This commit is contained in:
gferg 2003-03-07 19:58:41 +00:00
parent ac79b6464b
commit 44316ea86b
18 changed files with 4815 additions and 0 deletions

View File

@ -0,0 +1,49 @@
<sect1><title>Acknowledgements</title>
<para>Ori Pomerantz would like to thank Yoav Weiss for many helpful ideas
and discussions, as well as finding mistakes within this document before
its publication. Ori would also like to thank Frodo Looijaard from the
Netherlands, Stephen Judd from New Zealand, Magnus Ahltorp from Sweeden and
Emmanuel Papirakis from Quebec, Canada.</para>
<para>I'd like to thank Ori Pomerantz for authoring this guide in the first
place and then letting me maintain it. It was a tremendous effort on his
part. I hope he likes what I've done with this document.</para>
<para> I would also like to thank Jeff Newmiller and Rhonda Bailey for
teaching me. They've been patient with me and lent me their experience,
regardless of how busy they were. David Porter had the unenviable job of
helping convert the original LaTeX source into docbook. It was a long,
boring and dirty job. But someone had to do it. Thanks, David.</para>
<para> Thanks also goes to the fine people at <ulink
url="www.kernelnewbies.org">www.kernelnewbies.org</ulink>. In particular,
Mark McLoughlin and John Levon who I'm sure have much better things to do
than to hang out on kernelnewbies.org and teach the newbies. If this guide
teaches you anything, they are partially to blame.</para>
<para>Both Ori and I would like to thank Richard M. Stallman and Linus
Torvalds for giving us the opportunity to not only run a high-quality
operating system, but to take a close peek at how it works. I've never met
Linus, and probably never will, but he has made a profound difference in my
life.</para>
</sect1>
<sect1><title>Nota Bene</title>
<para>Ori's original document was good about supporting earlier versions of
Linux, going all the way back to the 2.0 days. I had originally intended
to keep with the program, but after thinking about it, opted out. My main
reason to keep with the compatibility was for Linux distributions like
LEAF, which tended to use older kernels. However, even LEAF uses 2.2 and
2.4 kernels these days.<para>
<para>Both Ori and I use the x86 platform. For the most part, the source
code and discussions should apply to other architectures, but I can't
promise anything. One exception is chapter (fixme), on interrupt handlers,
which should not work on any architecture except for x86.</para>
</sect1>

View File

@ -0,0 +1,206 @@
<sect1><title>What Is A Kernel Module?</title>
<para>So, you want to write a kernel module. You know C, you've written a few
normal programs to run as processes, and now you want to get to where the real
action is, to where a single wild pointer can wipe out your file system and a core
dump means a reboot.</para>
<para>What exactly is a kernel module? Modules are pieces of code that can be
loaded and unloaded into the kernel upon demand. They extend the functionality of
the kernel without the need to reboot the system. For example, one type of module
is the device driver, which allows the kernel to access hardware connected to the
system. Without modules, we would have to build monolithic kernels and add new
functionality directly into the kernel image. Besides having larger kernels, this
has the disadvantage of requiring us to rebuild and reboot the kernel every time we
want new functionality.</para>
</sect1>
<sect1><title>How Do Modules Get Into The Kernel?</title>
<indexterm><primary>/proc/modules</primary></indexterm>
<indexterm><primary>kmod</primary></indexterm>
<indexterm><primary>kerneld</primary></indexterm>
<indexterm><primary><filename>/etc/modules.conf</filename></primary></indexterm>
<indexterm><primary><filename>/etc/conf.modules</filename></primary></indexterm>
<para>You can see what modules are already loaded into the kernel by running
<command>lsmod</command>, which gets its information by reading the file
<filename>/proc/modules</filename>.</para>
<para>How do these modules find their way into the kernel? When the kernel needs a
feature that is not resident in the kernel, the kernel module daemon
kmod<footnote><para>In earlier versions of linux, this was known as
kerneld.</para></footnote> execs modprobe to load the module in. modprobe is
passed a string in one of two forms:</para>
<itemizedlist>
<listitem><para>A module name like <filename>softdog</filename> or
<filename>ppp</filename>.</listitem>
<listitem><para>A more generic identifier like
<varname>char-major-10-30</varname>.</listitem>
</itemizedlist>
<para>If modprobe is handed a generic identifier, it first looks for that string in
the file <filename>/etc/modules.conf</filename><footnote><para>This file used to be
called <filename>conf.modules</filename> before linux 2.0, but this name is now
deprecated.</para></footnote>. If it finds an alias line like:</para>
<screen>
alias char-major-10-30 softdog
</screen>
<para>it knows that the generic identifier refers to the module
<filename>softdog.o</filename>.</para>
<para>Next, modprobe looks through the file
<filename>/lib/modules/version/modules.dep</filename>, to see if other modules must
be loaded before the requested module may be loaded. This file is created by
<command>depmod -a</command> and contains module dependencies. For example,
<filename>msdos.o</filename> requires the <filename>fat.o</filename> module to be
already loaded into the kernel. The requested module has a dependancy on another
module if the other module defines symbols (variables or functions) that the
requested module uses.</para>
<para>Lastly, modprobe uses insmod to first load any prerequisite modules into the
kernel, and then the requested module. modprobe directs insmod to <filename
role="directory">/lib/modules/version/</filename><footnote><para>If you are
modifying the kernel, to avoid overwriting your existing modules you may want to
use the <varname>EXTRAVERSION</varname> variable in the kernel Makefile to create a
seperate directory.</para></footnote>, the standard directory for modules. insmod
is intended to be fairly dumb about the location of modules, whereas modprobe is
aware of the default location of modules. So for example, if you wanted to load
the msdos module, you'd have to either run:</para>
<screen>
insmod /lib/modules/2.5.1/kernel/fs/fat/fat.o
insmod /lib/modules/2.5.1/kernel/fs/msdos/msdos.o
</screen>
<para>or just run "<command>modprobe -a msdos</command>".</para>
<indexterm><primary>modules.conf</primary><secondary>keep</secondary></indexterm>
<indexterm><primary>modules.conf</primary><secondary>comment</secondary></indexterm>
<indexterm><primary>modules.conf</primary><secondary>alias</secondary></indexterm>
<indexterm><primary>modules.conf</primary><secondary>options</secondary></indexterm>
<indexterm><primary>modules.conf</primary><secondary>path</secondary></indexterm>
<para>Linux distros provide modprobe, insmod and depmod as a package called
modutils or mod-utils.</para>
<para>Before finishing this chapter, let's take a quick look at a piece of
<filename>/etc/modules.conf</filename>:</para>
<screen>
#This file is automatically generated by update-modules
path[misc]=/lib/modules/2.4.?/local
keep
path[net]=~p/mymodules
options mydriver irq=10
alias eth0 eepro
</screen>
<para>Lines beginning with a '#' are comments. Blank lines are ignored.</para>
<para>The <literal>path[misc]</literal> line tells modprobe to replace the search
path for misc modules with the directory <filename
role="directory">/lib/modules/2.4.?/local</filename>. As you can see, shell meta
characters are honored.</para>
<para>The <literal>path[net]</literal> line tells modprobe to look for net modules
in the directory <filename role="directory">~p/mymodules</filename>, however, the
"keep" directive preceding the <literal>path[net]</literal> directive tells modprobe
to add this directory to the standard search path of net modules as opposed to
replacing the standard search path, as we did for the misc modules.</para>
<para>The alias line says to load in <filename>eepro.o</filename> whenever kmod
requests that the generic identifier `eth0' be loaded.</para>
<para>You won't see lines like "alias block-major-2 floppy" in
<filename>/etc/modules.conf</filename> because modprobe already knows about the
standard drivers which will be used on most systems.</para>
<para>Now you know how modules get into the kernel. There's a bit more to the story
if you want to write your own modules which depend on other modules (we calling this
`stacking modules'). But this will have to wait for a future chapter. We have a
lot to cover before addressing this relatively high-level issue.</para>
<sect2><title>Before We Begin</title>
<para>Before we delve into code, there are a few issues we need to cover.
