Any ethernet frame transmitted to that address will be
received by the addressed host but ignored by all the other hosts connected to the
122 Chapter 10. Networks
network. These unique addresses are built into each ethernet device when they are
manufactured and it is usually kept in an SROM2 on the ethernet card. Ethernet
addresses are 6 bytes long, an example would be 08-00-2b-00-49-A4. Some ethernet
addresses are reserved for multicast purposes and ethernet frames sent with these
destination addresses will be received by all hosts on the network. As ethernet frames
can carry many dierent protocols (as data) they, like IP packets, contain a protocol
identier in their headers. This allows the ethernet layer to correctly receive IP
packets and to pass them onto the IP layer.
In order to send an IP packet via a multi-connection protocol such as ethernet, the
IP layer must nd the ethernet address of the IP host. This is because IP addresses
are simply an addressing concept, the ethernet devices themselves have their own
physical addresses. IP addresses on the other hand can be assigned and reassigned
by network administrators at will but the network hardware responds only to ethernet
frames with its own physical address or to special multicast addresses which
all machines must receive. Linux uses the Address Resolution Protocol (or ARP)
to allow machines to translate IP addresses into real hardware addresses such as
ethernet addresses. A host wishing to know the hardware address associated with
an IP address sends an ARP request packet containing the IP address that it wishes
translating to all nodes on the network by sending it to a multicast address. The
target host that owns the IP address, responds with an ARP reply that contains its
physical hardware address. ARP is not just restricted to ethernet devices, it can
resolve IP addresses for other physical media, for example FDDI. Those network
devices that cannot ARP are marked so that Linux does not attempt to ARP. There
is also the reverse function, Reverse ARP or RARP, which translates phsyical network
addresses into IP addresses. This is used by gateways, which respond to ARP
requests on behalf of IP addresses that are in the remote network.
10.2 The Linux TCP/IP Networking Layers
Just like the network protocols themselves, Figure 10.2 shows that Linux implements
the internet protocol address family as a series of connected layers of software. BSD
sockets are supported by a generic socket management software concerned only with
BSD sockets. Supporting this is the INET socket layer, this manages the communication
end points for the IP based protocols TCP and UDP. UDP (User Datagram
Protocol) is a connectionless protocol whereas TCP (Transmission Control Protocol)
is a reliable end to end protocol. When UDP packets are transmitted, Linux neither
knows nor cares if they arrive safely at their destination. TCP packets are numbered
and both ends of the TCP connection make sure that transmitted data is received
correctly. The IP layer contains code implementing the Internet Protocol. This code
prepends IP headers to transmitted data and understands how to route incoming IP
packets to either the TCP or UDP layers. Underneath the IP layer, supporting all
of Linux's networking are the network devices, for example PPP and ethernet. Network
devices do not always represent physical devices; some like the loopback device
are purely software devices. Unlike standard Linux devices that are created via the
mknod command, network devices appear only if the underlying software has found
and initialized them. You will only see /dev/eth0 when you have built a kernel with
the appropriate ethernet device driver in it. The ARP protocol sits between the IP
2Synchronous Read Only Memory
10.2. The Linux TCP/IP Networking Layers 123
User
Kernel
Network
Devices
Socket
Interface
Protocol
Layers
Sockets
BSD
INET
Sockets
TCP UDP
Applications
Network
Ethernet
IP
PPP SLIP
ARP
Figure 10.2: Linux Networking Layers
124 Chapter 10. Networks
layer and the protocols that support ARPing for addresses.
10.3 The BSD Socket Interface
This is a general interface which not only supports various forms of networking but
is also an inter-process communications mechanism. A socket describes one end
of a communications link, two communicating processes would each have a socket
describing their end of the communication link between them. Sockets could be
thought of as a special case of pipes but, unlike pipes, sockets have no limit on the
amount of data that they can contain. Linux supports several classes of socket and
these are known as address families. This is because each class has its own method of
addressing its communications. Linux supports the following socket address families
or domains:
UNIX Unix domain sockets,
INET The Internet address family supports communications via
TCP/IP protocols
AX25 Amateur radio X25
IPX Novell IPX
APPLETALK Appletalk DDP
X25 X25
There are several socket types and these represent the type of service that supports
the connection. Not all address families support all types of service. Linux BSD
sockets support a number of socket types:
Stream These sockets provide reliable two way sequenced data streams with a guarantee
that data cannot be lost, corrupted or duplicated in transit. Stream
sockets are supported by the TCP protocol of the Internet (INET) address
family.
Datagram These sockets also provide two way data transfer but, unlike stream
sockets, there is no guarantee that the messages will arrive. Even if they
do arrive there is no guarantee that they will arrive in order or even not be
duplicated or corrupted. This type of socket is supported by the UDP protocol
of the Internet address family.
Raw This allows processes direct (hence \raw") access to the underlying protocols.
It is, for example, possible to open a raw socket to an ethernet device and see
raw IP data trac.
Reliable Delivered Messages These are very like datagram sockets but the data
is guaranteed to arrive.
Sequenced Packets These are like stream sockets except that the data packet sizes
are xed.
Packet This is not a standard BSD socket type, it is a Linux specic extension that
allows processes to access packets directly at the device level.
Processes that communicate using sockets use a client server model. A server provides
a service and clients make use of that service. One example would be a Web Server,
which provides web pages and a web client, or browser, which reads those pages. A
10.4. The INET Socket Layer 125
server using sockets, rst creates a socket and then binds a name to it. The format
of this name is dependent on the socket's address family and it is, in eect, the local
address of the server. The socket's name or address is specied using the sockaddr
data structure. An INET socket would have an IP port address bound to it. The
registered port numbers can be seen in /etc/services; for example, the port number
for a web server is 80. Having bound an address to the socket, the server then listens
for incoming connection requests specifying the bound address. The originator of the
request, the client, creates a socket and makes a connection request on it, specifying
the target address of the server. For an INET socket the address of the server is its
IP address and its port number. These incoming requests must nd their way up
through the various protocol layers and then wait on the server's listening socket.
Once the server has received the incoming request it either accepts or rejects it. If
the incoming request is to be accepted, the server must create a new socket to accept
it on. Once a socket has been used for listening for incoming connection requests it
cannot be used to support a connection. With the connection established both ends
are free to send and receive data. Finally, when the connection is no longer needed it
can be shutdown. Care is taken to ensure that data packets in transit are correctly
dealt with.
The exact meaning of operations on a BSD socket depends on its underlying address
family. Setting up TCP/IP connections is very dierent from setting up an amateur
radio X.25 connection. Like the virtual lesystem, Linux abstracts the socket interface
with the BSD socket layer being concerned with the BSD socket interface to
the application programs which is in turn supported by independent address family
specic software. At kernel initialization time, the address families built into the
kernel register themselves with the BSD socket interface. Later on, as applications
create and use BSD sockets, an association is made between the BSD socket and
its supporting address family. This association is made via cross-linking data structures
and tables of address family specic support routines. For example there is an
address family specic socket creation routine which the BSD socket interface uses
when an application creates a new socket.
When the kernel is congured, a number of address families and protocols are built
into the protocols vector. Each is represented by its name, for example \INET"
and the address of its initialization routine. When the socket interface is initialized
at boot time each protocol's initialization routine is called. For the socket address
families this results in them registering a set of protocol operations. This is a set
of routines, each of which performs a a particular operation specic to that address
family. The registered protocol operations are kept in the pops vector, a vector of
pointers to proto ops data structures. The proto ops data structure consists of
See
include/linux/-
the address family type and a set of pointers to socket operation routines specic net.h
to a particular address family. The pops vector is indexed by the address family
identier, for example the Internet address family identier (AF INET is 2).
10.4 The INET Socket Layer
The INET socket layer supports the internet address family which contains the
TCP/IP protocols. As discussed above, these protocols are layered, one protocol
using the services of another. Linux's TCP/IP code and data structures re
ect this
layering. Its interface with the BSD socket layer is through the set of Internet ad126
Chapter 10. Networks
files_struct
count
close_on_exec
open_fs
fd[0]
fd[1]
fd[255]
type
ops
inode
data
socket
Address Family
socket operations
BSD Socket
File Operations
lseek
read
write
select
ioctl
close
fasync
socket
sock
type
protocol
file
f_mode
f_pos
f_flags
f_count
f_owner
f_version
f_op
f_inode
SOCK_STREAM
SOCK_STREAM
Figure 10.3: Linux BSD Socket Data Structures
10.4. The INET Socket Layer 127
dress family socket operations which it registers with the BSD socket layer during
network initialization. These are kept in the pops vector along with the other registered
address families. The BSD socket layer calls the INET layer socket support
routines from the registered INET proto ops data structure to perform work for it.
For example a BSD socket create request that gives the address family as INET will
use the underlying INET socket create function. The BSD socket layer passes the
socket data structure representing the BSD socket to the INET layer in each of these
operations. Rather than clutter the BSD socket wiht TCP/IP specic information,
the INET socket layer uses its own data structure, the sock which it links to the
See
include/net/-
BSD socket data structure. This linkage can be seen in Figure 10.3. It links the sock.h
sock data structure to the BSD socket data structure using the data pointer in
the BSD socket. This means that subsequent INET socket calls can easily retrieve
the sock data structure. The sock data structure's protocol operations pointer is
also set up at creation time and it depends on the protocol requested. If TCP is
requested, then the sock data structure's protocol operations pointer will point to
the set of TCP protocol operations needed for a TCP connection.
10.4.1 Creating a BSD Socket
The system call to create a new socket passes identiers for its address family, socket
type and protocol. Firstly the requested address family is used to search the pops
See
sys socket() in
vector for a matching address family. It may be that a particular address family is net/socket.c
implemented as a kernel module and, in this case, the kerneld daemon must load
the module before we can continue. A new socket data structure is allocated to
represent the BSD socket. Actually the socket data structure is physically part of
the VFS inode data structure and allocating a socket really means allocating a VFS
inode. This may seem strange unless you consider that sockets can be operated on
in just the same way that ordinairy les can. As all les are represented by a VFS
inode data structure, then in order to support le operations, BSD sockets must
also be represented by a VFS inode data structure.
The newly created BSD socket data structure contains a pointer to the address
family specic socket routines and this is set to the proto ops data structure retrieved
from the pops vector. Its type is set to the sccket type requested; one of
SOCK STREAM, SOCK DGRAM and so on. The address family specic creation
routine is called using the address kept in the proto ops data structure.
A free le descriptor is allocated from the current processes fd vector and the file
data structure that it points at is initialized. This includes setting the le operations
pointer to point to the set of BSD socket le operations supported by the BSD socket
interface. Any future operations will be directed to the socket interface and it will
in turn pass them to the supporting address family by calling its address family
operation routines.
10.4.2 Binding an Address to an INET BSD Socket
In order to be able to listen for incoming internet connection requests, each server
must create an INET BSD socket and bind its address to it. The bind operation is
mostly handled within the INET socket layer with some support from the underlying
TCP and UDP protocol layers. The socket having an address bound to cannot be
being used for any other communication. This means that the socket's state must
128 Chapter 10. Networks
be TCP CLOSE. The sockaddr pass to the bind operation contains the IP address to
be bound to and, optionally, a port number. Normally the IP address bound to
would be one that has been assigned to a network device that supports the INET
address family and whose interface is up and able to be used. You can see which
network interfaces are currently active in the system by using the ifcong command.
The IP address may also be the IP broadcast address of either all 1's or all 0's.
These are special addresses that mean \send to everybody"3. The IP address could
also be specied as any IP address if the machine is acting as a transparent proxy or
rewall, but only processes with superuser privileges can bind to any IP address. The
IP address bound to is saved in the sock data structure in the recv addr and saddr
elds. These are used in hash lookups and as the sending IP address respectively.
The port number is optional and if it is not specied the supporting network is
asked for a free one. By convention, port numbers less than 1024 cannot be used
by processes without superuser privileges. If the underlying network does allocate a
port number it always allocates ones greater than 1024.
As packets are being received by the underlying network devices they must be routed
to the correct INET and BSD sockets so that they can be processed. For this reason
UDP and TCP maintain hash tables which are used to lookup the addresses within
incoming IP messages and direct them to the correct socket/sock pair. TCP is a
connection oriented protocol and so there is more information involved in processing
TCP packets than there is in processing UDP packets.
UDP maintains a hash table of allocated UDP ports, the udp hash table. This
consists of pointers to sock data structures indexed by a hash function based on the
port number. As the UDP hash table is much smaller than the number of permissible
port numbers (udp hash is only 128 or UDP HTABLE SIZE entries long) some entries in
the table point to a chain of sock data structures linked together using each sock's
next pointer.
TCP is much more complex as it maintains several hash tables. However, TCP
does not actually add the binding sock data stucture into its hash tables during the
bind operation, it merely checks that the port number requested is not currently
being used. The sock data structure is added to TCP's hash tables during the listen
operation.
REVIEW NOTE: What about the route entered?
10.4.3 Making a Connection on an INET BSD Socket
Once a socket has been created and, provided it has not been used to listen for
inbound connection requests, it can be used to make outbound connection requests.
For connectionless protocols like UDP this socket operation does not do a whole lot
but for connection orientated protocols like TCP it involves building a virtual circuit
between two applications.
An outbound connection can only be made on an INET BSD socket that is in the
right state; that is to say one that does not already have a connection established and
one that is not being used for listening for inbound connections. This means that the
BSD socket data structure must be in state SS UNCONNECTED. The UDP protocol
does not establish virtual connections between applications, any messages sent are
datagrams, one o messages that may or may not reach their destinations. It does,
3duh? What used for?
10.4. The INET Socket Layer 129
however, support the connect BSD socket operation. A connection operation on a
UDP INET BSD socket simply sets up the addresses of the remote application; its IP
address and its IP port number. Additionally it sets up a cache of the routing table
entry so that UDP packets sent on this BSD socket do not need to check the routing
database again (unless this route becomes invalid). The cached routing information
is pointed at from the ip route cache pointer in the INET sock data structure. If
no addressing information is given, this cached routing and IP addressing information
will be automatically be used for messages sent using this BSD socket. UDP moves
the sock's state to TCP ESTABLISHED.
For a connect operation on a TCP BSD socket, TCP must build a TCP message
containing the connection information and send it to IP destination given. The TCP
message contains information about the connection, a unique starting message sequence
number, the maximum sized message that can be managed by the initiating
host, the transmit and receive window size and so on. Within TCP all messages
are numbered and the initial sequence number is used as the rst message number.
Linux chooses a reasonably random value to avoid malicious protocol attacks. Every
message transmitted by one end of the TCP connection and successfully received by
the other is acknowledged to say that it arrived successfully and uncorrupted. Unacknowledges
messages will be retransmitted. The transmit and receive window size is
the number of outstanding messages that there can be without an acknowledgement
being sent. The maximum message size is based on the network device that is being
used at the initiating end of the request. If the receiving end's network device supports
smaller maximum message sizes then the connection will use the minimum of
the two. The application making the outbound TCP connection request must now
wait for a response from the target application to accept or reject the connection
request. As the TCP sock is now expecting incoming messages, it is added to the
tcp listening hash so that incoming TCP messages can be directed to this sock
data structure. TCP also starts timers so that the outbound connection request can
be timed out if the target application does not respond to the request.
