1.1. What Is A Kernel Module?

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.

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.

1.2. How Do Modules Get Into The Kernel?

You can see what modules are already loaded into the kernel by running lsmod, which gets its information by reading the file /proc/modules.

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[1] execs modprobe to load the module in. modprobe is passed a string in one of two forms:

• A module name like softdog or ppp.

• A more generic identifier like char-major-10-30.

If modprobe is handed a generic identifier, it first looks for that string in the file /etc/modules.conf. If it finds an alias line like:

alias char-major-10-30 softdog
	

it knows that the generic identifier refers to the module softdog.o.

Next, modprobe looks through the file /lib/modules/version/modules.dep, to see if other modules must be loaded before the requested module may be loaded. This file is created by depmod -a and contains module dependencies. For example, msdos.o requires the fat.o 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.

Lastly, modprobe uses insmod to first load any prerequisite modules into the kernel, and then the requested module. modprobe directs insmod to /lib/modules/version/[2], 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:

insmod /lib/modules/2.5.1/kernel/fs/fat/fat.o
insmod /lib/modules/2.5.1/kernel/fs/msdos/msdos.o
	

or just run "modprobe -a msdos".

Linux distros provide modprobe, insmod and depmod as a package called modutils or mod-utils.

Before finishing this chapter, let's take a quick look at a piece of /etc/modules.conf:

#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
	

Lines beginning with a '#' are comments. Blank lines are ignored.

The path[misc] line tells modprobe to replace the search path for misc modules with the directory /lib/modules/2.4.?/local. As you can see, shell meta characters are honored.

The path[net] line tells modprobe to look for net modules in the directory ~p/mymodules, however, the "keep" directive preceding the path[net] 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.

The alias line says to load in eepro.o whenever kmod requests that the generic identifier `eth0' be loaded.

You won't see lines like "alias block-major-2 floppy" in /etc/modules.conf because modprobe already knows about the standard drivers which will be used on most systems.

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.

2.1. Hello, World (part 1): The Simplest Module

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.

Here's the simplest module possible. Don't compile it yet; we'll cover module compilation in the next section.

/*  
 *  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");
}

Kernel modules must have at least two functions: a "start" (initialization) function called init_module() which is called when the module is insmoded into the kernel, and an "end" (cleanup) function called cleanup_module() which is called just before it is rmmoded. Actually, things have changed starting with kernel 2.3.13. You can now use whatever name you like for the start and end functions of a module, and you'll learn how to do this in Section 2.3. In fact, the new method is the preferred method. However, many people still use init_module() and cleanup_module() for their start and end functions.

Typically, init_module() 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 cleanup_module() function is supposed to undo whatever init_module() did, so the module can be unloaded safely.

Lastly, every kernel module needs to include linux/module.h. We needed to include linux/kernel.h only for the macro expansion for the printk() log level, KERN_ALERT, which you'll learn about in Section 2.1.1.

2.2. Compiling Kernel Modules

Kernel modules need to be compiled with certain gcc options to make them work. In addition, they also need to be compiled with certain symbols defined. Former kernel versions required us to care much about these settings, which are usually stored in Makefiles. Although hierarchically organized, many redundant settings accumulated in sublevel Makefiles and made them large and rather difficult to maintain. Fortunately, there is a new way of doing these things, called kbuild, and the build process for external loadable modules is now fully integrated into the standard kernel build mechanism. To learn more on how to compile modules which are not part of the official kernel (as ours), see file linux/Documentation/kbuild/modules.txt.

So, let's look at a simple Makefile for compiling a module named hello-1.c:

obj-m += hello-1.o

Now you can compile the module by issuing the command make -C /usr/src/linux-`uname -r` SUBDIRS=$PWD modules . You should obtain an output which resembles the following:

[root@pcsenonsrv test_module]# make -C /usr/src/linux-`uname -r` SUBDIRS=$PWD modules
make: Entering directory `/usr/src/linux-2.6.x
  CC [M]  /root/test_module/hello-1.o
  Building modules, stage 2.
  MODPOST
  CC      /root/test_module/hello-1.mod.o
  LD [M]  /root/test_module/hello-1.ko
make: Leaving directory `/usr/src/linux-2.6.x
	

Please note that kernel 2.6 introduces a new file naming convention: kernel modules now have a .ko extension (in place of the old .o extension) which easily distinguishes them from conventional object files. Additional details about Makefiles for kernel modules are available in linux/Documentation/kbuild/makefiles.txt. Be sure to read this and the related files before starting to dig into Makefiles.