Everyone's system is different and everyone has their own groove. Getting your
first "hello world" program to compile and load correctly can sometimes be a
trick. Rest assured, after you get over the initial hurdle of doing it for the
first time, it will be smooth sailing thereafter.</para>
<sect3><title>Modversioning</title>
<para>A module compiled for one kernel won't load if you boot a different
kernel unless you enable <literal>CONFIG_MODVERSION</literal> in the kernel.
We won't go into module versioning until later in this guide. Until we
cover modversions, the examples in the guide may not work if you're running
a kernel with modversioning turned on. However, most stock Linux distro
kernels come with it turned on. If you're having trouble loading the
modules because of versioning errors, compile a kernel with modversioning
turned off.</para>
</sect3>
<sect3><title>Using X</title>
<para>It is highly recommended that you type in, compile and load all the
examples this guide discusses. It's also highly recommended you do this
from a console. You should not be working on this stuff in X.</para>
<para>Modules can't print to the screen like <function>printf()</function>
can, but they can log information and warnings, which ends up being printed
on your screen, but only on a console. If you insmod a module from an
xterm, the information and warnings will be logged, but only to your log
files. You won't see it unless you look through your log files. To have
immediate access to this information, do all your work from console.</para>
</sect3>
<sect3><title>Compiling Issues and Kernel Version</title>
<para>Very often, Linux distros will distribute kernel source that has been
patched in various non-standard ways, which may cause trouble.</para>
<para>A more common problem is that some Linux distros distribute incomplete
kernel headers. You'll need to compile your code using various header files
from the Linux kernel. Murphy's Law states that the headers that are
missing are exactly the ones that you'll need for your module work.</para>
<para>To avoid these two problems, I highly recommend that you download,
compile and boot into a fresh, stock Linux kernel which can be downloaded
from any of the Linux kernel mirror sites. See the Linux Kernel HOWTO for
more details.</para>
<para>Ironically, this can also cause a problem. By default, gcc on your
system may look for the kernel headers in their default location rather than
where you installed the new copy of the kernel (usually in <filename
role="directory">/usr/src/</filename>. This can be fixed by using gcc's
<literal>-I</literal> switch.</para>
</sect3>
</sect2>
</sect1>
<!--
vim: tw=86
-->

View File

@ -0,0 +1,564 @@
<sect1><title>Hello, World (part 1): The Simplest Module</title>
<para>When the first caveman programmer chiseled the first program on the walls of
the first cave computer, it was a program to paint the string `Hello, world' in
Antelope pictures. Roman programming textbooks began with the `Salut, Mundi'
program. I don't know what happens to people who break with this tradition, but I
think it's safer not to find out. We'll start with a series of hello world programs
that demonstrate the different aspects of the basics of writing a kernel
module.</para>
<para>Here's the simplest module possible. Don't compile it yet; we'll cover module
compilation in the next section.</para>
<example><title>Hello World (part 1)</title>
<programlisting><![CDATA[
/* hello-1.c - The simplest kernel module.
*/
#include <linux/module.h> /* Needed by all modules */
#include <linux/kernel.h> /* Needed for KERN_ALERT */
int init_module(void)
{
printk("<1>Hello world 1.\n");
/* A non 0 return means init_module failed; module can't be loaded. */
return 0;
}
void cleanup_module(void)
{
printk(KERN_ALERT "Goodbye world 1.\n");
}
]]></programlisting>
</example>
<indexterm><primary><function>init_module()</function></primary></indexterm>
<indexterm><primary><function>cleanup_module()</function></primary></indexterm>
<indexterm><primary><function>printk()</function></primary></indexterm>
<para>A kernel module must have at least two functions: a "start" (initialization)
function called <function>init_module()</function> which is called when the module is
insmoded into the kernel, and an "end" (cleanup) function called
<function>cleanup_module()</function> which is called just before it is
rmmoded.</para>
<para>Typically, <function>init_module()</function> either registers a handler for
something with the kernel, or it replaces one of the kernel functions with its own
code (usually code to do something and then call the original function). The
<function>cleanup_module()</function> function is supposed to undo whatever
<function>init_module()</function> did, so the module can be unloaded safely.</para>
<para>Despite what you might think, <function>printk()</function> was not meant to
communicate information to the user, even though we use it for exactly this purpose
within this document! It happens to be a logging mechanism for the kernel, and is
used to log information or give warnings. Therefore, each
<function>printk()</function> statement comes with a priority, which is the
<varname>&lt;1&gt;</varname> you see. There are 8 priorities and the kernel has
macros for them, so you don't have to use cryptic numbers. We could've used a macro
instead of the explicit priority level: <function>printk(KERN_ALERT "Hello,
world.");</function> There are 8 priority levels and you can view them (and what
they mean) in the file <filename role="headerfile">linux/kernel.h</filename>. If you
don't specify a priority level, the default priority,
<varname>DEFAULT_MESSAGE_LOGLEVEL</varname>, will be used.</para>
<para>Take time to read through the priority macros. The header file also describes
what each priority means. In practise, don't use number, like
<literal>&lt;4&gt;</literal>. Always use the macro, like
<literal>KERN_WARNING</literal>.</para>
<para>If the priority is less than <varname>int console_loglevel</varname>, the
message is printed on your current terminal. If both <command>syslogd</command> and
<application>klogd</application> are running, then the message will also get appended
to <filename>/var/log/messages</filename>, whether it got printed to the console or
not. We use a high priority, like <literal>KERN_ALERT</literal>, to make sure the
<function>printk()</function> messages get printed to your console rather than just
logged to your logfile. When you write real modules, you'll want to use priorities
that are meaningful for the situation at hand.</para>
<para>There's more I want to show you using "hello world" type programs, but before
we move on, you need to learn how to compile them.</para>
</sect1>
<sect1><title>Compiling Kernel Modules</title>
<indexterm><primary>insmod</primary></indexterm>
<para>A kernel module is not an independant executable, but an object file which will
be linked into the kernel during runtime using insmod. As a result, modules should
be compiled with the <option>-c</option> flag. Also, because the kernel makes
extensive use of inline functions, modules must be compiled with the optimization
flag, <option>-O</option>, although heavy optimization like <option>-O2</option> is
not recommended. Without optimization, some of the assembler macros calls will be
mistaken by the compiler for function calls. This will cause loading the module to
fail, since insmod won't find those functions in the kernel.</para>
<para>Kernel modules also need to be compiled with certain symbols defined. This is
because the kernel header files need to behave differently, depending on whether
we're compiling a kernel module or an executable. You define symbols using gcc's
<option>-D</option> option, and here are a list of symbols that should be defined for
every module you compile:</para>
<indexterm><primary>MODULE</primary></indexterm>
<indexterm><primary>__KERNEL__</primary></indexterm>
<indexterm><primary>__SMP__</primary></indexterm>
<itemizedlist>
<listitem><para><varname>__KERNEL__</varname>: Tells the header files that the code
will be run in kernel mode, not as a user process.</para></listitem>
<listitem><para><varname>MODULE</varname>: Tells the header files to give the
appropriate definitions for a kernel module.</para></listitem>
<listitem><para><varname>__SMP__</varname>: This must be defined if the kernel was
compiled to support symmetrical multiprocessing, even if it's running just on one
CPU. Note how I did this in <filename>hello-1.c</filename>.</para></listitem>
</itemizedlist>
<para> So, let's look at a simple Makefile for compiling a module:</para>
<example><title>Makefile for a basic kernel module</title>
<screen><![CDATA[
# Makefile for a basic kernel module
CC=gcc
CFLAGS := -c -0 -W -Wall -Wstrict-prototypes -Wmissing-prototypes
MODFLAGS := -DMODULE -D__KERNEL__
hello-1.o: hello.c
${CC} ${MODCFLAGS} hello.c
]]></screen>
</example>
<para>Type <filename>hello-1.c</filename> in and compile it. Insert it into the
kernel with <command>insmod ./hello-1.o</command>. Neat, eh? All modules loaded
into the kernel are listed in <filename>/proc/modules</filename>. Go ahead and cat
that file to see that your module is really a part of the kernel. Congratulations,
you are now the author of Linux kernel code! When the novelty wares off, you can
remove your module from the kernel by using <command>rmmod hello-1</command>. Take
a look at <filename>/var/log/messages</filename> just to see that it got logged to
your system logfile.</para>
</sect1>
<sect1><title>Hello World (part 2): The <function>module_init()</function> and
<function>module_exit()</function> Macros</title>
<indexterm><primary>module_init</primary></indexterm>
<indexterm><primary>module_exit</primary></indexterm>
<para>As of Linux 2.4, you can rename the init and cleanup functions of your
modules; they no longer have to be called <function>init_module()</function> and
<function>cleanup_module()</function> respectively. This is done with the
<function>module_init()</function> and <function>module_exit()</function> macros.