10.4.4 Listening on an INET BSD Socket
Once a socket has had an address bound to it, it may listen for incoming connection
requests specifying the bound addresses. A network application can listen on a socket
without rst binding an address to it; in this case the INET socket layer nds an
unused port number (for this protocol) and automatically binds it to the socket. The
listen socket function moves the socket into state TCP LISTEN and does any network
specic work needed to allow incoming connections.
For UDP sockets, changing the socket's state is enough but TCP now adds the
socket's sock data structure into two hash tables as it is now active. These are the
tcp bound hash table and the tcp listening hash. Both are indexed via a hash
function based on the IP port number.
Whenever an incoming TCP connection request is received for an active listening
socket, TCP builds a new sock data structure to represent it. This sock data
structure will become the bottom half of the TCP connection when it is eventually
accepted. It also clones the incoming sk buff containing the connection request and
queues it onto the receive queue for the listening sock data structure. The clone
sk buff contains a pointer to the newly created sock data structure.
130 Chapter 10. Networks
10.4.5 Accepting Connection Requests
UDP does not support the concept of connections, accepting INET socket connection
requests only applies to the TCP protocol as an accept operation on a listening socket
causes a new socket data structure to be cloned from the original listening socket.
The accept operation is then passed to the supporting protocol layer, in this case
INET to accept any incoming connection requests. The INET protocol layer will
fail the accept operation if the underlying protocol, say UDP, does not support
connections. Otherwise the accept operation is passed through to the real protocol,
in this case TCP. The accept operation can be either blocking or non-blocking. In
the non-blocking case if there are no incoming connections to accept, the accept
operation will fail and the newly created socket data structure will be thrown away.
In the blocking case the network application performing the accept operation will
be added to a wait queue and then suspended until a TCP connection request is
received. Once a connection request has been received the sk buff containing the
request is discarded and the sock data structure is returned to the INET socket
layer where it is linked to the new socket data structure created earlier. The le
descriptor (fd) number of the new socket is returned to the network application,
and the application can then use that le descriptor in socket operations on the newly
created INET BSD socket.
10.5 The IP Layer
10.5.1 Socket Buers
One of the problems of having many layers of network protocols, each one using the
services of another, is that each protocol needs to add protocol headers and tails to
data as it is transmitted and to remove them as it processes received data. This make
passing data buers between the protocols dicult as each layer needs to nd where
its particular protocol headers and tails are. One solution is to copy buers at each
layer but that would be inecient. Instead, Linux uses socket buers or sk buffs
to pass data between the protocol layers and the network device drivers. sk buffs
contain pointer and length elds that allow each protocol layer to manipulate the
application data via standard functions or \methods".
Figure 10.4 shows the sk buff data structure; each sk buff has a block of data
See
include/linux/-
skbuff.h associated with it. The sk buff has four data pointers, which are used to manipulate
and manage the socket buer's data:
head points to the start of the data area in memory. This is xed when the sk buff
and its associated data block is allocated,
data points at the current start of the protocol data. This pointer varies depending
on the protocol layer that currently owns the sk buff,
tail points at the current end of the protocol data. Again, this pointer varies depending
on the owning protocol layer,
end points at the end of the data area in memory. This is xed when the sk buff
is allocated.
10.5. The IP Layer 131
head
data
tail
end
next
prev
dev
Packet
to be
transmitted
sk_buff
truesize len
Figure 10.4: The Socket Buer (sk bu)
There are two length elds len and truesize, which describe the length of the current
protocol packet and the total size of the data buer respectively. The sk buff
handling code provides standard mechanisms for adding and removing protocol headers
and tails to the application data. These safely manipulate the data, tail and
len elds in the sk buff:
push This moves the data pointer towards the start of the data area and increments
the len eld. This is used when adding data or protocol headers to the start
of the data to be transmitted,
See skb push()
in include/-
linux/skbuff.h
pull This moves the data pointer away from the start, towards the end of the data
area and decrements the len eld. This is used when removing data or protocol
headers from the start of the data that has been received,
See skb pull()
in include/-
linux/skbuff.h
put This moves the tail pointer towards the end of the data area and increments
the len eld. This is used when adding data or protocol information to the
end of the data to be transmitted,
See skb put() in
include/linux/-
skbuff.h
trim This moves the tail pointer towards the start of the data area and decrements
the len eld. This is used when removing data or protocol tails from the
received packet.
See skb trim()
in include/-
linux/skbuff.h
The sk buff data structure also contains pointers that are used as it is stored in
doubly linked circular lists of sk buff's during processing. There are generic sk buff
routines for adding sk buffs to the front and back of these lists and for removing
them.
10.5.2 Receiving IP Packets
Chapter 8 described how Linux's network drivers built are into the kernel and initialized.
This results in a series of device data structures linked together in the
132 Chapter 10. Networks
dev base list. Each device data structure describes its device and provides a set of
callback routines that the network protocol layers call when they need the network
driver to perform work. These functions are mostly concerned with transmitting
data and with the network device's addresses. When a network device receives packets
from its network it must convert the received data into sk buff data structures.
These received sk buff's are added onto the backlog queue by the network drivers as
they are received. If the backlog queue grows too large, then the received sk buff's
See netif rx()
in
net/core/dev.c are discarded. The network bottom half is
agged as ready to run as there is work
to do.
When the network bottom half handler is run by the scheduler it processes any
network packets waiting to be transmitted before processing the backlog queue of
sk buff's determining which protocol layer to pass the received packets See net bh() in to. As the
net/core/dev.c
Linux networking layers were initialized, each protocol registered itself by adding a
packet type data structure onto either the ptype all list or into the ptype base
hash table. The packet type data structure contains the protocol type, a pointer
to a network device, a pointer to the protocol's receive data processing routine and,
nally, a pointer to the next packet type data structure in the list or hash chain.
The ptype all chain is used to snoop all packets being received from any network
device and is not normally used. The ptype base hash table is hashed by protocol
identier and is used to decide which protocol should receive the incoming network
packet. The network bottom half matches the protocol types of incoming sk buff's
against one or more of the packet type entries in either table. The protocol may
match more than one entry, for example when snooping all network trac, and in this
case the sk buff will be cloned. The sk buff is passed to the matching protocol's
handling routine.
See ip recv() in
net/ipv4/-
ip input.c
10.5.3 Sending IP Packets
Packets are transmitted by applications exchanging data or else they are generated by
the network protocols as they support established connections or connections being
established. Whichever way the data is generated, an sk buff is built to contain the
data and various headers are added by the protocol layers as it passes through them.
The sk buff needs to be passed to a network device to be transmitted. First though
the protocol, for example IP, needs to decide which network device to use. This
depends on the best route for the packet. For computers connected by modem to
a single network, say via the PPP protocol, the routing choice is easy. The packet
should either be sent to the local host via the loopback device or to the gateway at
the end of the PPP modem connection. For computers connected to an ethernet the
choices are harder as there are many computers connected to the network.
For every IP packet transmitted, IP uses the routing tables to resolve the route for
the destination IP address. Each IP destination successfully looked up in the routing
tables returns a rtable data structure describing the route to use. This includes
See
include/net/-
route.h the source IP address to use, the address of the network device data structure and,
sometimes, a prebuilt hardware header. This hardware header is network device
specic and contains the source and destination physical addresses and other media
specic information. If the network device is an ethernet device, the hardware header
would be as shown in Figure 10.1 and the source and destination addresses would be
physical ethernet addresses. The hardware header is cached with the route because
it must be appended to each IP packet transmitted on this route and constructing
10.6. The Address Resolution Protocol (ARP) 133
it takes time. The hardware header may contain physical addresses that have to be
resolved using the ARP protocol. In this case the outgoing packet is stalled until the
address has been resolved. Once it has been resolved and the hardware header built,
the hardware header is cached so that future IP packets sent using this interface do
not have to ARP.
10.5.4 Data Fragmentation
Every network device has a maximum packet size and it cannot transmit or receive a
data packet bigger than this. The IP protocol allows for this and will fragment data
into smaller units to t into the packet size that the network device can handle. The
IP protocol header includes a fragment eld which contains a
ag and the fragment
oset.
When an IP packet is ready to be transmited, IP nds the network device to send
See
ip build xmit()
in net/ipv4/
the IP packet out on. This device is found from the IP routing tables. Each device ip output.c
has a eld describing its maximum transfer unit (in bytes), this is the mtu eld. If
the device's mtu is smaller than the packet size of the IP packet that is waiting to
be transmitted, then the IP packet must be broken down into smaller (mtu sized)
fragments. Each fragment is represented by an sk buff; its IP header marked to
show that it is a fragment and what oset into the data this IP packet contains. The
last packet is marked as being the last IP fragment. If, during the fragmentation, IP
cannot allocate an sk buff, the transmit will fail.
Receiving IP fragments is a little more dicult than sending them because the IP
fragments can be received in any order and they must all be received before they can
be reassembled. Each time an IP packet is received it is checked to see if it is an IP
See ip rcv() in
net/ipv4/-
fragment. The rst time that the fragment of a message is received, IP creates a new ip input.c
ipq data structure, and this is linked into the ipqueue list of IP fragments awaiting
recombination. As more IP fragments are received, the correct ipq data structure is
found and a new ipfrag data structure is created to describe this fragment. Each
ipq data structure uniquely describes a fragmented IP receive frame with its source
and destination IP addresses, the upper layer protocol identier and the identier for
this IP frame. When all of the fragments have been received, they are combined into
a single sk buff and passed up to the next protocol level to be processed. Each ipq
contains a timer that is restarted each time a valid fragment is received. If this timer
expires, the ipq data structure and its ipfrag's are dismantled and the message is
presumed to have been lost in transit. It is then up to the higher level protocols to
retransmit the message.
10.6 The Address Resolution Protocol (ARP)
The Address Resolution Protocol's role is to provide translations of IP addresses into
physical hardware addresses such as ethernet addresses. IP needs this translation
just before it passes the data (in the form of an sk buff) to the device driver for
transmission. It performs various checks to see if this device needs a hardware header
See
ip build xmit()
in
net/ipv4/-
ip output.c
and, if it does, if the hardware header for the packet needs to be rebuilt. Linux caches
hardware headers to avoid frequent rebuilding of them. If the hardware header
needs rebuilding, it calls the device specic hardware header rebuilding routine. All
ethernet devices use the same generic header rebuilding routine which in turn uses
See eth
rebuild header()
in net/-
ethernet/eth.c
134 Chapter 10. Networks
the ARP services to translate the destination IP address into a physical address.
The ARP protocol itself is very simple and consists of two message types, an ARP
request and an ARP reply. The ARP request contains the IP address that needs
translating and the reply (hopefully) contains the translated IP address, the hardware
address. The ARP request is broadcast to all hosts connected to the network, so,
for an ethernet network, all of the machines connected to the ethernet will see the
ARP request. The machine that owns the IP address in the request will respond to
the ARP request with an ARP reply containing its own physical address.
The ARP protocol layer in Linux is built around a table of arp table data structures
which each describe an IP to physical address translation. These entries are created
as IP addresses need to be translated and removed as they become stale over time.
Each arp table data structure has the following elds:
last used the time that this ARP entry was last used,
last updated the time that this ARP entry was last updated,

ags these describe this entry's state, if it is complete and so on,
IP address The IP address that this entry describes
hardware address The translated hardware address
hardware header This is a pointer to a cached hardware header,
timer This is a timer list entry used to time out ARP requests
that do not get a response,
retries The number of times that this ARP request has been
retried,
sk buff queue List of sk buff entries waiting for this IP address
to be resolved
The ARP table consists of a table of pointers (the arp tables vector) to chains of
arp table entries. The entries are cached to speed up access to them, each entry is
found by taking the last two bytes of its IP address to generate an index into the
table and then following the chain of entries until the correct one is found. Linux also
caches prebuilt hardware headers o the arp table entries in the form of hh cache
data structures.
When an IP address translation is requested and there is no corresponding arp table
entry, ARP must send an ARP request message. It creates a new arp table entry
in the table and queues the sk buff containing the network packet that needs the
address translation on the sk buff queue of the new entry. It sends out an ARP
request and sets the ARP expiry timer running. If there is no response then ARP
will retry the request a number of times and if there is still no response ARP will
remove the arp table entry. Any sk buff data structures queued waiting for the
IP address to be translated will be notied and it is up to the protocol layer that is
transmitting them to cope with this failure. UDP does not care about lost packets
but TCP will attempt to retransmit on an established TCP link. If the owner of
the IP address responds with its hardware address, the arp table entry is marked
as complete and any queued sk buff's will be removed from the queue and will go
on to be transmitted. The hardware address is written into the hardware header of
each sk buff.
The ARP protocol layer must also respond to ARP requests that specfy its IP address.
It registers its protocol type (ETH P ARP), generating a packet type data
structure. This means that it will be passed all ARP packets that are received by
the network devices. As well as ARP replies, this includes ARP requests. It gener10.7.
IP Routing 135
ates an ARP reply using the hardware address kept in the receiving device's device
data structure.
Network topologies can change over time and IP addresses can be reassigned to
dierent hardware addresses. For example, some dial up services assign an IP address
as each connection is established. In order that the ARP table contains up to date
entries, ARP runs a periodic timer which looks through all of the arp table entries
to see which have timed out. It is very careful not to remove entries that contain
one or more cached hardware headers. Removing these entries is dangerous as other
data structures rely on them. Some arp table entries are permanent and these are
marked so that they will not be deallocated. The ARP table cannot be allowed to
grow too large; each arp table entry consumes some kernel memory. Whenever the
a new entry needs to be allocated and the ARP table has reached its maximum size
the table is pruned by searching out the oldest entries and removing them.
10.7 IP Routing
The IP routing function determines where to send IP packets destined for a particular
IP address. There are many choices to be made when transmitting IP packets. Can
the destination be reached at all? If it can be reached, which network device should
be used to transmit it? If there is more than one network device that could be used
to reach the destination, which is the better one? The IP routing database maintains
information that gives answers to these questions. There are two databases, the most
important being the Forwarding Information Database. This is an exhaustive list of
known IP destinations and their best routes. A smaller and much faster database,
the route cache is used for quick lookups of routes for IP destinations. Like all caches,
it must contain only the frequently accessed routes; its contents are derived from the
Forwarding Information Database.
Routes are added and deleted via IOCTL requests to the BSD socket interface.
These are passed onto the protocol to process. The INET protocol layer only allows
processes with superuser privileges to add and delete IP routes. These routes can be
xed or they can be dynamic and change over time. Most systems use xed routes
unless they themselves are routers. Routers run routing protocols which constantly
check on the availability of routes to all known IP destinations. Systems that are
not routers are known as end systems. The routing protocols are implemented as
daemons, for example GATED, and they also add and delete routes via the IOCTL
BSD socket interface.