Now it is time to insert your freshly-compiled module it into the kernel with insmod ./hello-1.ko (ignore anything you see about tainted kernels; we'll cover that shortly).

All modules loaded into the kernel are listed in /proc/modules. 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, remove your module from the kernel by using rmmod hello-1. Take a look at /var/log/messages just to see that it got logged to your system logfile.

Here's another exercise to the reader. See that comment above the return statement in init_module()? Change the return value to something non-zero, recompile and load the module again. What happens?

2.3. Hello World (part 2)

As of Linux 2.4, you can rename the init and cleanup functions of your modules; they no longer have to be called init_module() and cleanup_module() respectively. This is done with the module_init() and module_exit() macros. These macros are defined in linux/init.h. 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:

/*  
 *  hello-2.c - Demonstrating the module_init() and module_exit() macros.
 *  This is preferred over using init_module() and cleanup_module().
 */
#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 __init hello_2_init(void)
{
	printk(KERN_ALERT "Hello, world 2\n");
	return 0;
}

static void __exit hello_2_exit(void)
{
	printk(KERN_ALERT "Goodbye, world 2\n");
}

module_init(hello_2_init);
module_exit(hello_2_exit);

So now we have two real kernel modules under our belt. Adding another module is as simple as this:

obj-m += hello-1.o
obj-m += hello-2.o

Now have a look at linux/drivers/char/Makefile for a real world example. As you can see, some things get hardwired into the kernel (obj-y) but where are all those obj-m gone? Those familiar with shell scripts will easily be able to spot them. For those not, the obj-$(CONFIG_FOO) entries you see everywhere expand into obj-y or obj-m, depending on whether the CONFIG_FOO variable has been set to y or m. While we are at it, those were exactly the kind of variables that you have set in the linux/.config file, the last time when you said make menuconfig or something like that.

2.4. Hello World (part 3): The __init and __exit Macros

This demonstrates a feature of kernel 2.2 and later. Notice the change in the definitions of the init and cleanup functions. The __init macro causes the init function to be discarded and its memory freed once the init function finishes for built-in drivers, but not loadable modules. If you think about when the init function is invoked, this makes perfect sense.

There is also an __initdata which works similarly to __init but for init variables rather than functions.

The __exit macro causes the omission of the function when the module is built into the kernel, and like __exit, has no effect for loadable modules. Again, if you consider when the cleanup function runs, this makes complete sense; built-in drivers don't need a cleanup function, while loadable modules do.

These macros are defined in linux/init.h and serve to free up kernel memory. When you boot your kernel and see something like Freeing unused kernel memory: 236k freed, this is precisely what the kernel is freeing.

/*  
 *  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 hello_3_init(void)
{
	printk(KERN_ALERT "Hello, world %d\n", hello3_data);
	return 0;
}

static void __exit hello_3_exit(void)
{
	printk(KERN_ALERT "Goodbye, world 3\n");
}

module_init(hello_3_init);
module_exit(hello_3_exit);

2.5. Hello World (part 4): Licensing and Module Documentation

If you're running kernel 2.4 or later, you might have noticed something like this when you loaded the previous example modules:

# insmod hello-3.o
Warning: loading hello-3.o will taint the kernel: no license
  See http://www.tux.org/lkml/#export-tainted for information about tainted modules
Hello, world 3
Module hello-3 loaded, with warnings
	

In kernel 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 MODULE_LICENSE() 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 license mechanism is defined and documented in linux/module.h:

/* * The following license idents are currently accepted as indicating free * software modules * * "GPL" [GNU Public License v2 or later] * "GPL v2" [GNU Public License v2] * "GPL and additional rights" [GNU Public License v2 rights and more] * "Dual BSD/GPL" [GNU Public License v2 * or BSD license choice] * "Dual MPL/GPL" [GNU Public License v2 * or Mozilla license choice] * * The following other idents are available * * "Proprietary" [Non free products] * * There are dual licensed components, but when running with Linux it is the * GPL that is relevant so this is a non issue. Similarly LGPL linked with GPL * is a GPL combined work. * * This exists for several reasons * 1. So modinfo can show license info for users wanting to vet their setup * is free * 2. So the community can ignore bug reports including proprietary modules * 3. So vendors can do likewise based on their own policies */

Similarly, MODULE_DESCRIPTION() is used to describe what the module does, MODULE_AUTHOR() declares the module's author, and MODULE_SUPPORTED_DEVICE() declares what types of devices the module supports.

These macros are all defined in linux/module.h and aren't used by the kernel itself. They're simply for documentation and can be viewed by a tool like objdump. As an exercise to the reader, try grepping through linux/drivers to see how module authors use these macros to document their modules.

/*  
 *  hello-4.c - Demonstrates module 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 init_hello_4(void)
{
	printk(KERN_ALERT "Hello, world 4\n");
	return 0;
}

static void __exit cleanup_hello_4(void)
{
	printk(KERN_ALERT "Goodbye, world 4\n");
}

module_init(init_hello_4);
module_exit(cleanup_hello_4);

/*  
 *  You can use strings, like this:
 */

/* 
 * Get rid of taint message by declaring code as GPL. 
 */
MODULE_LICENSE("GPL");

/*
 * Or with defines, like this:
 */
MODULE_AUTHOR(DRIVER_AUTHOR);	/* Who wrote this module? */
MODULE_DESCRIPTION(DRIVER_DESC);	/* What does this module do */

/*  
 *  This module uses /dev/testdevice.  The MODULE_SUPPORTED_DEVICE macro might
 *  be used in the future to help automatic configuration of modules, but is 
 *  currently unused other than for documentation purposes.
 */
MODULE_SUPPORTED_DEVICE("testdevice");

2.6. Passing Command Line Arguments to a Module

Modules can take command line arguments, but not with the argc/argv you might be used to.

To allow arguments to be passed to your module, declare the variables that will take the values of the command line arguments as global and then use the MODULE_PARM() macro, (defined in linux/module.h) to set the mechanism up. At runtime, insmod will fill the variables with any command line arguments that are given, like ./insmod mymodule.o myvariable=5. 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.

The MODULE_PARM() macro takes 2 arguments: the name of the variable and its type. The supported variable types are "b": single byte, "h": short int, "i": integer, "l": long int and "s": string, and the integer types can be signed as usual or unsigned. Strings should be declared as "char *" 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:

int myint = 3;
char *mystr;

MODULE_PARM(myint, "i");
MODULE_PARM(mystr, "s");
	

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:

int myshortArray[4];
MODULE_PARM (myintArray, "3-9i");
	

A good use for this is to have the module variable's default values set, like an port or IO address. If the variables contain the default values, then perform autodetection (explained elsewhere). Otherwise, keep the current value. This will be made clear later on.

Lastly, there's a macro function, MODULE_PARM_DESC(), that is used to document arguments that the module can take. It takes two parameters: a variable name and a free form string describing that variable.

/*
 *  hello-5.c - Demonstrates command line argument passing to a module.
 */
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/stat.h>

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Peter Jay Salzman");

static short int myshort = 1;
static int myint = 420;
static long int mylong = 9999;
static char *mystring = "blah";