These macros are defined in <filename role="header">linux/init.h</filename>. The
only caveat is that your init and cleanup functions must be defined before calling
the macros, otherwise you'll get compilation errors. Here's an example of this
technique:</para>
<example><title>Hello World (part 2)</title>
<programlisting><![CDATA[
/* hello-2.c - Demonstrating the module_init() and module_exit() macros.
*/
#include <linux/module.h> /* Needed by all modules */
#include <linux/kernel.h> /* Needed for KERN_ALERT */
#include <linux/init.h> /* Needed for the macros */
int my_wonderful_init(void)
{
printk(KERN_ALERT "Hello, world 2\n");
return 0;
}
void my_wonderful_cleanup(void)
{
printk(KERN_ALERT "Goodbye, world 2\n");
}
module_init(my_wonderful_init);
module_exit(my_wonderful_cleanup);
]]></programlisting>
</example>
</sect1>
<sect1><title>Hello World (part 3): The <literal>__init</literal> and
<literal>__exit</literal> Macros</title>
<indexterm><primary><function>__init</function></primary></indexterm>
<indexterm><primary><function>__initdata</function></primary></indexterm>
<indexterm><primary><function>__exit</function></primary></indexterm>
<indexterm><primary><function>__initfunction()</function></primary></indexterm>
<para>This demonstrates a feature of kernel 2.2 and later. Notice the change in the
definitions of the init and cleanup functions. The <function>__init</function> macro
will cause the init function to be discarded and its memory reclaimed for the kernel
once the init function finishes, but only for built-in drivers. It has no effect for
loadable modules.</para>
<para>There is also an <function>__initdata</function> which works similarly to
<function>__init</function> but for init variables rather than functions.</para>
<para>The <function>__exit</function> macro causes the omission of the function when
the module is built into the kernel. This only has an effect for built in modules
since they never exit (and hence don't need a cleanup function). It has no effect on
loadable modules since they need their cleanup function.</para>
<para>These macros are defined in <filename role="headerfile">linux/init.h</filename>
and serve to free up kernel memory. When you boot your kernel and see something like
<literal>Freeing unused kernel memory: 236k freed</literal>, this is precisely what
the kernel is freeing.</para>
<example><title>Hello World (part 3)</title>
<programlisting><![CDATA[
/* hello-3.c - Illustrating the __init, __initdata and __exit macros.
*/
#include <linux/module.h> /* Needed by all modules */
#include <linux/kernel.h> /* Needed for KERN_ALERT */
#include <linux/init.h> /* Needed for the macros */
static int hello3_data __initdata = 3;
static int __init hello3_init_function(void)
{
printk(KERN_ALERT "Hello, world %d\n", hello3_data);
return 0;
}
static void __exit hello3_cleanup_function(void)
{
printk(KERN_ALERT "Goodbye, world 3\n");
}
module_init(hello3_init_function);
module_exit(hello3_cleanup_function);
]]></programlisting>
</example>
<para>You may see a directive named "<function>__initfunction()</function>" in
drivers written for Linux 2.2 kernels:</para>
<screen><![CDATA[
__initfunction(int init_module(void))
{
printk(KERN_ALERT "Hi there.\n");
return 0;
}
]]></screen>
<para>This macro served the same purpose as <function>__init</function>, but is now
deprecated in favor of <function>__init</function>. Don't use
<function>__initfunction()</function> in your own code.</para>
</sect1>
<sect1><title>Hello World (part 4): Licensing and Module Documentation</title>
<indexterm><primary><literal>MODULE_LICENSE()</literal></primary></indexterm>
<indexterm><primary><literal>MODULE_DESCRIPTION()</literal></primary></indexterm>
<indexterm><primary><literal>MODULE_AUTHOR()</literal></primary></indexterm>
<indexterm><primary><literal>MODULE_SUPPORTED_DEVICE()</literal></primary></indexterm>
<para>If you're running kernel 2.4 or later, you might have noticed the message like:
"<literal>Warning: loading hello-1.o will taint the kernel: no license</literal>"
when you loaded the previous example modules. In 2.4 and later, a mechanism was
devised to identify code licensed under the GPL (and friends) so people can be warned
that the code is non open-source. This is accomplished by the
<literal>MODULE_LICENSE()</literal> macro which is demonstrated in the next piece of
code. By setting the license to GPL, you can keep the warning from being printed.
This mechanism is documented in <filename
role="headerfile">linux/module.h</filename>, and I recommend you read the comments
about this macro in the header file.</para>
<para>A similar mechanism is used to identify the module description
"<literal>MODULE_DESCRIPTION()</literal>", author
"<literal>MODULE_AUTHOR()</literal>" and what device the module supports
"<literal>MODULE_SUPPORTED_DEVICE()</literal>" and is defined in <filename
role="headerfile">linux/module.h</filename>. This info isn't really used by the
kernel itself; it's used as documentation and can be viewed by a tool like
objdump.</para>
<para>We haven't covered devices yet, so the last macro may be a mystery to you, but
keep it in mind. We'll cover char devices shortly.</para>
<example><title>Hello World (part 4)</title>
<programlisting><![CDATA[
/* hello-4.c - Demonstrates tainting messages and documentation.
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
#define DRIVER_AUTHOR "Peter Jay Salzman <p@dirac.org>"
#define DRIVER_DESC "A sample driver"
static int __init hello4_init_function(void)
{
printk(KERN_ALERT "Hello, world 4\n");
return 0;
}
static void __exit hello4_cleanup_function(void)
{
printk(KERN_ALERT "Goodbye, world 4\n");
}
module_init(hello4_init_function);
module_exit(hello4_cleanup_function);
/* You can use strings here or a define, as shown. It doesn't matter what you
* actually name the #define's, so "AUTHOR" is as good as "DRIVER_AUTHOR". */
MODULE_AUTHOR(DRIVER_AUTHOR);
MODULE_DESCRIPTION(DRIVER_DESC);
/* This gets rid of the "taint message" by declaring this code as GPL. */
MODULE_LICENSE("GPL");
/* This says that the module uses /dev/testdevice. It might be used in the
* future to help automatic configuration of modules, but is currently unused
* other than documentation purposes. */
MODULE_SUPPORTED_DEVICE("testdevice");
]]></programlisting>
</example>
</sect1>
<sect1><title>Passing Command Line Arguments to a Module</title>
<para>Modules can take command line arguments, but not with the argc/argv you might
be used to.</para>
<para>To allow arguments to be passed to your driver, declare the variables that will
take the values of the command line arguments as global and then use the MODULE_PARM
macro (defined in <filename role="headerfile">linux/module.h</filename>) to set the
mechanism up. At runtime, insmod will fill the variables with any command line
arguments that are given. The variable declarations and macros should be placed at
the beginning of the module for clarity. The example code should clear up my
admittedly lousy explanation.</para>
<para>The <literal>MODULE_PARM</literal> macro takes 2 arguments: the name of the
variable and its type. The supported variable types are "<literal>b</literal>":
single byte, "<literal>h</literal>": short int, "<literal>i</literal>": integer,
"<literal>l</literal>": long int and "<literal>s</literal>": string. Strings should
be declared as "<type>char *</type>" and insmod will allocate memory for them. You
should always try to give the variables an initial default value. This is kernel
code, and you should program defensively. For example:</para>
<screen>
int myint = 3;
char *mystr;
MODULE_PARM (myint, "i");
MODULE_PARM (mystr, "s");
</screen>
<para>Arrays are supported too. An integer value preceding the type in MODULE_PARM
will indicate an array of some maximum length. Two numbers separated by a '-' will
give the minimum and maximum number of values. For example, an array of shorts with
at least 2 and no more than 4 values could be declared as:</para>
<screen>
int myshortArray[4];
MODULE_PARM (myintArray, "2-4i");
</screen>
<para>A good use for this is to have the module variable's default values set, like
which IO port or IO memory to use. If the variables contain the default values, then
perform autodetection (explained elsewhere). Otherwise, keep the current value.