10.7.1 The Route Cache
Whenever an IP route is looked up, the route cache is rst checked for a matching
route. If there is no matching route in the route cache the Forwarding Information
Database is searched for a route. If no route can be found there, the IP packet will fail
to be sent and the application notied. If a route is in the Forwarding Information
Database and not in the route cache, then a new entry is generated and added into
the route cache for this route. The route cache is a table (ip rt hash table) that
contains pointers to chains of rtable data structures. The index into the route table
is a hash function based on the least signicant two bytes of the IP address. These
are the two bytes most likely to be dierent between destinations and provide the
136 Chapter 10. Networks
fz_list
fz_nent
fz_logmask
fz_hash_table
fz_next
fz_mask
fib_zone
fib_dst
fib_use
fib_info
fib_metric
fib_next
fib_tos
fib_node
fib_prev
fib_gateway
fib_dev
fib_refcnt
fib_window
fib_flags
fib_mtu
fib_irtt
fib_next
fib_info
fib_dst
fib_use
fib_info
fib_metric
fib_next
fib_tos
fib_node
fib_prev
fib_gateway
fib_dev
fib_refcnt
fib_window
fib_flags
fib_mtu
fib_irtt
fib_next
fib_info
fib_zones
Figure 10.5: The Forwarding Information Database
best spread of hash values. Each rtable entry contains information about the route;
the destination IP address, the network device to use to reach that IP address, the
maximum size of message that can be used and so on. It also has a reference count,
a usage count and a timestamp of the last time that they were used (in jiffies).
The reference count is incremented each time the route is used to show the number
of network connections using this route. It is decremented as applications stop using
the route. The usage count is incremented each time the route is looked up and is
used to order the rtable entry in its chain of hash entries. The last used timestamp
for all of the entries in the route cache is periodically checked to see if the rtable
is too old . If the route has not been recently used, it is discarded from the route
See ip rt
check expire()
in net/-
ipv4/route.c cache. If routes are kept in the route cache they are ordered so that the most used
entries are at the front of the hash chains. This means that nding them will be
quicker when routes are looked up.
10.7.2 The Forwarding Information Database
The forwarding information database (shown in Figure 10.5 contains IP's view of the
routes available to this system at this time. It is quite a complicated data structure
and, although it is reasonably eciently arranged, it is not a quick database to
consult. In particular it would be very slow to look up destinations in this database
for every IP packet transmitted. This is the reason that the route cache exists: to
speed up IP packet transmission using known good routes. The route cache is derived
from the forwarding database and represents its commonly used entries.
Each IP subnet is represented by a fib zone data structure. All of these are pointed
at from the fib zones hash table. The hash index is derived from the IP subnet
10.7. IP Routing 137
mask. All routes to the same subnet are described by pairs of fib node and fib info
data structures queued onto the fz list of each fib zone data structure. If the
number of routes in this subnet grows large, a hash table is generated to make
nding the fib node data structures easier.
Several routes may exist to the same IP subnet and these routes can go through one
of several gateways. The IP routing layer does not allow more than one route to a
subnet using the same gateway. In other words, if there are several routes to a subnet,
then each route is guaranteed to use a dierent gateway. Associated with each route
is its metric. This is a measure of how advantagious this route is. A route's metric
is, essentially, the number of IP subnets that it must hop across before it reaches the
destination subnet. The higher the metric, the worse the route.
138 Chapter 10. Networks
Chapter 11
Kernel Mechanisms
This chapter describes some of the general tasks and mechanisms that
the Linux kernel needs to supply so that other parts of the kernel work
eectively together.
11.1 Bottom Half Handling
There are often times in a kernel when you do not want to do work at this moment.
A good example of this is during interrupt processing. When the interrupt was
asserted, the processor stopped what it was doing and the operating system delivered
the interrupt to the appropriate device driver. Device drivers should not spend too
much time handling interrupts as, during this time, nothing else in the system can
bh_active
31 0
31 0
bh_mask
bh_base
31
0 Bottom half handler
(timers)
Figure 11.1: Bottom Half Handling Data Structures
139
140 Chapter 11. Kernel Mechanisms
run. There is often some work that could just as well be done later on. Linux's
bottom half handlers were invented so that device drivers and other parts of the
Linux kernel could queue work to be done later on. Figure 11.1 shows the kernel
data structures associated with bottom half handling. There can be up to 32 dierent
See
include/linux/-
interrupt.h bottom half handlers; bh base is a vector of pointers to each of the kernel's bottom
half handling routines. bh active and bh mask have their bits set according to what
handlers have been installed and are active. If bit N of bh mask is set then the
Nth element of bh base contains the address of a bottom half routine. If bit N of
bh active is set then the N'th bottom half handler routine should be called as soon
as the scheduler deems reasonable. These indices are statically dened; the timer
bottom half handler is the highest priority (index 0), the console bottom half handler
is next in priority (index 1) and so on. Typically the bottom half handling routines
have lists of tasks associated with them. For example, the immediate bottom half
handler works its way through the immediate tasks queue (tq immediate) which
contains tasks that need to be performed immediately.
Some of the kernel's bottom half handers are device specic, but others are more
generic:
TIMER This handler is marked as active each time the system's periodic timer
interrupts and is used to drive the kernel's timer queue mechanisms,
CONSOLE This handler is used to process console messages,
TQUEUE This handler is used to process tty messages,
NET This handler handles general network processing,
IMMEDIATE This is a generic handler used by several device drivers to queue
work to be done later.
Whenever a device driver, or some other part of the kernel, needs to schedule work to
be done later, it adds work to the appropriate system queue, for example the timer
queue, and then signals the kernel that some bottom half handling needs to be done.
It does this by setting the appropriate bit in bh active. Bit 8 is set if the driver
has queued something on the immediate queue and wishes the immediate bottom
half handler to run and process it. The bh active bitmask is checked at the end of
each system call, just before control is returned to the calling process. If it has any
bits set, the bottom half handler routines that are active are called. Bit 0 is checked
rst, then 1 and so on until bit 31. The bit in bh active is cleared as each bottom
See
do bottom half()
in kernel/-
softirq.c half handling routine is called. bh active is transient; it only has meaning between
calls to the scheduler and is a way of not calling bottom half handling routines when
there is no work for them to do.
11.2 Task Queues
Task queues are the kernel's way of deferring work until later. Linux has a generic
mechanism for queuing work on queues and for processing them later. Task queues
See
include/linux/-
tqueue.h are often used in conjunction with bottom half handlers; the timer task queue is
processed when the timer queue bottom half handler runs. A task queue is a simple
data structure, see gure 11.2 which consists of a singly linked list of tq struct data
structures each of which contains the address of a routine and a pointer to some data.
11.3. Timers 141
*data
next
*routine()
tq_struct
sync
*data
next
*routine()
tq_struct
sync
task queue
Figure 11.2: A Task Queue
The routine will be called when the element on the task queue is processed and it
will be passed a pointer to the data.
Anything in the kernel, for example a device driver, can create and use task queues
but there are three task queues created and managed by the kernel:
timer This queue is used to queue work that will be done as soon after the next
system clock tick as is possible. Each clock tick, this queue is checked to see if it
contains any entries and, if it does, the timer queue bottom half handler is made
active. The timer queue bottom half handler is processed, along with all the
other bottom half handlers, when the scheduler next runs. This queue should
not be confused with system timers, which are a much more sophisticated
mechanism.
immediate This queue is also processed when the scheduler processes the active
bottom half handlers. The immediate bottom half handler is not as high in
priority as the timer queue bottom half handler and so these tasks will be run
later.
scheduler This task queue is processed directly by the scheduler. It is used to
support other task queues in the system and, in this case, the task to be run
will be a routine that processes a task queue, say for a device driver.
When task queues are processed, the pointer to the rst element in the queue is
removed from the queue and replaced with a null pointer. In fact, this removal is
an atomic operation, one that cannot be interrupted. Then each element in the
queue has its handling routine called in turn. The elements in the queue are often
statically allocated data. However there is no inherent mechanism for discarding
allocated memory. The task queue processing routine simply moves onto the next
element in the list. It is the job of the task itself to ensure that it properly cleans up
any allocated kernel memory.
11.3 Timers
An operating system needs to be able to schedule an activity sometime in the future.
A mechanism is needed whereby activities can be scheduled to run at some relatively
precise time. Any microprocessor that wishes to support an operating system must
have a programmable interval timer that periodically interrupts the processor. This
periodic interrupt is known as a system clock tick and it acts like a metronome,
orchestrating the system's activities. Linux has a very simple view of what time it
is; it measures time in clock ticks since the system booted. All system times are
based on this measurement, which is known as jiffies after the globally available
variable of the same name.
142 Chapter 11. Kernel Mechanisms
31 0
timer_active
next
prev
expires
data
*function()
timer_list
next
prev
expires
data
*function()
timer_list
expires
*fn()
timer_struct
expires
*fn()
timer_struct
31
0
timer_table
next
prev
expires
data
*function()
timer_head
Figure 11.3: System Timers
Linux has two types of system timers, both queue routines to be called at some
system time but they are slightly dierent in their implementations. Figure 11.3
shows both mechanisms. The rst, the old timer mechanism, has a static array of 32
See
include/linux/-
timer.h pointers to timer struct data structures and a mask of active timers, timer active.
Where the timers go in the timer table is statically dened (rather like the bottom
half handler table bh base). Entries are added into this table mostly at system
initialization time. The second, newer, mechanism uses a linked list of timer list
data structures held in ascending expiry time order.
Both methods use the time in jiffies as an expiry time so that a timer that wished
to run in 5s would have to convert 5s to units of jiffies and add that to the current
system time to get the system time in jiffies when the timer should expire. Every
system clock tick the timer bottom half handler is marked as active so that the
when the scheduler next runs, the timer queues will be processed. The timer bottom
half handler processes both types of system timer. For the old system timers the
See timer bh()
in
kernel/sched.c timer active bit mask is check for bits that are set. If the expiry time for an active
See
run old timers()
in
kernel/sched.c
timer has expired (expiry time is less than the current system jiffies), its timer
routine is called and its active bit is cleared. For new system timers, the entries in
the linked list of timer list data structures are checked. Every expired timer is
See
run timer list()
in
kernel/sched.c
removed from the list and its routine is called. The new timer mechanism has the
advantage of being able to pass an argument to the timer routine.
11.4 Wait Queues
There are many times when a process must wait for a system resource. For example
a process may need the VFS inode describing a directory in the le system and that
inode may not be in the buer cache. In this case the process must wait for that
inode to be fetched from the physical media containing the le system before it can
carry on.
The Linux kernel uses a simple data structure, a wait queue (see gure 11.4), which
See
include/linux/-
wait.h
11.5. Buzz Locks 143
wait queue
*task
*next
Figure 11.4: Wait Queue
consists of a pointer to the processes task struct and a pointer to the next element
in the wait queue.
When processes are added to the end of a wait queue they can either be interruptible
or uninterruptible. Interruptible processes may be interrupted by events
such as timers expiring or signals being delivered whilst they are waiting on a wait
queue. The waiting processes state will re
ect this and either be INTERRUPTIBLE or
UNINTERRUPTIBLE. As this process can not now continue to run, the scheduler is run
and, when it selects a new process to run, the waiting process will be suspended. 1
When the wait queue is processed, the state of every process in the wait queue is
set to RUNNING. If the process has been removed from the run queue, it is put back
onto the run queue. The next time the scheduler runs, the processes that are on
the wait queue are now candidates to be run as they are now no longer waiting.
When a process on the wait queue is scheduled the rst thing that it will do is
remove itself from the wait queue. Wait queues can be used to synchronize access
to system resources and they are used by Linux in its implementation of semaphores
(see below).
11.5 Buzz Locks
These are better known as spin locks and they are a primitive way of protecting a
data structure or piece of code. They only allow one process at a time to be within
a critical region of code. They are used in Linux to restrict access to elds in data
structures, using a single integer eld as a lock. Each process wishing to enter the
region attempts to change the lock's initial value from 0 to 1. If its current value is
1, the process tries again, spinning in a tight loop of code. The access to the memory
location holding the lock must be atomic, the action of reading its value, checking
that it is 0 and then changing it to 1 cannot be interrupted by any other process.
Most CPU architectures provide support for this via special instructions but you can
also implement buzz locks using uncached main memory.
When the owning process leaves the critical region of code it decrements the buzz
lock, returning its value to 0. Any processes spinning on the lock will now read it as
0, the rst one to do this will increment it to 1 and enter the critical region.
11.6 Semaphores
Semaphores are used to protect critical regions of code or data structures. Remember
that each access of a critical piece of data such as a VFS inode describing a directory
is made by kernel code running on behalf of a process. It would be very dangerous
to allow one process to alter a critical data structure that is being used by another
1 REVIEW NOTE: What is to stop a task in state INTERRUPTIBLE being made to run the next
time the scheduler runs? Processes in a wait queue should never run until they are woken up.
144 Chapter 11. Kernel Mechanisms
process. One way to achieve this would be to use a buzz lock around the critical piece
of data is being accessed but this is a simplistic approach that would not give very
good system performance. Instead Linux uses semaphores to allow just one process
at a time to access critical regions of code and data; all other processes wishing to
access this resource will be made to wait until it becomes free. The waiting processes
are suspended, other processes in the system can continue to run as normal.
A Linux semaphore data structure contains the following information:
See
include/asm/-
semaphore.h
count This eld keeps track of the count of processes wishing to use this resource.
A positive value means that the resource is available. A negative or zero value
means that processes are waiting for it. An initial value of 1 means that one
and only one process at a time can use this resource. When processes want
this resource they decrement the count and when they have nished with this
resource they increment the count,
waking This is the count of processes waiting for this resource which is also the
number of process waiting to be woken up when this resource becomes free,
wait queue When processes are waiting for this resource they are put onto this
wait queue,
lock A buzz lock used when accessing the waking eld.
Suppose the initial count for a semaphore is 1, the rst process to come along will
see that the count is positive and decrement it by 1, making it 0. The process now
\owns" the critical piece of code or resource that is being protected by the semaphore.
When the process leaves the critical region it increments the semphore's count. The
most optimal case is where there are no other processes contending for ownership of
the critical region. Linux has implemented semaphores to work eciently for this,
the most common, case.
If another process wishes to enter the critical region whilst it is owned by a process
it too will decrement the count. As the count is now negative (-1) the process cannot
enter the critical region. Instead it must wait until the owning process exits it.
Linux makes the waiting process sleep until the owning process wakes it on exiting
the critical region. The waiting process adds itself to the semaphore's wait queue
and sits in a loop checking the value of the waking eld and calling the scheduler
until waking is non-zero.
The owner of the critical region increments the semaphore's count and if it is less
than or equal to zero then there are processes sleeping, waiting for this resource. In
the optimal case the semaphore's count would have been returned to its initial value
of 1 and no further work would be neccessary. The owning process increments the
waking counter and wakes up the process sleeping on the semaphore's wait queue.