/* 
 * module_param(foo, int, 0000)
 * The first param is the parameters name
 * The second param is it's data type
 * The final argument is the permissions bits, 
 * for exposing parameters in sysfs (if non-zero) at a later stage.
 */

module_param(myshort, short, S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP);
MODULE_PARM_DESC(myshort, "A short integer");
module_param(myint, int, S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH);
MODULE_PARM_DESC(myint, "An integer");
module_param(mylong, long, S_IRUSR);
MODULE_PARM_DESC(mylong, "A long integer");
module_param(mystring, charp, 0000);
MODULE_PARM_DESC(mystring, "A character string");

static int __init hello_5_init(void)
{
	printk(KERN_ALERT "Hello, world 5\n=============\n");
	printk(KERN_ALERT "myshort is a short integer: %hd\n", myshort);
	printk(KERN_ALERT "myint is an integer: %d\n", myint);
	printk(KERN_ALERT "mylong is a long integer: %ld\n", mylong);
	printk(KERN_ALERT "mystring is a string: %s\n", mystring);
	return 0;
}

static void __exit hello_5_exit(void)
{
	printk(KERN_ALERT "Goodbye, world 5\n");
}

module_init(hello_5_init);
module_exit(hello_5_exit);

I would recommend playing around with this code:

satan# insmod hello-5.o mystring="bebop" mybyte=255 myintArray=-1
mybyte is an 8 bit integer: 255
myshort is a short integer: 1
myint is an integer: 20
mylong is a long integer: 9999
mystring is a string: bebop
myintArray is -1 and 420

satan# rmmod hello-5
Goodbye, world 5

satan# insmod hello-5.o mystring="supercalifragilisticexpialidocious" \
> mybyte=256 myintArray=-1,-1
mybyte is an 8 bit integer: 0
myshort is a short integer: 1
myint is an integer: 20
mylong is a long integer: 9999
mystring is a string: supercalifragilisticexpialidocious
myintArray is -1 and -1

satan# rmmod hello-5
Goodbye, world 5

satan# insmod hello-5.o mylong=hello
hello-5.o: invalid argument syntax for mylong: 'h'

2.7. Modules Spanning Multiple Files

Sometimes it makes sense to divide a kernel module between several source files. In this case, you need to:

1. In all the source files but one, add the line #define __NO_VERSION__. This is important because module.h normally includes the definition of kernel_version, a global variable with the kernel version the module is compiled for. If you need version.h, you need to include it yourself, because module.h won't do it for you with __NO_VERSION__.

2. Compile all the source files as usual.

3. Combine all the object files into a single one. Under x86, use ld -m elf_i386 -r -o <module name.o> <1st src file.o> <2nd src file.o>.

The makefile will, once again, save us from having to get our hands dirty with compiling and linking the object files.

Here's an example of such a kernel module.

/*
 *  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;
}

The next file:

/*
 *  stop.c - Illustration of multi filed modules
 */

#include <linux/kernel.h>	/* We're doing kernel work */
#include <linux/module.h>	/* Specifically, a module  */

void cleanup_module()
{
	printk("<1>Short is the life of a kernel module\n");
}

And finally, the makefile:

obj-m += hello-1.o
obj-m += hello-2.o
obj-m += hello-3.o
obj-m += hello-4.o
obj-m += hello-5.o
obj-m += startstop.o
startstop-objs := start.o stop.o

2.8. Building modules for a precompiled kernel

Obviously, we strongly suggest you to recompile your kernel, so that you can enable a number of useful debugging features, such as forced module unloading (MODULE_FORCE_UNLOAD): when this option is enabled, you can force the kernel to unload a module even when it believes it is unsafe, via a rmmod -f module command. This option can save you a lot of time and a number of reboots during the development of a module.

Nevertheless, there is a number of cases in which you may want to load your module into a precompiled running kernel, such as the ones shipped with common Linux distributions, or a kernel you have compiled in the past. In certain circumstances you could require to compile and insert a module into a running kernel which you are not allowed to recompile, or on a machine that you prefer not to reboot. If you can't think of a case that will force you to use modules for a precompiled kernel you might want to skip this and treat the rest of this chapter as a big footnote.