This will be made clear later on. For now, I just want to demonstrate passing
arguments to a module.</para>
<example><title>Hello World (part 5)</title>
<programlisting><![CDATA[
/* hello-5.c - Demonstrates command line argument passing to a module.
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
static int myint = 0;
static char *mystring = "blah";
MODULE_PARM (myint, "i");
MODULE_PARM (mystring, "s");
static int __init hello5_init_function(void)
{
printk(KERN_ALERT "Hello, world 5\n");
printk(KERN_ALERT "integer: %i\n", myint);
printk(KERN_ALERT "string: %s\n", mystring);
return 0;
}
static void __exit hello5_cleanup_function(void)
{
printk(KERN_ALERT "Goodbye, world 5\n");
}
module_init(hello5_init_function);
module_exit(hello5_cleanup_function);
]]></programlisting>
</example>
</sect1>
<sect1><title>Modules Spanning Multiple Files</title>
<indexterm><primary>source files</primary><secondary>multiple</secondary></indexterm>
<indexterm><primary>__NO_VERSION__</primary></indexterm>
<indexterm><primary>module.h</primary></indexterm>
<indexterm><primary>version.h</primary></indexterm>
<indexterm><primary>kernel\_version</primary></indexterm>
<indexterm><primary>ld</primary></indexterm>
<indexterm><primary>elf_i386</primary></indexterm>
<para>Sometimes it makes sense to divide a kernel module between several
source files. In this case, you need to:</para>
<orderedlist>
<listitem><para>In all the source files but one, add the line
<command>#define __NO_VERSION__</command>. This is important because
<filename role="headerfile">module.h</filename> normally includes the
definition of <varname>kernel_version</varname>, a global variable with
the kernel version the module is compiled for. If you need <filename
role="headerfile">version.h</filename>, you need to include it yourself,
because <filename role="headerfile">module.h</filename> won't do it for
you with <varname>__NO_VERSION__</varname>.</para></listitem>
<listitem><para>Compile all the source files as usual.</para></listitem>
<listitem><para>Combine all the object files into a single one. Under x86,
use <command>ld -m elf_i386 -r -o &lt;module name.o&gt; &lt;1st src
file.o&gt; &lt;2nd src file.o&gt;</command>.</para></listitem>
</orderedlist>
<para>Here's an example of such a kernel module.</para>
<example><title>start.c</title>
<programlisting><![CDATA[
/* start.c - Illustration of multi filed modules
*/
#include <linux/kernel.h> /* We're doing kernel work */
#include <linux/module.h> /* Specifically, a module */
int init_module(void)
{
printk("Hello, world - this is the kernel speaking\n");
return 0;
}
]]></programlisting>
</example>
<para>The next file:</para>
<example><title>stop.c</title>
<programlisting><![CDATA[
/* stop.c - Illustration of multi filed modules
*/
#if defined(CONFIG_MODVERSIONS) && ! defined(MODVERSIONS)
#include <linux/modversions.h> /* Will be explained later */
#define MODVERSIONS
#endif
#include <linux/kernel.h> /* We're doing kernel work */
#include <linux/module.h> /* Specifically, a module */
#define __NO_VERSION__ /* It's not THE file of the kernel module */
#include <linux/version.h> /* Not included by module.h because of
__NO_VERSION__ */
void cleanup_module()
{
printk("<1>Short is the life of a kernel module\n");
}
]]></programlisting>
</example>
<para>And finally, the makefile:</para>
<example><title>Makefile for a multi-filed module</title>
<screen><![CDATA[
CC=gcc
MODCFLAGS := -O -Wall -DMODULE -D__KERNEL__
hello.o: hello2_start.o hello2_stop.o
ld -m elf_i386 -r -o hello2.o hello2_start.o hello2_stop.o
start.o: hello2_start.c
${CC} ${MODCFLAGS} -c hello2_start.c
stop.o: hello2_stop.c
${CC} ${MODCFLAGS} -c hello2_stop.c
]]></screen>
</example>
</sect1>
<!--
vim: tw=87
-->

View File

@ -0,0 +1,317 @@
<sect1><title>Modules vs Programs</title>
<sect2><title>How modules begin and end</title>
<para>A program usually begins with a <function>main()</function> function, executes a
bunch of instructions and terminates upon completion of those instructions. Kernel
modules work a bit differently. A module always begin with either the
<function>init_module</function> or the function you specify with
<function>module_init</function> call. This is the entry function for modules; it tells
the kernel what functionality the module provides and sets up the kernel to run the
module's functions when they're needed. Once it does this, entry function returns and the
module does nothing until the kernel wants to do something with the code that the module
provides.</para>
<para>All modules end by calling either <function>cleanup_module</function> or the
function you specify with the <function>module_exit</function> call. This is the exit
function for modules; it undoes whatever entry function did. It unregisters the
functionality that the entry function registered.</para>
<para>Every module must have an entry function and an exit function. Since there's more
than one way to specify entry and exit functions, I'll try my best to use the terms `entry
function' and `exit function', but if I slip and simply refer to them as
<function>init_module</function> and <function>cleanup_module</function>, I think you'll
know what I mean.</para>
</sect2>
<sect2><title>Functions available to modules</title>
<indexterm><primary>library function</primary></indexterm>
<indexterm><primary>system call</primary></indexterm>
<indexterm><primary><filename>/proc/ksyms</filename></primary></indexterm>
<para>Programmers use functions they don't define all the time. A prime example of this
is <function>printf()</function>. You use these library functions which are provided by
the standard C library, libc. The definitions for these functions don't actually enter
your program until the linking stage, which insures that the code (for
<function>printf()</function> for example) is available, and fixes the call instruction
to point to that code.</para>
<para>Kernel modules are different here, too. In the hello world example, you might
have noticed that we used a function, <function>printk()</function> but didn't include a
standard I/O library. That's because modules are object files whose symbols get
resolved upon insmod'ing. The definition for the symbols comes from the kernel itself;
the only external functions you can use are the ones provided by the kernel. If you're
curious about what symbols have been exported by your kernel, take a look at
<filename>/proc/ksyms</filename>.</para>
<para>One point to keep in mind is the difference between library functions and system
calls. Library functions are higher level, run completely in user space and provide a
more convenient interface for the programmer to the functions that do the real
work---system calls. System calls run in kernel mode on the user's behalf and are
provided by the kernel itself. The library function <function>printf()</function> may
look like a very general printing function, but all it really does is format the data
into strings and write the string data using the low-level system call
<function>write()</function>, which then sends the data to standard output.</para>
<para> Would you like to see what system calls are made by
<function>printf()</function>? It's easy! Compile the following program: </para>
<screen>
#include &lt;stdio.h&gt;
int main(void)
{ printf("hello"); return 0; }
</screen>
<indexterm><primary>strace</primary></indexterm>
<para>with <command>gcc -Wall -o hello hello.c</command>. Run the exectable with
<command>strace hello</command>. Are you impressed? Every line you see corresponds to
a system call. strace<footnote><para>It's an invaluable tool for figuring out things
like what files a program is trying to access. Ever have a program bail silently
because it couldn't find a file? It's a PITA!</para></footnote> is a handy program that
gives you details about what system calls a program is making, including which call is
made, what its arguments are what it returns. It's an invaluable tool for figuring out
things like what files a program is trying to access. Towards the end, you'll see a
line which looks like <function>write(1, "hello", 5hello)</function>. There it is. The
face behind the <function>printf()</function> mask. You may not be familiar with write,
since most people use library functions for file I/O (like fopen, fputs, fclose). If
that's the case, try looking at <command>man 2 write</command>. The 2nd man section is
devoted to system calls (like <function>kill()</function> and
<function>read()</function>. The 3rd man section is devoted to library calls, which you
would probably be more familiar with (like <function>cosh()</function> and
<function>random()</function>).</para>
<para>You can even write modules to replace the kernel's system calls, which we'll do
shortly. Crackers often make use of this sort of thing for backdoors or trojans, but
you can write your own modules to do more benign things, like have the kernel write
<emphasis>Tee hee, that tickles!</emphasis> everytime someone tries to delete a file on
your system.</para>
</sect2>
<sect2><title>User Space vs Kernel Space</title>
<para>A kernel is all about access to resources, whether the resource in question happens
to be a video card, a hard drive or even memory. Programs often compete for the same
resource. As I just saved this document, updatedb started updating the locate database.