When the waiting process wakes up, the waking counter is now 1 and it knows that
it it may now enter the critical region. It decrements the waking counter, returning
it to a value of zero, and continues. All access to the waking eld of semaphore are
protected by a buzz lock using the semaphore's lock.
Chapter 12
Modules
This chapter describes how the Linux kernel can dynamically load functions,
for example lesystems, only when they are needed.
Linux is a monolithic kernel; that is, it is one, single, large program where all the
functional components of the kernel have access to all of its internal data structures
and routines. The alternative is to have a micro-kernel structure where the
functional pieces of the kernel are broken out into separate units with strict communication
mechanisms between them. This makes adding new components into
the kernel via the conguration process rather time consuming. Say you wanted to
use a SCSI driver for an NCR 810 SCSI and you had not built it into the kernel.
You would have to congure and then build a new kernel before you could use the
NCR 810. There is an alternative, Linux allows you to dynamically load and unload
components of the operating system as you need them. Linux modules are lumps of
code that can be dynamically linked into the kernel at any point after the system has
booted. They can be unlinked from the kernel and removed when they are no longer
needed. Mostly Linux kernel modules are device drivers, pseudo-device drivers such
as network drivers, or le-systems.
You can either load and unload Linux kernel modules explicitly using the insmod and
rmmod commands or the kernel itself can demand that the kernel daemon (kerneld)
loads and unloads the modules as they are needed. Dynamically loading code as it
is needed is attractive as it keeps the kernel size to a minimum and makes the kernel
very
exible. My current Intel kernel uses modules extensively and is only 406Kbytes
long. I only occasionally use VFAT le systems and so I build my Linux kernel to
automatically load the VFAT le system module as I mount a VFAT partition. When
I have unmounted the VFAT partition the system detects that I no longer need the
VFAT le system module and removes it from the system. Modules can also be useful
145
146 Chapter 12. Modules
for trying out new kernel code without having to rebuild and reboot the kernel every
time you try it out. Nothing, though, is for free and there is a slight performance
and memory penalty associated with kernel modules. There is a little more code that
a loadable module must provide and this and the extra data structures take a little
more memory. There is also a level of indirection introduced that makes accesses of
kernel resources slightly less ecient for modules.
Once a Linux module has been loaded it is as much a part of the kernel as any normal
kernel code. It has the same rights and responsibilities as any kernel code; in other
words, Linux kernel modules can crash the kernel just like all kernel code or device
drivers can.
So that modules can use the kernel resources that they need, they must be able to nd
them. Say a module needs to call kmalloc(), the kernel memory allocation routine.
At the time that it is built, a module does not know where in memory kmalloc() is,
so when the module is loaded, the kernel must x up all of the module's references
to kmalloc() before the module can work. The kernel keeps a list of all of the
kernel's resources in the kernel symbol table so that it can resolve references to those
resources from the modules as they are loaded. Linux allows module stacking, this
is where one module requires the services of another module. For example, the VFAT
le system module requires the services of the FAT le system module as the VFAT
le system is more or less a set of extensions to the FAT le system. One module
requiring services or resources from another module is very similar to the situation
where a module requires services and resources from the kernel itself. Only here
the required services are in another, previously loaded module. As each module is
loaded, the kernel modies the kernel symbol table, adding to it all of the resources
or symbols exported by the newly loaded module. This means that, when the next
module is loaded, it has access to the services of the already loaded modules.
When an attempt is made to unload a module, the kernel needs to know that the
module is unused and it needs some way of notifying the module that it is about to
be unloaded. That way the module will be able to free up any system resources that
it has allocated, for example kernel memory or interrupts, before it is removed from
the kernel. When the module is unloaded, the kernel removes any symbols that that
module exported into the kernel symbol table.
Apart from the ability of a loaded module to crash the operating system by being
badly written, it presents another danger. What happens if you load a module built
for an earlier or later kernel than the one that you are now running? This may
cause a problem if, say, the module makes a call to a kernel routine and supplies the
wrong arguments. The kernel can optionally protect against this by making rigorous
version checks on the module as it is loaded.
12.1 Loading a Module
There are two ways that a kernel module can be loaded. The rst way is to use the
insmod command to manually insert the it into the kernel. The second, and much
more clever way, is to load the module as it is needed; this is known as demand
loading. When the kernel discovers the need for a module, for example when the
user mounts a le system that is not in the kernel, the kernel will request that the
kernel daemon (kerneld) attempts to load the appropriate module.
kerneld is in the
modules package
along with
insmod, lsmod
and rmmod.
12.1. Loading a Module 147
ref
symtab
module
next
size
addr
*cleanup()
state
name
ref
symtab
module
next
size
addr
*cleanup()
state
name "vfat"
symbol_table
size
n_symbols
n_refs
symbols
references
symbol_table
size
n_symbols
n_refs
symbols
references
module_list
"fat"
Figure 12.1: The List of Kernel Modules
The kernel daemon is a normal user process albeit with super user privileges. When
it is started up, usually at system boot time, it opens up an Inter-Process Communication
(IPC) channel to the kernel. This link is used by the kernel to send messages
to the kerneld asking for various tasks to be performed. Kerneld's major function
See
include/linux/-
is to load and unload kernel modules but it is also capable of other tasks such as kerneld.h
starting up the PPP link over serial line when it is needed and closing it down when it
is not. Kerneld does not perform these tasks itself, it runs the neccessary programs
such as insmod to do the work. Kerneld is just an agent of the kernel, scheduling
work on its behalf.
The insmod utility must nd the requested kernel module that it is to load. Demand
loaded kernel modules are normally kept in /lib/modules/kernel-version. The
kernel modules are linked object les just like other programs in the system except
that they are linked as a relocatable images. That is, images that are not linked
to run from a particular address. They can be either a.out or elf format object
les. insmod makes a privileged system call to nd the kernel's exported symbols.
These are kept in pairs containing the symbol's name and its value, for example its
See sys get -
kernel syms() in
address. The kernel's exported symbol table is held in the rst module data structure kernel/module.c
in the list of modules maintained by the kernel and pointed at by the module list
pointer. Only specically entered symbols are added into the table, which is built
See
include/linux/-
when the kernel is compiled and linked, not every symbol in the kernel is exported module.h
to its modules. An example symbol is ``request irq'' which is the kernel routine
that must be called when a driver wishes to take control of a particular system
interrupt. In my current kernel, this has a value of 0x0010cd30. You can easily see
the exported kernel symbols and their values by looking at /proc/ksyms or by using
the ksyms utility. The ksyms utility can either show you all of the exported kernel
symbols or only those symbols exported by loaded modules. insmod reads the module
into its virtual memory and xes up its unresolved references to kernel routines and
148 Chapter 12. Modules
resources using the exported symbols from the kernel. This xing up takes the form
of patching the module image in memory. insmod physically writes the address of
the symbol into the appropriate place in the module.
When insmod has xed up the module's references to exported kernel symbols, it asks
the kernel for enough space to hold the new kernel, again using a privileged system
call. The kernel allocates a new module data structure and enough kernel memory
to hold the new module and puts it at the end of the kernel modules list. The new
module is marked as UNINITIALIZED. Figure 12.1 shows the list of kernel modules
See sys -
create module()
in kernel/-
module.c. after two modules, VFAT and VFAT have been loaded into the kernel. Not shown in the
diagram is the rst module on the list, which is a pseudo-module that is only there
to hold the kernel's exported symbol table. You can use the command lsmod to list
all of the loaded kernel modules and their interdependencies. lsmod simply reformats
/proc/modules which is built from the list of kernel module data structures. The
memory that the kernel allocates for it is mapped into the insmod process's address
space so that it can access it. insmod copies the module into the allocated space and
relocates it so that it will run from the kernel address that it has been allocated.
This must happen as the module cannot expect to be loaded at the same address
twice let alone into the same address in two dierent Linux systems. Again, this
relocation involves patching the module image with the appropriate addresses.
The new module also exports symbols to the kernel and insmod builds a table of
these exported images. Every kernel module must contain module initialization and
module cleanup routines and these symbols are deliberately not exported but insmod
must know the addresses of them so that it can pass them to the kernel. All being
well, insmod is now ready to initialize the module and it makes a privileged system call
passing the kernel the addresses of the module's initialization and cleanup routines.
See sys init -
module() in
kernel/-
module.c. When a new module is added into the kernel, it must update the kernel's set of
symbols and modify the modules that are being used by the new module. Modules
that have other modules dependent on them must maintain a list of references at the
end of their symbol table and pointed at by their module data structure. Figure 12.1
shows that the VFAT le system module is dependent on the FAT le system module.
So, the FAT module contains a reference to the VFAT module; the reference was
added when the VFAT module was loaded. The kernel calls the modules initialization
routine and, if it is successful it carries on installing the module. The module's
cleanup routine address is stored in it's module data structure and it will be called
by the kernel when that module is unloaded. Finally, the module's state is set to
RUNNING.
12.2 Unloading a Module
Modules can be removed using the rmmod command but demand loaded modules are
automatically removed from the system by kerneld when they are no longer being
used. Every time its idle timer expires, kerneld makes a system call requesting
that all unused demand loaded modules are removed from the system. The timer's
value is set when you start kerneld; my kerneld checks every 180 seconds. So,
for example, if you mount an iso9660 CD ROM and your iso9660 lesystem is a
loadable module, then shortly after the CD ROM is unmounted, the iso9660 module
will be removed from the kernel.
A module cannot be unloaded so long as other components of the kernel are depend12.2.
Unloading a Module 149
ing on it. For example, you cannot unload the VFAT module if you have one or more
VFAT le systems mounted. If you look at the output of lsmod, you will see that each
module has a count associated with it. For example:
Module: #pages: Used by:
msdos 5 1
vfat 4 1 (autoclean)
fat 6 [vfat msdos] 2 (autoclean)
The count is the number of kernel entities that are dependent on this module. In the
above example, the vfat and msdos modules are both dependent on the fat module
and so it has a count of 2. Both the vfat and msdos modules have 1 dependent,
which is a mounted le system. If I were to load another VFAT le system then the
vfat module's count would become 2. A module's count is held in the rst longword
of its image.
This eld is slightly overloaded as it also holds the AUTOCLEAN and VISITED
ags.
Both of these
ags are used for demand loaded modules. These modules are marked
as AUTOCLEAN so that the system can recognize which ones it may automatically
unload. The VISITED
ag marks the module as in use by one or more other system
components; it is set whenever another component makes use of the module. Each
time the system is asked by kerneld to remove unused demand loaded modules it
looks through all of the modules in the system for likely candidates. It only looks
at modules marked as AUTOCLEAN and in the state RUNNING. If the candidate has
its VISITED
ag cleared then it will remove the module, otherwise it will clear the
VISITED
ag and go on to look at the next module in the system.
Assuming that a module can be unloaded, its cleanup routine is called to allow it
to free up the kernel resources that it has allocated. The module data structure is
See sys -
delete module()
in
kernel/module.c marked as DELETED and it is unlinked from the list of kernel modules. Any other
modules that it is dependent on have their reference lists modied so that they no
longer have it as a dependent. All of the kernel memory that the module needed is
deallocated.
150 Chapter 12. Modules
Chapter 13
Processors
Linux runs on a number of processors; this chapter gives a brief outline
of each of them.
13.1 X86
TBD
13.2 ARM
The ARM processor implements a low power, high performance 32 bit RISC architecture.
It is being widely used in embedded devices such as mobile phones and PDAs
(Personal Data Assistants). It has 31 32 bit registers with 16 visible in any mode. Its
instructions are simple load and store instructions (load a value from memory, perform
an operation and store the result back into memory). One interesting feature
it has is that every instruction is conditional. For example, you can test the value
of a register and, until you next test for the same condition, you can conditionally
execute instructions as and when you like. Another interesting feature is that you
can perform arithmetic and shift operations on values as you load them. It operates
in several modes, including a system mode that can be entered from user mode via
a SWI (software interrupt).
It is a synthasisable core and ARM (the company) does not itself manufacture processors.
Instead the ARM partners (companies such as Intel or LSI for example)
implement the ARM architecture in silicon. It allows other processors to be tightly
151
152 Chapter 13. Processors
coupled via a co-processor interface and it has several memory management unit
variations. These range from simple memory protection schemes to complex page
hierarchies.
13.3 Alpha AXP Processor
The Alpha AXP architecture is a 64-bit load/store RISC architecture designed with
speed in mind. All registers are 64 bits in length; 32 integer registers and 32
oating
point registers. Integer register 31 and
oating point register 31 are used for null
operations. A read from them generates a zero value and a write to them has no eect.
All instructions are 32 bits long and memory operations are either reads or writes.
The architecture allows dierent implementations so long as the implementations
follow the architecture.
There are no instructions that operate directly on values stored in memory; all data
manipulation is done between registers. So, if you want to increment a counter
in memory, you rst read it into a register, then modify it and write it out. The
instructions only interact with each other by one instruction writing to a register
or memory location and another register reading that register or memory location.
One interesting feature of Alpha AXP is that there are instructions that can generate

ags, such as testing if two registers are equal, the result is not stored in a processor
status register, but is instead stored in a third register. This may seem strange at
rst, but removing this dependency from a status register means that it is much
easier to build a CPU which can issue multiple instructions every cycle. Instructions
on unrelated registers do not have to wait for each other to execute as they would if
there were a single status register. The lack of direct operations on memory and the
large number of registers also help issue multiple instructions.
The Alpha AXP architecture uses a set of subroutines, called privileged architecture
library code (PALcode). PALcode is specic to the operating system, the CPU
implementation of the Alpha AXP architecture and to the system hardware. These
subroutines provide operating system primitives for context switching, interrupts,
exceptions and memory management. These subroutines can be invoked by hardware
or by CALL PAL instructions. PALcode is written in standard Alpha AXP assembler
with some implementation specic extensions to provide direct access to low level
hardware functions, for example internal processor registers. PALcode is executed in
PALmode, a privileged mode that stops some system events happening and allows
the PALcode complete control of the physical system hardware.
Chapter 14
The Linux Kernel Sources
This chapter describes where in the Linux kernel sources you should start
looking for particular kernel functions.
This book does not depend on a knowledge of the 'C' programming language or
require that you have the Linux kernel sources available in order to understand how
the Linux kernel works. That said, it is a fruitful exercise to look at the kernel sources
to get an in-depth understanding of the Linux operating system. This chapter gives
an overview of the kernel sources; how they are arranged and where you might start
to look for particular code.
Where to Get The Linux Kernel Sources
All of the major Linux distributions (Craftworks, Debian, Slackware, Red Hat etcetera)
include the kernel sources in them. Usually the Linux kernel that got installed on
your Linux system was built from those sources. By their very nature these sources
tend to be a little out of date so you may want to get the latest sources from one of
the web sites mentioned in chapter B. They are kept on ftp://ftp.cs.helsinki.fi
and all of the other web sites shadow them. This makes the Helsinki web site the
most up to date, but sites like MIT and Sunsite are never very far behind.
If you do not have access to the web, there are many CD ROM vendors who oer
snapshots of the world's major web sites at a very reasonable cost. Some even oer a
subscription service with quarterly or even monthly updates. Your local Linux User
Group is also a good source of sources.