Now, if you just install a kernel source tree, use it to compile your kernel module and you try to insert your module into the kernel, in most cases you would obtain an error as follows:

insmod: error inserting 'poet_atkm.ko': -1 Invalid module format
	

Less cryptical information are logged to /var/log/messages:

Jun  4 22:07:54 localhost kernel: poet_atkm: version magic '2.6.5-1.358custom 686 
REGPARM 4KSTACKS gcc-3.3' should be '2.6.5-1.358 686 REGPARM 4KSTACKS gcc-3.3'
	

In other words, your kernel refuses to accept your module because version strings (more precisely, version magics) do not match. Incidentally, version magics are stored in the module object in the form of a static string, starting with vermagic:. Version data are inserted in your module when it is linked against the init/vermagic.o file. To inspect version magics and other strings stored in a given module, issue the modinfo module.ko command:

[root@pcsenonsrv 02-HelloWorld]# modinfo hello-4.ko 
license:        GPL
author:         Peter Jay Salzman <p@dirac.org>
description:    A sample driver
vermagic:       2.6.5-1.358 686 REGPARM 4KSTACKS gcc-3.3
depends:        
	

To overcome this problem we could resort to the --force-vermagic option, but this solution is potentially unsafe, and unquestionably inacceptable in production modules. Consequently, we want to compile our module in an environment which was identical to the one in which our precompiled kernel was built. How to do this, is the subject of the remainder of this chapter.

First of all, make sure that a kernel source tree is available, having exactly the same version as your current kernel. Then, find the configuration file which was used to compile your precompiled kernel. Usually, this is available in your current /boot directory, under a name like config-2.6.x. You may just want to copy it to your kernel source tree: cp /boot/config-`uname -r` /usr/src/linux-`uname -r`/.config.

Let's focus again on the previous error message: a closer look at the version magic strings suggests that, even with two configuration files which are exactly the same, a slight difference in the version magic could be possible, and it is sufficient to prevent insertion of the module into the kernel. That slight difference, namely the custom string which appears in the module's version magic and not in the kernel's one, is due to a modification with respect to the original, in the makefile that some distribution include. Then, examine your /usr/src/linux/Makefile, and make sure that the specified version information matches exactly the one used for your current kernel. For example, you makefile could start as follows:

VERSION = 2
PATCHLEVEL = 6
SUBLEVEL = 5
EXTRAVERSION = -1.358custom
...
	

In this case, you need to restore the value of symbol EXTRAVERSION to -1.358. We suggest to keep a backup copy of the makefile used to compile your kernel available in /lib/modules/2.6.5-1.358/build. A simple cp /lib/modules/`uname -r`/build/Makefile /usr/src/linux-`uname -r` should suffice. Additionally, if you already started a kernel build with the previous (wrong) Makefile, you should also rerun make, or directly modify symbol UTS_RELEASE in file /usr/src/linux-2.6.x/include/linux/version.h according to contents of file /lib/modules/2.6.x/build/include/linux/version.h, or overwrite the latter with the first.

Now, please run make to update configuration and version headers and objects:

[root@pcsenonsrv linux-2.6.x]# make
CHK     include/linux/version.h
UPD     include/linux/version.h
SYMLINK include/asm -> include/asm-i386
SPLIT   include/linux/autoconf.h -> include/config/*
HOSTCC  scripts/basic/fixdep
HOSTCC  scripts/basic/split-include
HOSTCC  scripts/basic/docproc
HOSTCC  scripts/conmakehash
HOSTCC  scripts/kallsyms
CC      scripts/empty.o
...
	

If you do not desire to actually compile the kernel, you can interrupt the build process (CTRL-C) just after the SPLIT line, because at that time, the files you need will be are ready. Now you can turn back to the directory of your module and compile it: It will be built exactly according your current kernel settings, and it will load into it without any errors.