My vim session and updatedb are both using the hard drive concurrently. The kernel needs
to keep things orderly, and not give users access to resources whenever they feel like it.
To this end, a <acronym>CPU</acronym> can run in different modes. Each mode gives a
different level of freedom to do what you want on the system. The Intel 80386
architecture has 4 of these modes, which are called rings. Unix uses only two rings; the
highest ring (ring 0, also known as `supervisor mode' where everything is allowed to
happen) and the lowest ring, which is called `user mode'.</para>
<para>Recall the discussion about library functions vs system calls. Typically, you use a
library function in user mode. The library function calls one or more system calls, and
these system calls execute on the library function's behalf, but do so in supervisor mode
since they are part of the kernel itself. Once the system call completes its task, it
returns and execution gets transfered back to user mode.</para>
</sect2>
<sect2><title>Name Space</title>
<indexterm><primary>symbol table</primary></indexterm>
<indexterm><primary>namespace pollution</primary></indexterm>
<indexterm><primary><filename>/proc/ksyms</filename></primary></indexterm>
<para>When you write a small C program, you use variables which are convenient and make
sense to the reader. If, on the other hand, you're writing routines which will be part
of a bigger problem, any global variables you have are part of a community of other
peoples' global variables; some of the variable names can clash. When a program has
lots of global variables which aren't meaningful enough to be distinguished, you get
<emphasis>namespace pollution</emphasis>. In large projects, effort must be made to
remember reserved names, and to find ways to develop a scheme for naming unique variable
names and symbols.</para>
<para>When writing kernel code, even the smallest module will be linked against the
entire kernel, so this is definitely an issue. The best way to deal with this is to
declare all your variables as <type>static</type> and to use a well-defined prefix for
your symbols. By convention, all kernel prefixes are lowercase. If you don't want to
declare everything as <type>static</type>, another option is to declare a
<varname>symbol table</varname> and register it with a kernel. We'll get to this
later.</para>
<para>The file <filename>/proc/ksyms</filename> holds all the symbols that the kernel
knows about and which are therefore accessible to your modules since they share the
kernel's codespace.</para>
</sect2>
<sect2><title>Code space</title>
<indexterm><primary>code space</primary></indexterm>
<indexterm><primary>monolithic kernel</primary></indexterm>
<indexterm><primary>Hurd</primary></indexterm>
<indexterm><primary>Neutrino</primary></indexterm>
<indexterm><primary>microkernel</primary></indexterm>
<para>Memory management is a very complicated subject---the majority of O'Reilly's
`Understanding The Linux Kernel' is just on memory management! We're not setting out to
be experts in memory managements, but we do need to know a couple of facts to even begin
worrying about writing real modules.</para>
<para>If you haven't thought about what a segfault really means, you may be surprised to
hear that pointers don't actually point to memory locations. Not real ones, anyway.
When a process is created, the kernel sets aside a portion of real physical memory and
hands it to the process to use for its executing code, variables, stack, heap and other
things which a computer scientist would know about<footnote><para>I'm a physicist, not a
computer scientist, Jim!</para></footnote>. This memory begins with $0$ and extends up
to whatever it needs to be. Since the memory space for any two processes don't overlap,
every process that can access a memory address, say <literal>0xbffff978</literal>, would
be accessing a different location in real physical memory! The processes would be
accessing an index named <literal>0xbffff978</literal> which points to some kind of
offset into the region of memory set aside for that particular process. For the most
part, a process like our Hello, World program can't access the space of another process,
although there are ways which we'll talk about later.</para>
<para>The kernel has its own space of memory as well. Since a module is code which can
be dynamically inserted and removed in the kernel (as opposed to a semi-autonomous
object), it shares the kernel's codespace rather than having its own. Therefore, if
your module segfaults, the kernel segfaults. And if you start writing over data because
of an off-by-one error, then you're trampling on kernel code. This is even worse than
it sounds, so try your best to be careful.</para>
<para>By the way, I would like to point out that the above discussion is true for any
operating system which uses a monolithic kernel<footnote><para>This isn't quite the same
thing as `building all your modules into the kernel', although the idea is the
same.</para></footnote>. There are things called microkernels which have modules which
get their own codespace. The GNU Hurd and QNX Neutrino are two examples of a
microkernel.</para>
</sect2>
<sect2><title>Device Drivers</title>
<para>One class of module is the device driver, which provides functionality for
hardware like a TV card or a serial port. On unix, each piece of hardware is
represented by a file located in <filename role=directory>/dev</filename> named a
<filename>device file</filename> which provides the means to communicate with the
hardware. The device driver provides the communication on behalf of a user program. So
the <filename>es1370.o</filename> sound card device driver might connect the <filename
role="devicefile">/dev/sound</filename> device file to the Ensoniq IS1370 sound card. A
userspace program like mp3blaster can use <filename
role="devicefile">/dev/sound</filename> without ever knowing what kind of sound card is
installed.</para>
<sect3><title>Major and Minor Numbers</title>
<indexterm><primary>major number</primary></indexterm>
<indexterm><primary>minor number</primary></indexterm>
<para>Let's look at some device files. Here are device files which represent the
first three partitions on the primary master IDE hard drive:</para>
<screen>
# ls -l /dev/hda[1-3]
brw-rw---- 1 root disk 3, 1 Jul 5 2000 /dev/hda1
brw-rw---- 1 root disk 3, 2 Jul 5 2000 /dev/hda2
brw-rw---- 1 root disk 3, 3 Jul 5 2000 /dev/hda3
</screen>
<para>Notice the column of numbers separated by a comma? The first number is
called the device's major number. The second number is the minor number. The
major number tells you which driver is used to access the hardware. Each driver
is assigned a unique major number; all device files with the same major number are
controlled by the same driver. All the above major numbers are 3, because they're
all controlled by the same driver.</para>
<para>The minor number is used by the driver to distinguish between the various
hardware it controls. Returning to the example above, although all three devices
are handled by the same driver they have unique minor numbers because the driver
sees them as being different pieces of hardware.</para>
<para> Devices are divided into two types: character devices and block devices.