The Linux kernel sources have a very simple numbering system. Any even number
kernel (for example 2.0.30) is a stable, released, kernel and any odd numbered
153
154 Chapter 14. The Linux Kernel Sources
kernel (for example 2.1.42 is a development kernel. This book is based on the
stable 2.0.30 source tree. Development kernels have all of the latest features and
support all of the latest devices. Although they can be unstable, which may not be
exactly what you want it, is important that the Linux community tries the latest
kernels. That way they are tested for the whole community. Remember that it is
always worth backing up your system thoroughly if you do try out non-production
kernels.
Changes to the kernel sources are distributed as patch les. The patch utility is used
to apply a series of edits to a set of source les. So, for example, if you have the
2.0.29 kernel source tree and you wanted to move to the 2.0.30 source tree, you would
obtain the 2.0.30 patch le and apply the patches (edits) to that source tree:
$ cd /usr/src/linux
$ patch -p1 < patch-2.0.30
This saves copying whole source trees, perhaps over slow serial connections. A good
source of kernel patches (ocial and unocial) is the http://www.linuxhq.com web
site.
How The Kernel Sources Are Arranged
At the very top level of the source tree /usr/src/linux you will see a number of
directories:
arch The arch subdirectory contains all of the architecture specic kernel code. It
has further subdirectories, one per supported architecture, for example i386
and alpha.
include The include subdirectory contains most of the include les needed to build
the kernel code. It too has further subdirectories including one for every architecture
supported. The include/asm subdirectory is a soft link to the real
include directory needed for this architecture, for example include/asm-i386.
To change architectures you need to edit the kernel makele and rerun the
Linux kernel conguration program.
init This directory contains the initialization code for the kernel and it is a very
good place to start looking at how the kernel works.
mm This directory contains all of the memory management code. The architecture
specic memory management code lives down in arch/*/mm/, for example
arch/i386/mm/fault.c.
drivers All of the system's device drivers live in this directory. They are further
sub-divided into classes of device driver, for example block.
ipc This directory contains the kernels inter-process communications code.
modules This is simply a directory used to hold built modules.
fs All of the le system code. This is further sub-divided into directories, one per
supported le system, for example vfat and ext2.
155
kernel The main kernel code. Again, the architecture specic kernel code is in
arch/*/kernel.
net The kernel's networking code.
lib This directory contains the kernel's library code. The architecture specic library
code can be found in arch/*/lib/.
scripts This directory contains the scripts (for example awk and tk scripts) that are
used when the kernel is congured.
Where to Start Looking
A large complex program like the Linux kernel can be rather daunting to look at.
It is rather like a large ball of string with no end showing. Looking at one part of
the kernel often leads to looking at several other related les and before long you
have forgotten what you were looking for. The next subsections give you a hint as
to where in the source tree the best place to look is for a given subject.
System Startup and Initialization
On an Intel based system, the kernel starts when either loadlin.exe or LILO has
loaded the kernel into memory and passed control to it. Look in arch/i386/-
kernel/head.S for this part. Head.S does some architecture specic setup and
then jumps to the main() routine in init/main.c.
Memory Management
This code is mostly in mm but the architecture specic code is in arch/*/mm. The
page fault handling code is in mm/memory.c and the memory mapping and page cache
code is in mm/filemap.c. The buer cache is implemented in mm/buffer.c and the
swap cache in mm/swap state.c and mm/swapfile.c.
Kernel
Most of the relevent generic code is in kernel with the architecture specic code
in arch/*/kernel. The scheduler is in kernel/sched.c and the fork code is in
kernel/fork.c. The bottom half handling code is in include/linux/interrupt.h.
The task struct data structure can be found in include/linux/sched.h.
PCI
The PCI pseudo driver is in drivers/pci/pci.c with the system wide denitions in
include/linux/pci.h. Each architecture has some specic PCI BIOS code, Alpha
AXP's is in arch/alpha/kernel/bios32.c.
156 Chapter 14. The Linux Kernel Sources
Interprocess Communication
This is all in ipc. All System V IPC objects include an ipc perm data structure and
this can be found in include/linux/ipc.h. System V messages are implemented in
ipc/msg.c, shared memory in ipc/shm.c and semaphores in ipc/sem.c. Pipes are
implemented in ipc/pipe.c.
Interrupt Handling
The kernel's interrupt handling code is almost all microprocessor (and often platform)
specic. The Intel interrupt handling code is in arch/i386/kernel/irq.c and its
denitions in include/asm-i386/irq.h.
Device Drivers
Most of the lines of the Linux kernel's source code are in its device drivers. All of
Linux's device driver sources are held in drivers but these are further broken out
by type:
/block block device drivers such as ide (in ide.c). If you want to look at how
all of the devices that could possibly contain le systems are initialized then
you should look at device setup() in drivers/block/genhd.c. It not only
initializes the hard disks but also the network as you need a network to mount
nfs le systems. Block devices include both IDE and SCSI based devices.
/char This the place to look for character based devices such as ttys, serial ports
and mice.
/cdrom All of the CDROM code for Linux. It is here that the special CDROM
devices (such as Soundblaster CDROM) can be found. Note that the ide CD
driver is ide-cd.c in drivers/block and that the SCSI CD driver is in scsi.c
in drivers/scsi.
/pci This are the sources for the PCI pseudo-driver. A good place to look at how
the PCI subsystem is mapped and initialized. The Alpha AXP PCI xup code
is also worth looking at in arch/alpha/kernel/bios32.c.
/scsi This is where to nd all of the SCSI code as well as all of the drivers for the
scsi devices supported by Linux.
/net This is where to look to nd the network device drivers such as the DECChip
21040 PCI ethernet driver which is in tulip.c.
/sound This is where all of the sound card drivers are.
File Systems
The sources for the EXT2 le system are all in the fs/ext2/ directory with data structure
denitions in include/linux/ext2 fs.h, ext2 fs i.h and ext2 fs sb.h. The
Virtual File System data structures are described in include/linux/fs.h and the
code is in fs/*. The buer cache is implemented in fs/buffer.c along with the
update kernel daemon.
157
Network
The networking code is kept in net with most of the include les in include/net.
The BSD socket code is in net/socket.c and the IP version 4 INET socket code is
in net/ipv4/af inet.c. The generic protocol support code (including the sk buff
handling routines) is in net/core with the TCP/IP networking code in net/ipv4.
The network device drivers are in drivers/net.
Modules
The kernel module code is partially in the kernel and partially in the modules package.
The kernel code is all in kernel/modules.c with the data structures and kernel
demon kerneld messages in include/linux/module.h and include/linux/-
kerneld.h respectively. You may want to look at the structure of an ELF object le
in include/linux/elf.h.
158 Chapter 14. The Linux Kernel Sources
Appendix A
Linux Data Structures
This appendix lists the major data structures that Linux uses and which are described
in this book. They have been edited slightly to t the paper.
block dev struct
block dev struct data structures are used to register block devices as available for
use by the buer cache. They are held together in the blk dev vector.
See
include/linux/
blkdev.h
struct blk_dev_struct {
void (*request_fn)(void);
struct request * current_request;
struct request plug;
struct tq_struct plug_tq;
};
buer head
The buffer head data structure holds information about a block buer in the buer
cache.
See
include/linux/
fs.h
/* bh state bits */
#define BH_Uptodate 0 /* 1 if the buffer contains valid data */
#define BH_Dirty 1 /* 1 if the buffer is dirty */
#define BH_Lock 2 /* 1 if the buffer is locked */
#define BH_Req 3 /* 0 if the buffer has been invalidated */
#define BH_Touched 4 /* 1 if the buffer has been touched (aging) */
#define BH_Has_aged 5 /* 1 if the buffer has been aged (aging) */
159
160 Appendix A. Linux Data Structures
#define BH_Protected 6 /* 1 if the buffer is protected */
#define BH_FreeOnIO 7 /* 1 to discard the buffer_head after IO */
struct buffer_head {
/* First cache line: */
unsigned long b_blocknr; /* block number */
kdev_t b_dev; /* device (B_FREE = free) */
kdev_t b_rdev; /* Real device */
unsigned long b_rsector; /* Real buffer location on disk */
struct buffer_head *b_next; /* Hash queue list */
struct buffer_head *b_this_page; /* circular list of buffers in one
page */
/* Second cache line: */
unsigned long b_state; /* buffer state bitmap (above) */
struct buffer_head *b_next_free;
unsigned int b_count; /* users using this block */
unsigned long b_size; /* block size */
/* Non-performance-critical data follows. */
char *b_data; /* pointer to data block */
unsigned int b_list; /* List that this buffer appears */
unsigned long b_flushtime; /* Time when this (dirty) buffer
* should be written */
unsigned long b_lru_time; /* Time when this buffer was
* last used. */
struct wait_queue *b_wait;
struct buffer_head *b_prev; /* doubly linked hash list */
struct buffer_head *b_prev_free; /* doubly linked list of buffers */
struct buffer_head *b_reqnext; /* request queue */
};
device
Every network device in the system is represented by a device data structure.
See
include/linux/
netdevice.h
struct device
{
/*
* This is the first field of the "visible" part of this structure
* (i.e. as seen by users in the "Space.c" file). It is the name
* the interface.
*/
char *name;
/* I/O specific fields */
unsigned long rmem_end; /* shmem "recv" end */
unsigned long rmem_start; /* shmem "recv" start */
unsigned long mem_end; /* shared mem end */
unsigned long mem_start; /* shared mem start */
unsigned long base_addr; /* device I/O address */
unsigned char irq; /* device IRQ number */
161
/* Low-level status flags. */
volatile unsigned char start, /* start an operation */
interrupt; /* interrupt arrived */
unsigned long tbusy; /* transmitter busy */
struct device *next;
/* The device initialization function. Called only once. */
int (*init)(struct device *dev);
/* Some hardware also needs these fields, but they are not part of
the usual set specified in Space.c. */
unsigned char if_port; /* Selectable AUI,TP, */
unsigned char dma; /* DMA channel */
struct enet_statistics* (*get_stats)(struct device *dev);
/*
* This marks the end of the "visible" part of the structure. All
* fields hereafter are internal to the system, and may change at
* will (read: may be cleaned up at will).
*/
/* These may be needed for future network-power-down code. */
unsigned long trans_start; /* Time (jiffies) of
last transmit */
unsigned long last_rx; /* Time of last Rx */
unsigned short flags; /* interface flags (BSD)*/
unsigned short family; /* address family ID */
unsigned short metric; /* routing metric */
unsigned short mtu; /* MTU value */
unsigned short type; /* hardware type */
unsigned short hard_header_len; /* hardware hdr len */
void *priv; /* private data */
/* Interface address info. */
unsigned char broadcast[MAX_ADDR_LEN];
unsigned char pad;
unsigned char dev_addr[MAX_ADDR_LEN];
unsigned char addr_len; /* hardware addr len */
unsigned long pa_addr; /* protocol address */
unsigned long pa_brdaddr; /* protocol broadcast addr*/
unsigned long pa_dstaddr; /* protocol P-P other addr*/
unsigned long pa_mask; /* protocol netmask */
unsigned short pa_alen; /* protocol address len */
struct dev_mc_list *mc_list; /* M'cast mac addrs */
int mc_count; /* No installed mcasts */
struct ip_mc_list *ip_mc_list; /* IP m'cast filter chain */
__u32 tx_queue_len; /* Max frames per queue */
/* For load balancing driver pair support */
unsigned long pkt_queue; /* Packets queued */
struct device *slave; /* Slave device */
162 Appendix A. Linux Data Structures
struct net_alias_info *alias_info; /* main dev alias info */
struct net_alias *my_alias; /* alias devs */
/* Pointer to the interface buffers. */
struct sk_buff_head buffs[DEV_NUMBUFFS];
/* Pointers to interface service routines. */
int (*open)(struct device *dev);
int (*stop)(struct device *dev);
int (*hard_start_xmit) (struct sk_buff *skb,
struct device *dev);
int (*hard_header) (struct sk_buff *skb,
struct device *dev,
unsigned short type,
void *daddr,
void *saddr,
unsigned len);
int (*rebuild_header)(void *eth,
struct device *dev,
unsigned long raddr,
struct sk_buff *skb);
void (*set_multicast_list)(struct device *dev);
int (*set_mac_address)(struct device *dev,
void *addr);
int (*do_ioctl)(struct device *dev,
struct ifreq *ifr,
int cmd);
int (*set_config)(struct device *dev,
struct ifmap *map);
void (*header_cache_bind)(struct hh_cache **hhp,
struct device *dev,
unsigned short htype,
__u32 daddr);
void (*header_cache_update)(struct hh_cache *hh,
struct device *dev,
unsigned char * haddr);
int (*change_mtu)(struct device *dev,
int new_mtu);
struct iw_statistics* (*get_wireless_stats)(struct device *dev);
};
device struct
device struct data structures are used to register character and block devices (they
hold its name and the set of le operations that can be used for this device). Each
valid member of the chrdevs and blkdevs vectors represents a character or block
device respectivelySee .
fs/devices.c
struct device_struct {
const char * name;
struct file_operations * fops;
};
163
le
Each open le, socket etcetera is represented by a file data structure.
See
include/linux/
fs.h
struct file {
mode_t f_mode;
loff_t f_pos;
unsigned short f_flags;
unsigned short f_count;
unsigned long f_reada, f_ramax, f_raend, f_ralen, f_rawin;
struct file *f_next, *f_prev;
int f_owner; /* pid or -pgrp where SIGIO should be sent */
struct inode * f_inode;
struct file_operations * f_op;
unsigned long f_version;
void *private_data; /* needed for tty driver, and maybe others */
};
les struct
The files struct data structure describes the les that a process has open.
See
include/linux/
sched.h
struct files_struct {
int count;
fd_set close_on_exec;
fd_set open_fds;
struct file * fd[NR_OPEN];
};
fs struct See
include/linux/
sched.h
struct fs_struct {
int count;
unsigned short umask;
struct inode * root, * pwd;
};
gendisk
The gendisk data structure holds information about a hard disk. They are used
during initialization when the disks are found and then probed for partitions.
See
include/linux/
genhd.h
struct hd_struct {
long start_sect;
long nr_sects;
};
struct gendisk {
int major; /* major number of driver */
const char *major_name; /* name of major driver */
int minor_shift; /* number of times minor is shifted to
get real minor */
int max_p; /* maximum partitions per device */
164 Appendix A. Linux Data Structures
int max_nr; /* maximum number of real devices */
void (*init)(struct gendisk *);
/* Initialization called before we
do our thing */
struct hd_struct *part; /* partition table */
int *sizes; /* device size in blocks, copied to
blk_size[] */
int nr_real; /* number of real devices */
void *real_devices; /* internal use */
struct gendisk *next;
};
inode
The VFS inode data structure holds information about a le or directory on disk.