3.1. Modules vs Programs

#include <stdio.h>
int main(void)
{ printf("hello"); return 0; }
				

# 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
					

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
					

% 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
					

4.1. Character Device Drivers

struct file_operations {
	struct module *owner;
	 loff_t(*llseek) (struct file *, loff_t, int);
	 ssize_t(*read) (struct file *, char __user *, size_t, loff_t *);
	 ssize_t(*aio_read) (struct kiocb *, char __user *, size_t, loff_t);
	 ssize_t(*write) (struct file *, const char __user *, size_t, loff_t *);
	 ssize_t(*aio_write) (struct kiocb *, const char __user *, 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 (*aio_fsync) (struct kiocb *, 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 *);
	 ssize_t(*sendfile) (struct file *, loff_t *, size_t, read_actor_t,
			     void __user *);
	 ssize_t(*sendpage) (struct file *, struct page *, int, size_t,
			     loff_t *, int);
	unsigned long (*get_unmapped_area) (struct file *, unsigned long,
					    unsigned long, unsigned long,
					    unsigned long);
};
			

struct file_operations fops = {
	read: device_read,
	write: device_write,
	open: device_open,
	release: device_release
};
			

struct file_operations fops = {
	.read = device_read,
	.write = device_write,
	.open = device_open,
	.release = device_release
};
			

int register_chrdev(unsigned int major, const char *name, struct file_operations *fops);
			

/*
 *  chardev.c: Creates a read-only char device that says how many times
 *  you've read from the dev file
 */

#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 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;
	try_module_get(THIS_MODULE);

	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. 
	 */
	module_put(THIS_MODULE);

	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,	/* see include/linux/fs.h   */
			   char *buffer,	/* buffer to fill with data */
			   size_t length,	/* length of the buffer     */
			   loff_t * offset)
{
	/*
	 * 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 so "*" 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;
}

5.1. The /proc File System

In Linux there is an additional mechanism for the kernel and kernel modules to send information to processes --- the /proc 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 /proc/modules which has the list of modules and /proc/meminfo which has memory usage statistics.

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 /proc file, including pointers to any handler functions (in our case there is only one, the one called when somebody attempts to read from the /proc file). Then, init_module registers the structure with the kernel and cleanup_module unregisters it.

The reason we use proc_register_dynamic[7] 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 /proc 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.

Because we don't get called when the file is opened or closed, there's no where for us to put try_module_get and try_module_put 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 /proc files which will allow us to protect against this problem as well.

/*
 *  procfs.c -  create a "file" in /proc
 */

#include <linux/module.h>	/* Specifically, a module */
#include <linux/kernel.h>	/* We're doing kernel work */
#include <linux/proc_fs.h>	/* Necessary because we use the proc fs */

struct proc_dir_entry *Our_Proc_File;

/* 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. Write a "1" here to indicate EOF.
 * 6. A pointer to data (useful in case one common 
 *    read for multiple /proc/... entries)
 *
 * Usage and Return Value
 * ======================
 * 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.
 */
ssize_t
procfile_read(char *buffer,
	      char **buffer_location,
	      off_t offset, int buffer_length, int *eof, void *data)
{
	printk(KERN_INFO "inside /proc/test : procfile_read\n");

	int len = 0;		/* The number of bytes actually used */
	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) {
		printk(KERN_INFO "offset %d : /proc/test : procfile_read, \
		       wrote %d Bytes\n", (int)(offset), len);
		*eof = 1;
		return len;
	}

	/* 
	 * Fill the buffer and get its length 
	 */
	len = sprintf(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++;

	/* 
	 * Return the length 
	 */
	printk(KERN_INFO
	       "leaving /proc/test : procfile_read, wrote %d Bytes\n", len);
	return len;
}

int init_module()
{
	int rv = 0;
	Our_Proc_File = create_proc_entry("test", 0644, NULL);
	Our_Proc_File->read_proc = procfile_read;
	Our_Proc_File->owner = THIS_MODULE;
	Our_Proc_File->mode = S_IFREG | S_IRUGO;
	Our_Proc_File->uid = 0;
	Our_Proc_File->gid = 0;
	Our_Proc_File->size = 37;

	printk(KERN_INFO "Trying to create /proc/test:\n");

	if (Our_Proc_File == NULL) {
		rv = -ENOMEM;
		remove_proc_entry("test", &proc_root);
		printk(KERN_INFO "Error: Could not initialize /proc/test\n");
	} else {
		printk(KERN_INFO "Success!\n");
	}

	return rv;
}

void cleanup_module()
{
	remove_proc_entry("test", &proc_root);
	printk(KERN_INFO "/proc/test removed\n");
}

6.1. Using /proc For Input

So far we have two ways to generate output from kernel modules: we can register a device driver and mknod a device file, or we can create a /proc file. This allows the kernel module to tell us anything it likes. The only problem is that there is no way for us to talk back. The first way we'll send input to kernel modules will be by writing back to the /proc file.