The difference is that block devices have a buffer for requests, so they can
choose the best order in which to respond to the requests. This is important in
the case of storage devices, where it's faster to read or write sectors which are
close to each other, rather than those which are further apart. Another
difference is that block devices can only accept input and return output in blocks
(whose size can vary according to the device), whereas character devices are
allowed to use as many or as few bytes as they like. Most devices in the world
are character, because they don't need this type of buffering, and they don't
operate with a fixed block size. You can tell whether a device file is for a
block device or a character device by looking at the first character in the output
of <command>ls -l</command>. If it's `b' then it's a block device, and if it's `c'
then it's a character device. The devices you see above are block devices. Here
are some character devices (the serial ports):</para>
<screen>
crw-rw---- 1 root dial 4, 64 Feb 18 23:34 /dev/ttyS0
crw-r----- 1 root dial 4, 65 Nov 17 10:26 /dev/ttyS1
crw-rw---- 1 root dial 4, 66 Jul 5 2000 /dev/ttyS2
crw-rw---- 1 root dial 4, 67 Jul 5 2000 /dev/ttyS3
</screen>
<para> If you want to see which major numbers have been assigned, you can look at
<filename>/usr/src/linux/Documentation/devices.txt</filename>. </para>
<indexterm><primary>mknod</primary></indexterm>
<indexterm><primary>coffee</primary></indexterm>
<para>When the system was installed, all of those device files were created by the
<command>mknod</command> command. To create a new char device named `coffee' with
major/minor number <literal>12</literal> and <literal>2</literal>, simply do
<command>mknod /dev/coffee c 12 2</command>. You don't <emphasis>have</emphasis>
to put your device files into <filename role="directory">/dev</filename>, but it's
done by convention. Linus put his device files in <filename> /dev</filename>, and
so should you. However, when creating a device file for testing purposes, it's
probably OK to place it in your working directory where you compile the kernel
module. Just be sure to put it in the right place when you're done writing the
device driver.</para>
<para>I would like to make a few last points which are implicit from the above
discussion, but I'd like to make them explicit just in case. When a device file
is accessed, the kernel uses the major number of the file to determine which
driver should be used to handle the access. This means that the kernel doesn't
really need to use or even know about the minor number. The driver itself is the
only thing that cares about the minor number. It uses the minor number to
distinguish between different pieces of hardware.</para>
<para>By the way, when I say `hardware', I mean something a bit more abstract than
a PCI card that you can hold in your hand. Look at these two device
files:</para>
<screen>
% ls -l /dev/fd0 /dev/fd0u1680
brwxrwxrwx 1 root floppy 2, 0 Jul 5 2000 /dev/fd0
brw-rw---- 1 root floppy 2, 44 Jul 5 2000 /dev/fd0u1680
</screen>
<para>By now you can look at these two device files and know instantly that they
are block devices and are handled by same driver (block major
<literal>2</literal>). You might even be aware that these both represent your
floppy drive, even if you only have one floppy drive. Why two files? One
represents the floppy drive with <literal>1.44</literal> <acronym>MB</acronym> of
storage. The other is the <emphasis>same</emphasis> floppy drive with
<literal>1.68</literal> <acronym>MB</acronym> of storage, and corresponds to what
some people call a `superformatted' disk. One that holds more data than a
standard formatted floppy. So here's a case where two device files with different
minor number actually represent the same piece of physical hardware. So just be
aware that the word `hardware' in our discussion can mean something very
abstract.</para>
</sect3>
</sect2>
</sect1>
<!--
vim:textwidth=96
-->

View File

@ -0,0 +1,456 @@
<sect1><title>Character Device Drivers</title>
<indexterm><primary>device file</primary><secondary>character</secondary>
</indexterm>
<sect2><title>The <type>file_operations</type> Structure</title>
<indexterm><primary>file_operations</primary></indexterm>
<para>The <type>file_operations</type> structure is defined in <filename
role="headerfile">linux/fs.h</filename>, and holds pointers to functions
defined by the driver that perform various operations on the device. Each
field of the structure corresponds to the address of some function defined
by the driver to handle a requested operation.</para>
<para> For example, every character driver needs to define a function that
reads from the device. The <type>file_operations</type> structure holds
the address of the module's function that performs that operation. Here is
what the definition looks like for kernel <literal>2.4.2</literal>:</para>
<screen>
struct file_operations {
struct module *owner;
loff_t (*llseek) (struct file *, loff_t, int);
ssize_t (*read) (struct file *, char *, size_t, loff_t *);
ssize_t (*write) (struct file *, const char *, size_t, loff_t *);
int (*readdir) (struct file *, void *, filldir_t);
unsigned int (*poll) (struct file *, struct poll_table_struct *);
int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
int (*mmap) (struct file *, struct vm_area_struct *);
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *);
int (*release) (struct inode *, struct file *);
int (*fsync) (struct file *, struct dentry *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
ssize_t (*readv) (struct file *, const struct iovec *, unsigned long,
loff_t *);
ssize_t (*writev) (struct file *, const struct iovec *, unsigned long,
loff_t *);
};
</screen>
<para>Some operations are not implemented by a driver. For example, a
driver that handles a video card won't need to read from a directory
structure. The corresponding entries in the <type>file_operations</type>
structure should be set to <varname>NULL</varname>.</para>
<para>There is a gcc extension that makes assigning to this structure more
convenient. You'll see it in modern drivers, and may catch you by
surprise. This is what the new way of assigning to the structure looks
like:</para>
<screen>
struct file_operations fops = {
read: device_read,
write: device_write,
open: device_open,
release: device_release
};
</screen>
<para>However, there's also a C99 way of assigning to elements of a
structure. The version of gcc I'm currently using,
<literal>2.95</literal>, supports the new C99 syntax. You should use this
syntax in case someone wants to port your driver. It will help with
compatibility:</para>
<screen>
struct file_operations fops = {
.read = device_read,
.write = device_write,
.open = device_open,
.release = device_release
};
</screen>
<para>The meaning is clear, and you should be aware that any member of the
structure which you don't explicitly assign will be initialized to
<varname>NULL</varname> by gcc.</para>
<para>A pointer to a <type>struct file_operations</type> is commonly named
<varname>fops</varname>.</para>
</sect2>
<sect2><title>The <type>file</type> structure</title>
<indexterm><primary>file</primary></indexterm>
<indexterm><primary>inode</primary></indexterm>
<para>Each device is represented in the kernel by a <type>file</type>
structure, which is defined in <filename
role="header">linux/fs.h</filename>. Be aware that a <type>file</type> is
a kernel level structure and never appears in a user space program. It's
not the same thing as a <type>FILE</type>, which is defined by glibc and
would never appear in a kernel space function. Also, its name is a bit
misleading; it represents an abstract open `file', not a file on a disk,
which is represented by a structure named <type>inode</type>.</para>
<para>A pointer to a <varname>struct file</varname> is commonly named
<function>filp</function>. You'll also see it refered to as
<varname>struct file file</varname>. Resist the temptation.</para>
<para>Go ahead and look at the definition of <function>file</function>.
Most of the entries you see, like <function>struct dentry</function> aren't
used by device drivers, and you can ignore them. This is because drivers
don't fill <varname>file</varname> directly; they only use structures
contained in <varname>file</varname> which are created elsewhere.</para>
</sect2>
<sect2><title>Registering A Device</title>
<indexterm><primary>register_chrdev</primary></indexterm>
<indexterm><primary>major number</primary>
<secondary>dynamic allocation</secondary></indexterm>
<para>As discussed earlier, char devices are accessed through device files,
usually located in <filename
role="direcotry">/dev</filename><footnote><para>This is by convention.
When writing a driver, it's OK to put the device file in your current
directory. Just make sure you place it in <filename
role="directory">/dev</filename> for a production driver</para></footnote>.
The major number tells you which driver handles which device file. The
minor number is used only by the driver itself to differentiate which
device it's operating on, just in case the driver handles more than one
device.</para>
<para> Adding a driver to your system means registering it with the kernel.
This is synonymous with assigning it a major number during the module's
initialization. You do this by using the
<function>register_chrdev</function> function, defined by <filename
role="headerfile">linux/fs.h</filename>.</para>
<screen>
int register_chrdev(unsigned int major, const char *name,
struct file_operations *fops);
</screen>
<para>where <varname>unsigned int major</varname> is the major number you
want to request, <varname>const char *name</varname> is the name of the
device as it'll appear in <filename>/proc/devices</filename> and
<varname>struct file_operations *fops</varname> is a pointer to the
<varname>file_operations</varname> table for your driver. A negative
return value means the registertration failed. Note that we didn't pass
the minor number to <function>register_chrdev</function>. That's because
the kernel doesn't care about the minor number; only our driver uses it.
</para>
<para>Now the question is, how do you get a major number without hijacking
one that's already in use? The easiest way would be to look through
<filename>Documentation/devices.txt</filename> and pick an unused one.