See
include/linux/
fs.h
struct inode {
kdev_t i_dev;
unsigned long i_ino;
umode_t i_mode;
nlink_t i_nlink;
uid_t i_uid;
gid_t i_gid;
kdev_t i_rdev;
off_t i_size;
time_t i_atime;
time_t i_mtime;
time_t i_ctime;
unsigned long i_blksize;
unsigned long i_blocks;
unsigned long i_version;
unsigned long i_nrpages;
struct semaphore i_sem;
struct inode_operations *i_op;
struct super_block *i_sb;
struct wait_queue *i_wait;
struct file_lock *i_flock;
struct vm_area_struct *i_mmap;
struct page *i_pages;
struct dquot *i_dquot[MAXQUOTAS];
struct inode *i_next, *i_prev;
struct inode *i_hash_next, *i_hash_prev;
struct inode *i_bound_to, *i_bound_by;
struct inode *i_mount;
unsigned short i_count;
unsigned short i_flags;
unsigned char i_lock;
unsigned char i_dirt;
unsigned char i_pipe;
unsigned char i_sock;
unsigned char i_seek;
unsigned char i_update;
165
unsigned short i_writecount;
union {
struct pipe_inode_info pipe_i;
struct minix_inode_info minix_i;
struct ext_inode_info ext_i;
struct ext2_inode_info ext2_i;
struct hpfs_inode_info hpfs_i;
struct msdos_inode_info msdos_i;
struct umsdos_inode_info umsdos_i;
struct iso_inode_info isofs_i;
struct nfs_inode_info nfs_i;
struct xiafs_inode_info xiafs_i;
struct sysv_inode_info sysv_i;
struct affs_inode_info affs_i;
struct ufs_inode_info ufs_i;
struct socket socket_i;
void *generic_ip;
} u;
};
ipc perm
The ipc perm data structure describes the access permissions of a System V IPC
object .
See
include/linux/
ipc.h
struct ipc_perm
{
key_t key;
ushort uid; /* owner euid and egid */
ushort gid;
ushort cuid; /* creator euid and egid */
ushort cgid;
ushort mode; /* access modes see mode flags below */
ushort seq; /* sequence number */
};
irqaction
The irqaction data structure is used to describe the system's interrupt handlers.
See
include/linux/
interrupt.h
struct irqaction {
void (*handler)(int, void *, struct pt_regs *);
unsigned long flags;
unsigned long mask;
const char *name;
void *dev_id;
struct irqaction *next;
};
linux binfmt
Each binary le format that Linux understands is represented by a linux binfmt
data structure.
See
include/linux/
binfmts.h
166 Appendix A. Linux Data Structures
struct linux_binfmt {
struct linux_binfmt * next;
long *use_count;
int (*load_binary)(struct linux_binprm *, struct pt_regs * regs);
int (*load_shlib)(int fd);
int (*core_dump)(long signr, struct pt_regs * regs);
};
mem map t
The mem map t data structure (also known as page) is used to hold information about
each page of physical memory.
See
include/linux/
mm.h
typedef struct page {
/* these must be first (free area handling) */
struct page *next;
struct page *prev;
struct inode *inode;
unsigned long offset;
struct page *next_hash;
atomic_t count;
unsigned flags; /* atomic flags, some possibly
updated asynchronously */
unsigned dirty:16,
age:8;
struct wait_queue *wait;
struct page *prev_hash;
struct buffer_head *buffers;
unsigned long swap_unlock_entry;
unsigned long map_nr; /* page->map_nr == page - mem_map */
} mem_map_t;
mm struct
The mm struct data structure is used to describe the virtual memory of a task or
process.
See
include/linux/
sched.h
struct mm_struct {
int count;
pgd_t * pgd;
unsigned long context;
unsigned long start_code, end_code, start_data, end_data;
unsigned long start_brk, brk, start_stack, start_mmap;
unsigned long arg_start, arg_end, env_start, env_end;
unsigned long rss, total_vm, locked_vm;
unsigned long def_flags;
struct vm_area_struct * mmap;
struct vm_area_struct * mmap_avl;
struct semaphore mmap_sem;
};
167
pci bus
Every PCI bus in the system is represented by a pci bus data structure.
See
include/linux/
pci.h
struct pci_bus {
struct pci_bus *parent; /* parent bus this bridge is on */
struct pci_bus *children; /* chain of P2P bridges on this bus */
struct pci_bus *next; /* chain of all PCI buses */
struct pci_dev *self; /* bridge device as seen by parent */
struct pci_dev *devices; /* devices behind this bridge */
void *sysdata; /* hook for sys-specific extension */
unsigned char number; /* bus number */
unsigned char primary; /* number of primary bridge */
unsigned char secondary; /* number of secondary bridge */
unsigned char subordinate; /* max number of subordinate buses */
};
pci dev
Every PCI device in the system, including PCI-PCI and PCI-ISA bridge devices is
represented by a pci dev data structure.
See
include/linux/
pci.h
/*
* There is one pci_dev structure for each slot-number/function-number
* combination:
*/
struct pci_dev {
struct pci_bus *bus; /* bus this device is on */
struct pci_dev *sibling; /* next device on this bus */
struct pci_dev *next; /* chain of all devices */
void *sysdata; /* hook for sys-specific extension */
unsigned int devfn; /* encoded device & function index */
unsigned short vendor;
unsigned short device;
unsigned int class; /* 3 bytes: (base,sub,prog-if) */
unsigned int master : 1; /* set if device is master capable */
/*
* In theory, the irq level can be read from configuration
* space and all would be fine. However, old PCI chips don't
* support these registers and return 0 instead. For example,
* the Vision864-P rev 0 chip can uses INTA, but returns 0 in
* the interrupt line and pin registers. pci_init()
* initializes this field with the value at PCI_INTERRUPT_LINE
* and it is the job of pcibios_fixup() to change it if
* necessary. The field must not be 0 unless the device
* cannot generate interrupts at all.
*/
unsigned char irq; /* irq generated by this device */
};
168 Appendix A. Linux Data Structures
request
request data structures are used to make requests to the block devices in the system.
The requests are always to read or write blocks of data to or from the buer cache.
See
include/linux/
blkdev.h
struct request {
volatile int rq_status;
#define RQ_INACTIVE (-1)
#define RQ_ACTIVE 1
#define RQ_SCSI_BUSY 0xffff
#define RQ_SCSI_DONE 0xfffe
#define RQ_SCSI_DISCONNECTING 0xffe0
kdev_t rq_dev;
int cmd; /* READ or WRITE */
int errors;
unsigned long sector;
unsigned long nr_sectors;
unsigned long current_nr_sectors;
char * buffer;
struct semaphore * sem;
struct buffer_head * bh;
struct buffer_head * bhtail;
struct request * next;
};
rtable
Each rtable data structure holds information about the route to take in order to
send packets to an IP host. rtable data structures are used within the IP route
cache.
See
include/net/
route.h
struct rtable
{
struct rtable *rt_next;
__u32 rt_dst;
__u32 rt_src;
__u32 rt_gateway;
atomic_t rt_refcnt;
atomic_t rt_use;
unsigned long rt_window;
atomic_t rt_lastuse;
struct hh_cache *rt_hh;
struct device *rt_dev;
unsigned short rt_flags;
unsigned short rt_mtu;
unsigned short rt_irtt;
unsigned char rt_tos;
};
semaphore
Semaphores are used to protect critical data structures and regions of code. y
See
include/asm/
semaphore.h
169
struct semaphore {
int count;
int waking;
int lock ; /* to make waking testing atomic */
struct wait_queue *wait;
};
sk bu
The sk buff data structure is used to describe network data as it moves between
the layers of protocol.
See
include/linux/
skbuff.h
struct sk_buff
{
struct sk_buff *next; /* Next buffer in list */
struct sk_buff *prev; /* Previous buffer in list */
struct sk_buff_head *list; /* List we are on */
int magic_debug_cookie;
struct sk_buff *link3; /* Link for IP protocol level buffer chains */
struct sock *sk; /* Socket we are owned by */
unsigned long when; /* used to compute rtt's */
struct timeval stamp; /* Time we arrived */
struct device *dev; /* Device we arrived on/are leaving by */
union
{
struct tcphdr *th;
struct ethhdr *eth;
struct iphdr *iph;
struct udphdr *uh;
unsigned char *raw;
/* for passing file handles in a unix domain socket */
void *filp;
} h;
union
{
/* As yet incomplete physical layer views */
unsigned char *raw;
struct ethhdr *ethernet;
} mac;
struct iphdr *ip_hdr; /* For IPPROTO_RAW */
unsigned long len; /* Length of actual data */
unsigned long csum; /* Checksum */
__u32 saddr; /* IP source address */
__u32 daddr; /* IP target address */
__u32 raddr; /* IP next hop address */
__u32 seq; /* TCP sequence number */
__u32 end_seq; /* seq [+ fin] [+ syn] + datalen */
__u32 ack_seq; /* TCP ack sequence number */
unsigned char proto_priv[16];
volatile char acked, /* Are we acked ? */
used, /* Are we in use ? */
free, /* How to free this buffer */
170 Appendix A. Linux Data Structures
arp; /* Has IP/ARP resolution finished */
unsigned char tries, /* Times tried */
lock, /* Are we locked ? */
localroute, /* Local routing asserted for this frame */
pkt_type, /* Packet class */
pkt_bridged, /* Tracker for bridging */
ip_summed; /* Driver fed us an IP checksum */
#define PACKET_HOST 0 /* To us */
#define PACKET_BROADCAST 1 /* To all */
#define PACKET_MULTICAST 2 /* To group */
#define PACKET_OTHERHOST 3 /* To someone else */
unsigned short users; /* User count - see datagram.c,tcp.c */
unsigned short protocol; /* Packet protocol from driver. */
unsigned int truesize; /* Buffer size */
atomic_t count; /* reference count */
struct sk_buff *data_skb; /* Link to the actual data skb */
unsigned char *head; /* Head of buffer */
unsigned char *data; /* Data head pointer */
unsigned char *tail; /* Tail pointer */
unsigned char *end; /* End pointer */
void (*destructor)(struct sk_buff *); /* Destruct function */
__u16 redirport; /* Redirect port */
};
sock
Each sock data structure holds protocol specic information about a BSD socket.
For example, for an INET (Internet Address Domain) socket this data structure
would hold all of the TCP/IP and UDP/IP specic information.
See
include/linux/
net.h
struct sock
{
/* This must be first. */
struct sock *sklist_next;
struct sock *sklist_prev;
struct options *opt;
atomic_t wmem_alloc;
atomic_t rmem_alloc;
unsigned long allocation; /* Allocation mode */
__u32 write_seq;
__u32 sent_seq;
__u32 acked_seq;
__u32 copied_seq;
__u32 rcv_ack_seq;
unsigned short rcv_ack_cnt; /* count of same ack */
__u32 window_seq;
__u32 fin_seq;
__u32 urg_seq;
__u32 urg_data;
__u32 syn_seq;
int users; /* user count */
/*
* Not all are volatile, but some are, so we
171
* might as well say they all are.
*/
volatile char dead,
urginline,
intr,
blog,
done,
reuse,
keepopen,
linger,
delay_acks,
destroy,
ack_timed,
no_check,
zapped,
broadcast,
nonagle,
bsdism;
unsigned long lingertime;
int proc;
struct sock *next;
struct sock **pprev;
struct sock *bind_next;
struct sock **bind_pprev;
struct sock *pair;
int hashent;
struct sock *prev;
struct sk_buff *volatile send_head;
struct sk_buff *volatile send_next;
struct sk_buff *volatile send_tail;
struct sk_buff_head back_log;
struct sk_buff *partial;
struct timer_list partial_timer;
long retransmits;
struct sk_buff_head write_queue,
receive_queue;
struct proto *prot;
struct wait_queue **sleep;
__u32 daddr;
__u32 saddr; /* Sending source */
__u32 rcv_saddr; /* Bound address */
unsigned short max_unacked;
unsigned short window;
__u32 lastwin_seq; /* sequence number when we last
updated the window we offer */
__u32 high_seq; /* sequence number when we did
current fast retransmit */
volatile unsigned long ato; /* ack timeout */
volatile unsigned long lrcvtime; /* jiffies at last data rcv */
volatile unsigned long idletime; /* jiffies at last rcv */
unsigned int bytes_rcv;
/*
* mss is min(mtu, max_window)
172 Appendix A. Linux Data Structures
*/
unsigned short mtu; /* mss negotiated in the syn's */
volatile unsigned short mss; /* current eff. mss - can change */
volatile unsigned short user_mss; /* mss requested by user in ioctl */
volatile unsigned short max_window;
unsigned long window_clamp;
unsigned int ssthresh;
unsigned short num;
volatile unsigned short cong_window;
volatile unsigned short cong_count;
volatile unsigned short packets_out;
volatile unsigned short shutdown;
volatile unsigned long rtt;
volatile unsigned long mdev;
volatile unsigned long rto;
volatile unsigned short backoff;
int err, err_soft; /* Soft holds errors that don't
cause failure but are the cause
of a persistent failure not
just 'timed out' */
unsigned char protocol;
volatile unsigned char state;
unsigned char ack_backlog;
unsigned char max_ack_backlog;
unsigned char priority;
unsigned char debug;
int rcvbuf;
int sndbuf;
unsigned short type;
unsigned char localroute; /* Route locally only */
/*
* This is where all the private (optional) areas that don't
* overlap will eventually live.
*/
union
{
struct unix_opt af_unix;
#if defined(CONFIG_ATALK) || defined(CONFIG_ATALK_MODULE)
struct atalk_sock af_at;
#endif
#if defined(CONFIG_IPX) || defined(CONFIG_IPX_MODULE)
struct ipx_opt af_ipx;
#endif
#ifdef CONFIG_INET
struct inet_packet_opt af_packet;
#ifdef CONFIG_NUTCP
struct tcp_opt af_tcp;
#endif
#endif
} protinfo;
/*
* IP 'private area'
*/
173
int ip_ttl; /* TTL setting */
int ip_tos; /* TOS */
struct tcphdr dummy_th;
struct timer_list keepalive_timer; /* TCP keepalive hack */
struct timer_list retransmit_timer; /* TCP retransmit timer */
struct timer_list delack_timer; /* TCP delayed ack timer */
int ip_xmit_timeout; /* Why the timeout is running */
struct rtable *ip_route_cache; /* Cached output route */
unsigned char ip_hdrincl; /* Include headers ? */
#ifdef CONFIG_IP_MULTICAST
int ip_mc_ttl; /* Multicasting TTL */
int ip_mc_loop; /* Loopback */
char ip_mc_name[MAX_ADDR_LEN]; /* Multicast device name */
struct ip_mc_socklist *ip_mc_list; /* Group array */
#endif
/*
* This part is used for the timeout functions (timer.c).
*/
int timeout; /* What are we waiting for? */
struct timer_list timer; /* This is the TIME_WAIT/receive
* timer when we are doing IP
*/
struct timeval stamp;
/*
* Identd
*/
struct socket *socket;
/*
* Callbacks
*/
void (*state_change)(struct sock *sk);
void (*data_ready)(struct sock *sk,int bytes);
void (*write_space)(struct sock *sk);
void (*error_report)(struct sock *sk);
};
socket
Each socket data structure holds information about a BSD socket. It does not exist
independently; it is, instead, part of the VFS inode data structure.