Because the proc filesystem was written mainly to allow the kernel to report its situation to processes, there are no special provisions for input. The struct proc_dir_entry doesn't include a pointer to an input function, the way it includes a pointer to an output function. Instead, to write into a /proc file, we need to use the standard filesystem mechanism.

In Linux there is a standard mechanism for file system registration. Since every file system has to have its own functions to handle inode and file operations[8], there is a special structure to hold pointers to all those functions, struct inode_operations, which includes a pointer to struct file_operations. In /proc, whenever we register a new file, we're allowed to specify which struct inode_operations will be used for access to it. This is the mechanism we use, a struct inode_operations which includes a pointer to a struct file_operations which includes pointers to our module_input and module_output functions.

It's important to note that the standard roles of read and write are reversed in the kernel. Read functions are used for output, whereas write functions are used for input. The reason for that is that read and write refer to the user's point of view --- if a process reads something from the kernel, then the kernel needs to output it, and if a process writes something to the kernel, then the kernel receives it as input.

Another interesting point here is the module_permission function. This function is called whenever a process tries to do something with the /proc file, and it can decide whether to allow access or not. Right now it is only based on the operation and the uid of the current user (as available in current, a pointer to a structure which includes information on the currently running process), but it could be based on anything we like, such as what other processes are doing with the same file, the time of day, or the last input we received.

The reason for put_user and get_user is that Linux memory (under Intel architecture, it may be different under some other processors) is segmented. This means that a pointer, by itself, does not reference a unique location in memory, only a location in a memory segment, and you need to know which memory segment it is to be able to use it. There is one memory segment for the kernel, and one of each of the processes.

The only memory segment accessible to a process is its own, so when writing regular programs to run as processes, there's no need to worry about segments. When you write a kernel module, normally you want to access the kernel memory segment, which is handled automatically by the system. However, when the content of a memory buffer needs to be passed between the currently running process and the kernel, the kernel function receives a pointer to the memory buffer which is in the process segment. The put_user and get_user macros allow you to access that memory.

/* 
 *  procfs.c -  create a "file" in /proc, which allows both input and output.
 */
#include <linux/kernel.h>	/* We're doing kernel work */
#include <linux/module.h>	/* Specifically, a module */
#include <linux/proc_fs.h>	/* Necessary because we use proc fs */
#include <asm/uaccess.h>	/* for get_user and put_user */

/* 
 * Here we keep the last message received, to prove
 * that we can process our input 
 */
#define MESSAGE_LENGTH 80
static char Message[MESSAGE_LENGTH];
static struct proc_dir_entry *Our_Proc_File;

#define PROC_ENTRY_FILENAME "rw_test"

static ssize_t module_output(struct file *filp,	/* see include/linux/fs.h   */
			     char *buffer,	/* buffer to fill with data */
			     size_t length,	/* length of the buffer     */
			     loff_t * offset)
{
	static int finished = 0;
	int i;
	char message[MESSAGE_LENGTH + 30];

	/* 
	 * We return 0 to indicate end of file, that we have
	 * no more information. Otherwise, processes will
	 * continue to read from us in an endless loop. 
	 */
	if (finished) {
		finished = 0;
		return 0;
	}

	/* 
	 * We use put_user to copy the string from the kernel's
	 * memory segment to the memory segment of the process
	 * that called us. get_user, BTW, is
	 * used for the reverse. 
	 */
	sprintf(message, "Last input:%s", Message);
	for (i = 0; i < length && message[i]; i++)
		put_user(message[i], buffer + i);