That's a bad way of doing things because you'll never be sure if the number
you picked will be assigned later. The answer is that you can ask the
kernel to assign you a dynamic major number.</para>
<para> If you pass a major number of 0 to
<function>register_chrdev</function>, the return value will be the
dynamically allocated major number. The downside is that you can't make a
device file in advance, since you don't know what the major number will be.
There are a couple of ways to do this. First, the driver itself can print
the newly assigned number and we can make the device file by hand. Second,
the newly registered device will have an entry in
<filename>/proc/devices</filename>, and we can either make the device file
by hand or write a shell script to read the file in and make the device
file. The third method is we can have our driver make the the device file
using the <function>mknod</function> system call after a successful
registration and rm during the call to
<function>cleanup_module</function>.</para>
</sect2>
<sect2><title>Unregistering A Device</title>
<indexterm><primary>rmmod</primary><secondary>preventing</secondary>
</indexterm>
<para>We can't allow the kernel module to be
<application>rmmod</application>'ed whenever root feels like it. If the
device file is opened by a process and then we remove the kernel module,
using the file would cause a call to the memory location where the
appropriate function (read/write) used to be. If we're lucky, no other
code was loaded there, and we'll get an ugly error message. If we're
unlucky, another kernel module was loaded into the same location, which
means a jump into the middle of another function within the kernel. The
results of this would be impossible to predict, but they can't be very
positive.</para>
<para> Normally, when you don't want to allow something, you return an
error code (a negative number) from the function which is supposed to do
it. With <function>cleanup_module</function> that's impossible because
it's a void function. However, there's a counter which keeps track of how
many processes are using your module. You can see what it's value is by
looking at the 3rd field of <filename>/proc/modules</filename>. If this
number isn't zero, <function>rmmod</function> will fail. Note that you
don't have to check the counter from within
<function>cleanup_module</function> because the check will be performed for
you by the system call <function>sys_delete_module</function>, defined in
<filename>linux/module.c</filename>. You shouldn't use this counter
directly, but there are macros defined in <filename
role="headerfile">linux/modules.h</filename> which let you increase,
decrease and display this counter:</para>
<itemizedlist>
<listitem><para><varname>MOD_INC_USE_COUNT</varname>: Increment the use
count.</para></listitem>
<listitem><para><varname>MOD_DEC_USE_COUNT</varname>: Decrement the use count.
</para></listitem>
<listitem><para><varname>MOD_IN_USE</varname>: Display the use count.
</para></listitem>
</itemizedlist>
<para>It's important to keep the counter accurate; if you ever do lose
track of the correct usage count, you'll never be able to unload the
module; it's now reboot time, boys and girls. This is bound to happen to
you sooner or later during a module's development.</para>
<indexterm><primary>MOD_INC_USE_COUNT</primary></indexterm>
<indexterm><primary>MOD_DEC_USE_COUNT</primary></indexterm>
<indexterm><primary>MOD_IN_USE</primary></indexterm>
</sect2>
<sect2><title>chardev.c</title>
<para>The next code sample creates a char driver named
<filename>chardev</filename>. You can <filename>cat</filename> its device
file (or <filename>open</filename> the file with a program) and the driver
will put the number of times the device file has been read from into the
file. We don't support writing to the file (like <command>echo "hi" >
/dev/hello</command>), but catch these attempts and tell the user that the
operation isn't supported. Don't worry if you don't see what we do with
the data we read into the buffer; we don't do much with it. We simply read
in the data and print a message acknowledging that we received it.
<example><title>chardev.c</title>
<programlisting><![CDATA[
/*
* chardev.c: Creates a read-only char device that says how many times
* you've read from the dev file
*/
#if defined(CONFIG_MODVERSIONS) && ! defined(MODVERSIONS)
#include <linux/modversions.h>
#define MODVERSIONS
#endif
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/fs.h>
#include <asm/uaccess.h> /* for put_user */
/* Prototypes - this would normally go in a .h file
*/
int init_module(void);
void cleanup_module(void);
static int device_open(struct inode *, struct file *);
static int device_release(struct inode *, struct file *);
static ssize_t device_read(struct file *, char *, size_t, loff_t *);
static ssize_t device_write(struct file *, const char *, size_t, loff_t *);
#define SUCCESS 0
#define DEVICE_NAME "chardev" /* Dev name as it appears in /proc/devices */
#define BUF_LEN 80 /* Max length of the message from the device */
/* Global variables are declared as static, so are global within the file. */
static int Major; /* Major number assigned to our device driver */
static int Device_Open = 0; /* Is device open? Used to prevent multiple */
access to the device */
static char msg[BUF_LEN]; /* The msg the device will give when asked */
static char *msg_Ptr;
static struct file_operations fops = {
.read = device_read,
.write = device_write,
.open = device_open,
.release = device_release
};
/* Functions
*/
int init_module(void)
{
Major = register_chrdev(0, DEVICE_NAME, &fops);
if (Major > 0) {
printk ("Registering the character device failed with %d\n", Major);
return Major;
}
printk("<1>I was assigned major number %d. To talk to\n", Major);
printk("<1>the driver, create a dev file with\n");
printk("'mknod /dev/hello c %d 0'.\n", Major);
printk("<1>Try various minor numbers. Try to cat and echo to\n");
printk("the device file.\n");
printk("<1>Remove the device file and module when done.\n");
return 0;
}
void cleanup_module(void)
{
/* Unregister the device */
int ret = unregister_chrdev(Major, DEVICE_NAME);
if (ret < 0) printk("Error in unregister_chrdev: %d\n", ret);
}
/* Methods
*/
/* Called when a process tries to open the device file, like
* "cat /dev/mycharfile"
*/
static int device_open(struct inode *inode, struct file *file)
{
static int counter = 0;
if (Device_Open) return -EBUSY;
Device_Open++;
sprintf(msg,"I already told you %d times Hello world!\n", counter++");
msg_Ptr = msg;
MOD_INC_USE_COUNT;
return SUCCESS;
}
/* Called when a process closes the device file.
*/
static int device_release(struct inode *inode, struct file *file)
{
Device_Open --; /* We're now ready for our next caller */
/* Decrement the usage count, or else once you opened the file, you'll
never get get rid of the module. */
MOD_DEC_USE_COUNT;
return 0;
}
/* Called when a process, which already opened the dev file, attempts to
read from it.