See
include/linux/
net.h
struct socket {
short type; /* SOCK_STREAM, ... */
socket_state state;
long flags;
struct proto_ops *ops; /* protocols do most everything */
void *data; /* protocol data */
struct socket *conn; /* server socket connected to */
struct socket *iconn; /* incomplete client conn.s */
struct socket *next;
struct wait_queue **wait; /* ptr to place to wait on */
struct inode *inode;
174 Appendix A. Linux Data Structures
struct fasync_struct *fasync_list; /* Asynchronous wake up list */
struct file *file; /* File back pointer for gc */
};
task struct
Each task struct data structure describes a process or task in the system.
See
include/linux/
sched.h
struct task_struct {
/* these are hardcoded - don't touch */
volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
long counter;
long priority;
unsigned long signal;
unsigned long blocked; /* bitmap of masked signals */
unsigned long flags; /* per process flags, defined below */
int errno;
long debugreg[8]; /* Hardware debugging registers */
struct exec_domain *exec_domain;
/* various fields */
struct linux_binfmt *binfmt;
struct task_struct *next_task, *prev_task;
struct task_struct *next_run, *prev_run;
unsigned long saved_kernel_stack;
unsigned long kernel_stack_page;
int exit_code, exit_signal;
/* ??? */
unsigned long personality;
int dumpable:1;
int did_exec:1;
int pid;
int pgrp;
int tty_old_pgrp;
int session;
/* boolean value for session group leader */
int leader;
int groups[NGROUPS];
/*
* pointers to (original) parent process, youngest child, younger sibling,
* older sibling, respectively. (p->father can be replaced with
* p->p_pptr->pid)
*/
struct task_struct *p_opptr, *p_pptr, *p_cptr,
*p_ysptr, *p_osptr;
struct wait_queue *wait_chldexit;
unsigned short uid,euid,suid,fsuid;
unsigned short gid,egid,sgid,fsgid;
unsigned long timeout, policy, rt_priority;
unsigned long it_real_value, it_prof_value, it_virt_value;
unsigned long it_real_incr, it_prof_incr, it_virt_incr;
struct timer_list real_timer;
long utime, stime, cutime, cstime, start_time;
/* mm fault and swap info: this can arguably be seen as either
mm-specific or thread-specific */
175
unsigned long min_flt, maj_flt, nswap, cmin_flt, cmaj_flt, cnswap;
int swappable:1;
unsigned long swap_address;
unsigned long old_maj_flt; /* old value of maj_flt */
unsigned long dec_flt; /* page fault count of the last time */
unsigned long swap_cnt; /* number of pages to swap on next pass */
/* limits */
struct rlimit rlim[RLIM_NLIMITS];
unsigned short used_math;
char comm[16];
/* file system info */
int link_count;
struct tty_struct *tty; /* NULL if no tty */
/* ipc stuff */
struct sem_undo *semundo;
struct sem_queue *semsleeping;
/* ldt for this task - used by Wine. If NULL, default_ldt is used */
struct desc_struct *ldt;
/* tss for this task */
struct thread_struct tss;
/* filesystem information */
struct fs_struct *fs;
/* open file information */
struct files_struct *files;
/* memory management info */
struct mm_struct *mm;
/* signal handlers */
struct signal_struct *sig;
#ifdef __SMP__
int processor;
int last_processor;
int lock_depth; /* Lock depth.
We can context switch in and out
of holding a syscall kernel lock... */
#endif
};
timer list
timer list data structure's are used to implement real time timers for processes.
See
include/linux/
timer.h
struct timer_list {
struct timer_list *next;
struct timer_list *prev;
unsigned long expires;
unsigned long data;
void (*function)(unsigned long);
};
tq struct
Each task queue (tq struct) data structure holds information about work that has
been queued. This is usually a task needed by a device driver but which does not
176 Appendix A. Linux Data Structures
have to be See done immediately.
include/linux/
tqueue.h struct tq_struct {
struct tq_struct *next; /* linked list of active bh's */
int sync; /* must be initialized to zero */
void (*routine)(void *); /* function to call */
void *data; /* argument to function */
};
vm area struct
Each vm area struct data structure describes an area of virtual memory for a process.
See
include/linux/
mm.h
struct vm_area_struct {
struct mm_struct * vm_mm; /* VM area parameters */
unsigned long vm_start;
unsigned long vm_end;
pgprot_t vm_page_prot;
unsigned short vm_flags;
/* AVL tree of VM areas per task, sorted by address */
short vm_avl_height;
struct vm_area_struct * vm_avl_left;
struct vm_area_struct * vm_avl_right;
/* linked list of VM areas per task, sorted by address */
struct vm_area_struct * vm_next;
/* for areas with inode, the circular list inode->i_mmap */
/* for shm areas, the circular list of attaches */
/* otherwise unused */
struct vm_area_struct * vm_next_share;
struct vm_area_struct * vm_prev_share;
/* more */
struct vm_operations_struct * vm_ops;
unsigned long vm_offset;
struct inode * vm_inode;
unsigned long vm_pte; /* shared mem */
};
Appendix B
Useful Web and FTP Sites
The following World Wide Web and ftp sites are useful:
http://www.azstarnet.com/~axplinux This is David Mosberger-Tang's Alpha
AXP Linux web site and it is the place to go for all of the Alpha AXP HOWTOs.
It also has a large number of pointers to Linux and Alpha AXP specic
information such as CPU data sheets.
http://www.redhat.com/ Red Hat's web site. This has a lot of useful pointers.
ftp://sunsite.unc.edu This is the major site for a lot of free software. The Linux
specic software is held in pub/Linux.
http://www.intel.com Intel's web site and a good place to look for Intel chip
information.
http://www.ssc.com/lj/index.html The Linux Journal is a very good Linux
magazine and well worth the yearly subscription for its excellent articles.
http://www.blackdown.org/java-linux.html This is the primary site for information
on Java on Linux.
ftp://tsx-11.mit.edu/ ~ftp/pub/linux MIT's Linux ftp site.
ftp://ftp.cs.helsinki./pub/Software/Linux/Kernel Linus's kernel sources.
http://www.linux.org.uk The UK Linux User Group.
http://sunsite.unc.edu/mdw/linux.html Home page for the Linux Documentation
Project,
177
178 Appendix B. Useful Web and FTP Sites
http://www.digital.com Digital Equipment Corporation's main web page.
http://altavista.digital.com DIGITAL's Altavista search engine. A very good
place to search for information within the web and news groups.
http://www.linuxhq.com The Linux HQ web site holds up to date ocial and
unocal patches as well as advice and web pointers that help you get the best
set of kernel sources possible for your system.
http://www.amd.com The AMD web site.
http://www.cyrix.com Cyrix's web site.
http://www.arm.com ARM's web site.
Appendix C
Linux Documentation Project
Manifesto
This is the Linux Documentation Project \Manifesto"
Last Revision 21 September 1998, by Michael K. Johnson
This le describes the goals and current status of the Linux Documentation
Project, including names of projects, volunteers, FTP sites, and so
on.
C.1 Overview
The Linux Documentation Project is working on developing good, reliable docs for
the Linux operating system. The overall goal of the LDP is to collaborate in taking
care of all of the issues of Linux documentation, ranging from online docs (man
pages, texinfo docs, and so on) to printed manuals covering topics such as installing,
using, and running Linux. The LDP is essentially a loose team of volunteers with
little central organization; anyone who is interested in helping is welcome to join in
the eort. We feel that working together and agreeing on the direction and scope
of Linux documentation is the best way to go, to reduce problems with con
icting
eorts|two people writing two books on the same aspect of Linux wastes someone's
time along the way.
The LDP is set out to produce the canonical set of Linux online and printed documentation.
Because our docs will be freely available (like software licensed under
the terms of the GNU GPL) and distributed on the net, we are able to easily update
179
180 Appendix C. Linux Documentation Project Manifesto
the documentation to stay on top of the many changes in the Linux world. If you
are interested in publishing any of the LDP works, see the section \Publishing LDP
Manuals", below.
C.2 Getting Involved
Send mail to linux-howto@metalab.unc.edu
Of course, you'll also need to get in touch with the coordinator of whatever LDP
projects you're interested in working on; see the next section.
C.3 Current Projects
For a list of current projects, see the LDP Homepage at http://sunsite.unc.edu/LDP/ldp.html.
The best way to get involved with one of these projects is to pick up the current
version of the manual and send revisions, editions, or suggestions to the coordinator.
You probably want to coordinate with the author before sending revisions so that
you know you are working together.
C.4 FTP sites for LDP works
LDP works can be found on sunsite.unc.edu in the directory /pub/Linux/docs. LDP
manuals are found in /pub/Linux/docs/LDP, HOWTOs and other documentation
found in /pub/Linux/docs/HOWTO.
C.5 Documentation Conventions
Here are the conventions that are currently used by LDP manuals. If you are interested
in writing another manual using dierent conventions, please let us know of
your plans rst.
The man pages { the Unix standard for online manuals { are created with the Unix
standard nro man (or BSD mdoc) macros.
The guides { full books produced by the LDP { have historically been done in LaTeX,
as their primary goal has been to printed documentation. However, guide authors
have been moving towards SGML with the DocBook DTD, because it allows them
to create more dierent kinds of output, both printed and on-line. If you use LaTeX,
we have a style le you can use to keep your printed look consistent with other LDP
documents, and we suggest that you use it.
The HOWTO documents are all required to be in SGML format. Currently, they
use the linuxdoc DTD, which is quite simple. There is a move afoot to switch to the
DocBook DTD over time.
LDP documents must be freely redistributable without fees paid to the authors. It
is not required that the text be modiable, but it is encouraged. You can come up
with your own license terms that satisfy this constraint, or you can use a previously
prepared license. The LDP provides a boilerplate license that you can use, some
people like to use the GPL, and others write their own.
C.6. Copyright and License 181
The copyright for each manual should be in the name of the head writer or coordinator
for the project. \The Linux Documentation Project" isn't a formal entity and
shouldn't be used to copyright the docs.
C.6 Copyright and License
Here is a \boilerplate" license you may apply to your work. It has not been reviewed
by a lawyer; feel free to have your own lawyer review it (or your modication of it) for
its applicability to your own desires. Remember that in order for your document to
be part of the LDP, you must allow unlimited reproduction and distribution without
fee.
This manual may be reproduced and distributed in whole or in part, without fee,
subject to the following conditions:
 The copyright notice above and this permission notice must be preserved complete
on all complete or partial copies.
 Any translation or derived work must be approved by the author in writing
before distribution.
 If you distribute this work in part, instructions for obtaining the complete
version of this manual must be included, and a means for obtaining a complete
version provided.
 Small portions may be reproduced as illustrations for reviews or quotes in other
works without this permission notice if proper citation is given.
Exceptions to these rules may be granted for academic purposes: Write to the author
and ask. These restrictions are here to protect us as authors, not to restrict you as
learners and educators.
All source code in this document is placed under the GNU General Public License,
available via anonymous FTP from prep.ai.mit.edu:/pub/gnu/COPYING.
C.7 Publishing LDP Manuals
If you're a publishing company interested in distributing any of the LDP manuals,
read on.
By the license requirements given previously, anyone is allowed to publish and distribute
verbatim copies of the Linux Documentation Project manuals. You don't
need our explicit permission for this. However, if you would like to distribute a
translation or derivative work based on any of the LDP manuals, you may need to
obtain permission from the author, in writing, before doing so, if the license requires
that.
You may, of course, sell the LDP manuals for prot. We encourage you to do so.
Keep in mind, however, that because the LDP manuals are freely distributable,
anyone may photocopy or distribute printed copies free of charge, if they wish to do
so.
We do not require to be paid royalties for any prot earned from selling LDP manuals.
However, we would like to suggest that if you do sell LDP manuals for prot, that
182 Appendix C. Linux Documentation Project Manifesto
you either oer the author royalties, or donate a portion of your earnings to the
author, the LDP as a whole, or to the Linux development community. You may also
wish to send one or more free copies of the LDP manuals that you are distributing
to the authors. Your show of support for the LDP and the Linux community will be
very much appreciated.
We would like to be informed of any plans to publish or distribute LDP manuals,
just so we know how they're becoming available. If you are publishing or planning
to publish any LDP manuals, please send mail to ldp-l@linux.org.au. It's nice to
know who's doing what.
We encourage Linux software distributors to distribute the LDP manuals (such as
the Installation and Getting Started Guide) with their software. The LDP manuals
are intended to be used as the "ocial" Linux documentation, and we are glad to see
mail-order distributors bundling the LDP manuals with the software. As the LDP
manuals mature, hopefully they will fulll this goal more and more adequately.
Appendix D
The GNU General Public
License
Printed below is the GNU General Public License (the GPL or copyleft), under
which Linux is licensed. It is reproduced here to clear up some of the confusion
about Linux's copyright status|Linux is not shareware, and it is not in the public
domain. The bulk of the Linux kernel is copyright c
1993 by Linus Torvalds, and
other software and parts of the kernel are copyrighted by their authors. Thus, Linux
is copyrighted, however, you may redistribute it under the terms of the GPL printed
below.
GNU GENERAL PUBLIC LICENSE
Version 2, June 1991
Copyright (C) 1989, 1991 Free Software Foundation, Inc. 675 Mass Ave, Cambridge,
MA 02139, USA Everyone is permitted to copy and distribute verbatim copies of
this license document, but changing it is not allowed.
D.1 Preamble
The licenses for most software are designed to take away your freedom to share and
change it. By contrast, the GNU General Public License is intended to guarantee
your freedom to share and change free software{to make sure the software is free
for all its users. This General Public License applies to most of the Free Software
Foundation's software and to any other program whose authors commit to using
183
184 Appendix D. The GNU General Public License
it. (Some other Free Software Foundation software is covered by the GNU Library
General Public License instead.) You can apply it to your programs, too.
When we speak of free software, we are referring to freedom, not price. Our General
Public Licenses are designed to make sure that you have the freedom to distribute
copies of free software (and charge for this service if you wish), that you receive
source code or can get it if you want it, that you can change the software or use
pieces of it in new free programs; and that you know you can do these things.
To protect your rights, we need to make restrictions that forbid anyone to deny
you these rights or to ask you to surrender the rights. These restrictions translate
to certain responsibilities for you if you distribute copies of the software, or if you
modify it.
For example, if you distribute copies of such a program, whether gratis or for a fee,
you must give the recipients all the rights that you have. You must make sure that
they, too, receive or can get the source code. And you must show them these terms
so they know their rights.
We protect your rights with two steps: (1) copyright the software, and (2) oer you
this license which gives you legal permission to copy, distribute and/or modify the
software.
Also, for each author's protection and ours, we want to make certain that everyone
understands that there is no warranty for this free software. If the software is modied
by someone else and passed on, we want its recipients to know that what they have
is not the original, so that any problems introduced by others will not re
ect on the
original authors' reputations.