*/
static ssize_t device_read(struct file *filp,
char *buffer, /* The buffer to fill with data */
size_t length, /* The length of the buffer */
loff_t *offset) /* Our offset in the file */
{
/* Number of bytes actually written to the buffer */
int bytes_read = 0;
/* If we're at the end of the message, return 0 signifying end of file */
if (*msg_Ptr == 0) return 0;
/* Actually put the data into the buffer */
while (length && *msg_Ptr) {
/* The buffer is in the user data segment, not the kernel segment;
* assignment won't work. We have to use put_user which copies data from
* the kernel data segment to the user data segment. */
put_user(*(msg_Ptr++), buffer++);
length--;
bytes_read++;
}
/* Most read functions return the number of bytes put into the buffer */
return bytes_read;
}
/* Called when a process writes to dev file: echo "hi" > /dev/hello */
static ssize_t device_write(struct file *filp,
const char *buff,
size_t len,
loff_t *off)
{
printk ("<1>Sorry, this operation isn't supported.\n");
return -EINVAL;
}
]]></programlisting>
</example>
</sect2>
<sect2><title>Writing Modules for Multiple Kernel Versions</title>
<indexterm><primary>kernel versions</primary></indexterm>
<indexterm><primary>LINUX_VERSION_CODE</primary></indexterm>
<indexterm><primary>KERNEL_VERSION</primary></indexterm>
<para>The system calls, which are the major interface the kernel shows to
the processes, generally stay the same across versions. A new system call
may be added, but usually the old ones will behave exactly like they used
to. This is necessary for backward compatibility -- a new kernel version is
not supposed to break regular processes. In most cases, the device files
will also remain the same. On the other hand, the internal interfaces
within the kernel can and do change between versions.</para>
<para>The Linux kernel versions are divided between the stable versions
(n.$&lt;$even number$&gt;$.m) and the development versions (n.$&lt;$odd
number$&gt;$.m). The development versions include all the cool new ideas,
including those which will be considered a mistake, or reimplemented, in
the next version. As a result, you can't trust the interface to remain the
same in those versions (which is why I don't bother to support them in this
book, it's too much work and it would become dated too quickly). In the
stable versions, on the other hand, we can expect the interface to remain
the same regardless of the bug fix version (the m number).</para>
<para>There are differences between different kernel versions, and if you
want to support multiple kernel versions, you'll find yourself having to
code conditional compilation directives. The way to do this to compare the
macro <varname>LINUX_VERSION_CODE</varname> to the macro
<varname>KERNEL_VERSION</varname>. In version <varname>a.b.c</varname> of
the kernel, the value of this macro would be $2^{16}a+2^{8}b+c$. Be aware
that this macro is not defined for kernel 2.0.35 and earlier, so if you
want to write modules that support really old kernels, you'll have to
define it yourself, like:</para>
<example><title>some title</title>
<programlisting>
#if LINUX_KERNEL_VERSION >= KERNEL_VERSION(2,2,0)
#define KERNEL_VERSION(a,b,c) ((a)*65536+(b)*256+(c))
#endif
</programlisting>
</example>
<para>Of course since these are macros, you can also use <command>#ifndef
KERNEL_VERSION</command> to test the existence of the macro, rather than
testing the version of the kernel.</para>
</sect2>
</sect1>

View File

@ -0,0 +1,228 @@
<sect1><title>The /proc File System</title>
<!-- \index{proc file system} \index{/proc file system}
\index{file system\\/proc} -->
<para>In Linux there is an additional mechanism for the kernel and kernel
modules to send information to processes --- the <filename
role="directory">/proc</filename> file system. Originally designed to allow
easy access to information about processes (hence the name), it is now used
by every bit of the kernel which has something interesting to report, such as
<filename>/proc/modules</filename> which has the list of modules and
<filename>/proc/meminfo</filename> which has memory usage statistics.</para>
<!-- \index{/proc/modules} \index{/proc/meminfo} -->
<para>The method to use the proc file system is very similar to the one used
with device drivers --- you create a structure with all the information
needed for the <filename role="directory">/proc</filename> file, including
pointers to any handler functions (in our case there is only one, the one
called when somebody attempts to read from the <filename
role="directory">/proc</filename> file). Then,
<function>init_module</function> registers the structure with the kernel and
<function>cleanup_module</function> unregisters it.</para>
<para>The reason we use
<function>proc_register_dynamic</function><footnote><para>In version 2.0, in
version 2.2 this is done for us automatically if we set the inode to
zero.</para></footnote> is because we don't want to determine the inode
number used for our file in advance, but to allow the kernel to determine it
to prevent clashes. Normal file systems are located on a disk, rather than
just in memory (which is where <filename role="directory">/proc</filename>
is), and in that case the inode number is a pointer to a
disk location where the file's index-node (inode for short) is located. The
inode contains information about the file, for example the file's
permissions, together with a pointer to the disk location or locations where
the file's data can be found.</para>
<!-- \index{proc\_register\_dynamic} \index{proc\_register} \index{inode} -->
<para>Because we don't get called when the file is opened or closed, there's
no where for us to put <varname>MOD_INC_USE_COUNT</varname> and
<varname>MOD_DEC_USE_COUNT</varname> in this module, and if the file is
opened and then the module is removed, there's no way to avoid the
consequences. In the next chapter we'll see a harder to implement, but more
flexible, way of dealing with <filename role="directory">/proc</filename>
files which will allow us to protect against this problem as well.</para>
<example><title>procfs.c</title>
<programlisting><![CDATA[
/* procfs.c - create a "file" in /proc
* Copyright (C) 2001 by Peter Jay Salzman
*/
/* The necessary header files */
/* Standard in kernel modules */
#include <linux/kernel.h> /* We're doing kernel work */
#include <linux/module.h> /* Specifically, a module */
/* Deal with CONFIG_MODVERSIONS */
#if CONFIG_MODVERSIONS==1
#define MODVERSIONS
#include <linux/modversions.h>
#endif
/* Necessary because we use the proc fs */
#include <linux/proc_fs.h>
/* In 2.2.3 /usr/include/linux/version.h includes a
* macro for this, but 2.0.35 doesn't - so I add it
* here if necessary. */
#ifndef KERNEL_VERSION
#define KERNEL_VERSION(a,b,c) ((a)*65536+(b)*256+(c))
#endif
/* Put data into the proc fs file.
Arguments
=========
1. The buffer where the data is to be inserted, if
you decide to use it.
2. A pointer to a pointer to characters. This is
useful if you don't want to use the buffer
allocated by the kernel.
3. The current position in the file.
4. The size of the buffer in the first argument.
5. Zero (for future use?).
Usage and Return Value
======================
If you use your own buffer, like I do, put its
location in the second argument and return the
number of bytes used in the buffer.
A return value of zero means you have no further
information at this time (end of file). A negative
return value is an error condition.
For More Information
====================
The way I discovered what to do with this function
wasn't by reading documentation, but by reading the
code which used it. I just looked to see what uses
the get_info field of proc_dir_entry struct (I used a
combination of find and grep, if you're interested),
and I saw that it is used in <kernel source
directory>/fs/proc/array.c.
If something is unknown about the kernel, this is
usually the way to go. In Linux we have the great
advantage of having the kernel source code for
free - use it.
*/
int procfile_read(char *buffer,
char **buffer_location,
off_t offset,
int buffer_length,
int zero)
{
int len; /* The number of bytes actually used */
/* This is static so it will still be in memory
* when we leave this function */
static char my_buffer[80];
static int count = 1;
/* We give all of our information in one go, so if the
* user asks us if we have more information the
* answer should always be no.
*
* This is important because the standard read
* function from the library would continue to issue
* the read system call until the kernel replies
* that it has no more information, or until its
* buffer is filled.
*/
if (offset > 0)
return 0;
/* Fill the buffer and get its length */
len = sprintf(my_buffer,
"For the %d%s time, go away!\n", count,
(count % 100 > 10 && count % 100 < 14) ? "th" :
(count % 10 == 1) ? "st" :
(count % 10 == 2) ? "nd" :
(count % 10 == 3) ? "rd" : "th" );
count++;
/* Tell the function which called us where the
* buffer is */
*buffer_location = my_buffer;
/* Return the length */
return len;
}
struct proc_dir_entry Our_Proc_File =
{
0, /* Inode number - ignore, it will be filled by
* proc_register[_dynamic] */
4, /* Length of the file name */
"test", /* The file name */
S_IFREG | S_IRUGO, /* File mode - this is a regular
* file which can be read by its
* owner, its group, and everybody
* else */
1, /* Number of links (directories where the
* file is referenced) */
0, 0, /* The uid and gid for the file - we give it
* to root */
80, /* The size of the file reported by ls. */
NULL, /* functions which can be done on the inode
* (linking, removing, etc.) - we don't
* support any. */
procfile_read, /* The read function for this file,
* the function called when somebody
* tries to read something from it. */
NULL /* We could have here a function to fill the
* file's inode, to enable us to play with
* permissions, ownership, etc. */
};
/* Initialize the module - register the proc file */
int init_module()
{
/* Success if proc_register[_dynamic] is a success,
* failure otherwise. */
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,2,0)
/* In version 2.2, proc_register assign a dynamic
* inode number automatically if it is zero in the
* structure , so there's no more need for
* proc_register_dynamic
*/
return proc_register(&proc_root, &Our_Proc_File);
#else
return proc_register_dynamic(&proc_root, &Our_Proc_File);
#endif
/* proc_root is the root directory for the proc
* fs (/proc). This is where we want our file to be
* located.
*/
}
/* Cleanup - unregister our file from /proc */
void cleanup_module()
{
proc_unregister(&proc_root, Our_Proc_File.low_ino);
}
]]></programlisting>