Finally, any free program is threatened constantly by software patents. We wish to
avoid the danger that redistributors of a free program will individually obtain patent
licenses, in eect making the program proprietary. To prevent this, we have made it
clear that any patent must be licensed for everyone's free use or not licensed at all.
The precise terms and conditions for copying, distribution and modication follow.
D.2 Terms and Conditions for Copying, Distribution,
and Modication
0. This License applies to any program or other work which contains a notice
placed by the copyright holder saying it may be distributed under the terms of
this General Public License. The \Program", below, refers to any such program
or work, and a \work based on the Program" means either the Program or
any derivative work under copyright law: that is to say, a work containing
the Program or a portion of it, either verbatim or with modications and/or
translated into another language. (Hereinafter, translation is included without
limitation in the term \modication".) Each licensee is addressed as \you".
Activities other than copying, distribution and modication are not covered
by this License; they are outside its scope. The act of running the Program is
not restricted, and the output from the Program is covered only if its contents
constitute a work based on the Program (independent of having been made by
running the Program). Whether that is true depends on what the Program
does.
D.2. Terms and Conditions 185
1. You may copy and distribute verbatim copies of the Program's source code
as you receive it, in any medium, provided that you conspicuously and appropriately
publish on each copy an appropriate copyright notice and disclaimer
of warranty; keep intact all the notices that refer to this License and to the
absence of any warranty; and give any other recipients of the Program a copy
of this License along with the Program.
You may charge a fee for the physical act of transferring a copy, and you may
at your option oer warranty protection in exchange for a fee.
2. You may modify your copy or copies of the Program or any portion of it,
thus forming a work based on the Program, and copy and distribute such
modications or work under the terms of Section 1 above, provided that you
also meet all of these conditions:
a. You must cause the modied les to carry prominent notices stating that
you changed the les and the date of any change.
b. You must cause any work that you distribute or publish, that in whole or
in part contains or is derived from the Program or any part thereof, to
be licensed as a whole at no charge to all third parties under the terms of
this License.
c. If the modied program normally reads commands interactively when run,
you must cause it, when started running for such interactive use in the
most ordinary way, to print or display an announcement including an
appropriate copyright notice and a notice that there is no warranty (or
else, saying that you provide a warranty) and that users may redistribute
the program under these conditions, and telling the user how to view a
copy of this License. (Exception: if the Program itself is interactive but
does not normally print such an announcement, your work based on the
Program is not required to print an announcement.)
These requirements apply to the modied work as a whole. If identiable
sections of that work are not derived from the Program, and can be reasonably
considered independent and separate works in themselves, then this License,
and its terms, do not apply to those sections when you distribute them as
separate works. But when you distribute the same sections as part of a whole
which is a work based on the Program, the distribution of the whole must be
on the terms of this License, whose permissions for other licensees extend to
the entire whole, and thus to each and every part regardless of who wrote it.
Thus, it is not the intent of this section to claim rights or contest your rights
to work written entirely by you; rather, the intent is to exercise the right to
control the distribution of derivative or collective works based on the Program.
In addition, mere aggregation of another work not based on the Program with
the Program (or with a work based on the Program) on a volume of a storage
or distribution medium does not bring the other work under the scope of this
License.
3. You may copy and distribute the Program (or a work based on it, under Section
2) in object code or executable form under the terms of Sections 1 and 2 above
provided that you also do one of the following:
186 Appendix D. The GNU General Public License
a. Accompany it with the complete corresponding machine-readable source
code, which must be distributed under the terms of Sections 1 and 2 above
on a medium customarily used for software interchange; or,
b. Accompany it with a written oer, valid for at least three years, to give any
third party, for a charge no more than your cost of physically performing
source distribution, a complete machine-readable copy of the corresponding
source code, to be distributed under the terms of Sections 1 and 2
above on a medium customarily used for software interchange; or,
c. Accompany it with the information you received as to the oer to distribute
corresponding source code. (This alternative is allowed only for
noncommercial distribution and only if you received the program in object
code or executable form with such an oer, in accord with Subsection b
above.)
The source code for a work means the preferred form of the work for making
modications to it. For an executable work, complete source code means all
the source code for all modules it contains, plus any associated interface de-
nition les, plus the scripts used to control compilation and installation of the
executable. However, as a special exception, the source code distributed need
not include anything that is normally distributed (in either source or binary
form) with the major components (compiler, kernel, and so on) of the operating
system on which the executable runs, unless that component itself accompanies
the executable.
If distribution of executable or object code is made by oering access to copy
from a designated place, then oering equivalent access to copy the source code
from the same place counts as distribution of the source code, even though third
parties are not compelled to copy the source along with the object code.
4. You may not copy, modify, sublicense, or distribute the Program except as
expressly provided under this License. Any attempt otherwise to copy, modify,
sublicense or distribute the Program is void, and will automatically terminate
your rights under this License. However, parties who have received copies, or
rights, from you under this License will not have their licenses terminated so
long as such parties remain in full compliance.
5. You are not required to accept this License, since you have not signed it. However,
nothing else grants you permission to modify or distribute the Program
or its derivative works. These actions are prohibited by law if you do not accept
this License. Therefore, by modifying or distributing the Program (or any
work based on the Program), you indicate your acceptance of this License to
do so, and all its terms and conditions for copying, distributing or modifying
the Program or works based on it.
6. Each time you redistribute the Program (or any work based on the Program),
the recipient automatically receives a license from the original licensor to copy,
distribute or modify the Program subject to these terms and conditions. You
may not impose any further restrictions on the recipients' exercise of the rights
granted herein. You are not responsible for enforcing compliance by third
parties to this License.
D.2. Terms and Conditions 187
7. If, as a consequence of a court judgment or allegation of patent infringement
or for any other reason (not limited to patent issues), conditions are imposed
on you (whether by court order, agreement or otherwise) that contradict the
conditions of this License, they do not excuse you from the conditions of this
License. If you cannot distribute so as to satisfy simultaneously your obligations
under this License and any other pertinent obligations, then as a consequence
you may not distribute the Program at all. For example, if a patent license
would not permit royalty-free redistribution of the Program by all those who
receive copies directly or indirectly through you, then the only way you could
satisfy both it and this License would be to refrain entirely from distribution
of the Program.
If any portion of this section is held invalid or unenforceable under any particular
circumstance, the balance of the section is intended to apply and the
section as a whole is intended to apply in other circumstances.
It is not the purpose of this section to induce you to infringe any patents or
other property right claims or to contest validity of any such claims; this section
has the sole purpose of protecting the integrity of the free software distribution
system, which is implemented by public license practices. Many people have
made generous contributions to the wide range of software distributed through
that system in reliance on consistent application of that system; it is up to the
author/donor to decide if he or she is willing to distribute software through
any other system and a licensee cannot impose that choice.
This section is intended to make thoroughly clear what is believed to be a
consequence of the rest of this License.
8. If the distribution and/or use of the Program is restricted in certain countries
either by patents or by copyrighted interfaces, the original copyright holder
who places the Program under this License may add an explicit geographical
distribution limitation excluding those countries, so that distribution is permitted
only in or among countries not thus excluded. In such case, this License
incorporates the limitation as if written in the body of this License.
9. The Free Software Foundation may publish revised and/or new versions of the
General Public License from time to time. Such new versions will be similar in
spirit to the present version, but may dier in detail to address new problems
or concerns.
Each version is given a distinguishing version number. If the Program species
a version number of this License which applies to it and \any later version", you
have the option of following the terms and conditions either of that version or
of any later version published by the Free Software Foundation. If the Program
does not specify a version number of this License, you may choose any version
ever published by the Free Software Foundation.
10. If you wish to incorporate parts of the Program into other free programs whose
distribution conditions are dierent, write to the author to ask for permission.
For software which is copyrighted by the Free Software Foundation, write to the
Free Software Foundation; we sometimes make exceptions for this. Our decision
will be guided by the two goals of preserving the free status of all derivatives of
our free software and of promoting the sharing and reuse of software generally.
NO WARRANTY
188 Appendix D. The GNU General Public License
11. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE
IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED
BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED
IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
PROVIDE THE PROGRAM \AS IS" WITHOUT WARRANTY OF ANY
KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED
TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO
THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH
YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME
THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
12. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAWOR AGREED
TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER
PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE PROGRAM
AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING
ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL
DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE
THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA
OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED
BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO
OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER
OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
END OF TERMS AND CONDITIONS
D.3 Appendix: How to Apply These Terms to Your
New Programs
If you develop a new program, and you want it to be of the greatest possible use to
the public, the best way to achieve this is to make it free software which everyone
can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest to attach them
to the start of each source le to most eectively convey the exclusion of warranty;
and each le should have at least the \copyright" line and a pointer to where the full
notice is found.
hone line to give the program's name and a brief idea of what it does.i
Copyright c
19yy hname of authori
This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2 of the License, or (at your
option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
or FITNESS FOR A PARTICULAR PURPOSE. See
the GNU General Public License for more details.
D.3. How to Apply These Terms 189
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
675 Mass Ave, Cambridge, MA 02139, USA.
Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this when it starts
in an interactive mode:
Gnomovision version 69, Copyright (C) 19yy name of author Gnomovision
comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This
is free software, and you are welcome to redistribute it under certain
conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate
parts of the General Public License. Of course, the commands you use may be called
something other than `show w' and `show c'; they could even be mouse-clicks or
menu items{whatever suits your program.
You should also get your employer (if you work as a programmer) or your school, if
any, to sign a \copyright disclaimer" for the program, if necessary. Here is a sample;
alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the program
`Gnomovision' (which makes passes at compilers) written by James Hacker.
hsignature of Ty Cooni, 1 April 1989 Ty Coon, President of Vice
This General Public License does not permit incorporating your program into proprietary
programs. If your program is a subroutine library, you may consider it more
useful to permit linking proprietary applications with the library. If this is what you
want to do, use the GNU Library General Public License instead of this License.
190 Appendix D. The GNU General Public License
Glossary
Argument Functions and routines are passed arguments to process.
ARP Address Resolution Protocol. Used to translate IP addresses into physical
hardware addresses.
Ascii American Standard Code for Information Interchange. Each letter of the
alphabet is represented by an 8 bit code. Ascii is most often used to store
written characters.
Bit A single bit of data that represents either 1 or 0 (on or o).
Bottom Half Handler Handlers for work queued within the kernel.
Byte 8 bits of data,
C A high level programming language. Most of the Linux kernel is written in C.
CISC Complex Instruction Set Computer. The opposite of RISC, a processor which
supports a large number of often complex assembly instructions. The X86
architecture is a CISC architecture.
CPU Central Processing Unit. The main engine of the computer, see also microprocessor
and processor.
Data Structure This is a set of data in memory comprised of elds,
Device Driver The software controlling a particular device, for example the NCR
810 device driver controls the NCR 810 SCSI device.
DMA Direct Memory Access.
ELF Executable and Linkable Format. This object le format designed by the Unix
System Laboratories is now rmly established as the most commonly used
format in Linux.
191
192 Appendix D. Glossary
EIDE Extended IDE.
Executable image A structured le containing machine instructions and data.
This le can be loaded into a process's virtual memory and executed. See
also program.
Function A piece of software that performs an action. For example, returning the
bigger of two numbers.
IDE Integrated Disk Electronics.
Image See executable image.
IP Internet Protocol.
IPC Interprocess Communiction.
Interface A standard way of calling routines and passing data structures. For
example, the interface between two layers of code might be expressed in terms
of routines that pass and return a particular data structure. Linux's VFS is a
good example of an interface.
IRQ Interrupt Request Queue.
ISA Industry Standard Architecture. This is a standard, although now rather dated,
data bus interface for system components such as
oppy disk drivers.
Kernel Module A dynamically loaded kernel function such as a lesystem or a
device driver.
Kilobyte A thousand bytes of data, often written as Kbyte,
Megabyte A million bytes of data, often written as Mbyte,
Microprocessor A very integrated CPU. Most modern CPUs are Microprocessors.
Module A le containing CPU instructions in the form of either assembly language
instructions or a high level language like C.
Object le A le containing machine code and data that has not yet been linked
with other object les or libraries to become an executable image.
Page Physical memory is divided up into equal sized pages.
Pointer A location in memory that contains the address of another location in
memory,
Process This is an entity which can execute programs. A process could be thought
of as a program in action.
Processor Short for Microprocessor, equivalent to CPU.
PCI Peripheral Component Interconnect. A standard describing how the peripheral
components of a computer system may be connected together.
Peripheral An intelligent processor that does work on behalf of the system's CPU.
For example, an IDE controller chip,
193
Program A coherent set of CPU instructions that performs a task, such as printing
\hello world". See also executable image.
Protocol A protocol is a networking language used to transfer application data
between two cooperating processes or network layers.
Register A location within a chip, used to store information or instructions.
Register File The set of registers in a processor.
RISC Reduced Instruction Set Computer. The opposite of CISC, that is a processor
with a small number of assembly instructions, each of which performs simple
operations. The ARM and Alpha processors are both RISC architectures.
Routine Similar to a function except that, strictly speaking, routines do not return
values.
SCSI Small Computer Systems Interface.
Shell This is a program which acts as an interface between the operating system
and a human user. Also called a command shell, the most commonly used shell
in Linux is the bash shell.
SMP Symmetrical multiprocessing. Systems with more than one processor which
fairly share the work amongst those processors.
Socket A socket represents one end of a network connection, Linux supports the
BSD Socket interface.
Software CPU instructions (both assembler and high level languages like C) and
data. Mostly interchangable with Program.
System V A variant of UnixTM produced in 1983, which included, amongst other
things, System V IPC mechanisms.
TCP Transmission Control Protocol.
Task Queue A mechanism for deferring work in the Linux kernel.
UDP User Datagram Protocol.
Virtual memory A hardware and software mechanism for making the physical
memory in a system appear larger than it actually is.
194 Appendix D. Glossary
Bibliography
[1] Richard L. Sites. Alpha Architecture Reference Manual Digital Press
[2] Matt Welsh and Lar Kaufman. Running Linux O'Reilly & Associates, Inc, ISBN
1-56592-100-3
[3] PCI Special Interest Group PCI Local Bus Specication
[4] PCI Special Interest Group PCI BIOS ROM Specication
[5] PCI Special Interest Group PCI to PCI Bridge Architecture Specication
[6] Intel Peripheral Components Intel 296467, ISBN 1-55512-207-8
[7] Brian W. Kernighan and Dennis M. Richie The C Programming Language Prentice
Hall, ISBN 0-13-110362-8
[8] Steven Levy Hackers Penguin, ISBN 0-14-023269-9
[9] Intel Intel486 Processor Family: Programmer's Reference Manual Intel
[10] Comer D. E. Interworking with TCP/IP, Volume 1 - Principles, Protocols and
Architecture Prentice Hall International Inc
[11] David Jagger ARM Architectural Reference Manual Prentice Hall, ISBN 0-13-
736299-4
195