	/* 
	 * Notice, we assume here that the size of the message
	 * is below len, or it will be received cut. In a real
	 * life situation, if the size of the message is less
	 * than len then we'd return len and on the second call
	 * start filling the buffer with the len+1'th byte of
	 * the message. 
	 */
	finished = 1;

	return i;		/* Return the number of bytes "read" */
}

static ssize_t
module_input(struct file *filp, const char *buff, size_t len, loff_t * off)
{
	int i;
	/* 
	 * Put the input into Message, where module_output
	 * will later be able to use it 
	 */
	for (i = 0; i < MESSAGE_LENGTH - 1 && i < len; i++)
		get_user(Message[i], buff + i);

	Message[i] = '\0';	/* we want a standard, zero terminated string */
	return i;
}

/* 
 * This function decides whether to allow an operation
 * (return zero) or not allow it (return a non-zero
 * which indicates why it is not allowed).
 *
 * The operation can be one of the following values:
 * 0 - Execute (run the "file" - meaningless in our case)
 * 2 - Write (input to the kernel module)
 * 4 - Read (output from the kernel module)
 *
 * This is the real function that checks file
 * permissions. The permissions returned by ls -l are
 * for referece only, and can be overridden here.
 */

static int module_permission(struct inode *inode, int op, struct nameidata *foo)
{
	/* 
	 * We allow everybody to read from our module, but
	 * only root (uid 0) may write to it 
	 */
	if (op == 4 || (op == 2 && current->euid == 0))
		return 0;

	/* 
	 * If it's anything else, access is denied 
	 */
	return -EACCES;
}

/* 
 * The file is opened - we don't really care about
 * that, but it does mean we need to increment the
 * module's reference count. 
 */
int module_open(struct inode *inode, struct file *file)
{
	try_module_get(THIS_MODULE);
	return 0;
}

/* 
 * The file is closed - again, interesting only because
 * of the reference count. 
 */
int module_close(struct inode *inode, struct file *file)
{
	module_put(THIS_MODULE);
	return 0;		/* success */
}

static struct file_operations File_Ops_4_Our_Proc_File = {
	.read = module_output,
	.write = module_input,
	.open = module_open,
	.release = module_close,
};

/* 
 * Inode operations for our proc file. We need it so
 * we'll have some place to specify the file operations
 * structure we want to use, and the function we use for
 * permissions. It's also possible to specify functions
 * to be called for anything else which could be done to
 * an inode (although we don't bother, we just put
 * NULL). 
 */

static struct inode_operations Inode_Ops_4_Our_Proc_File = {
	.permission = module_permission,	/* check for permissions */
};

/* 
 * Module initialization and cleanup 
 */
int init_module()
{
	int rv = 0;
	Our_Proc_File = create_proc_entry(PROC_ENTRY_FILENAME, 0644, NULL);
	Our_Proc_File->owner = THIS_MODULE;
	Our_Proc_File->proc_iops = &Inode_Ops_4_Our_Proc_File;
	Our_Proc_File->proc_fops = &File_Ops_4_Our_Proc_File;
	Our_Proc_File->mode = S_IFREG | S_IRUGO | S_IWUSR;
	Our_Proc_File->uid = 0;
	Our_Proc_File->gid = 0;
	Our_Proc_File->size = 80;

	if (Our_Proc_File == NULL) {
		rv = -ENOMEM;
		remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
		printk(KERN_INFO "Error: Could not initialize /proc/test\n");
	}

	return rv;
}

void cleanup_module()
{
	remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
}

Still hungry for procfs examples? Well, first of all keep in mind, there are rumors around, claiming that procfs is on it's way out, consider using sysfs instead. Second, if you really can't get enough, there's a highly recommendable bonus level for procfs below linux/Documentation/DocBook/ . Use make help in your toplevel kernel directory for instructions about how to convert it into your favourite format. Example: make htmldocs . Consider using this mechanism, in case you want to document something kernel related yourself.

7.1. Talking to Device Files (writes and IOCTLs)}

Device files are supposed to represent physical devices. Most physical devices are used for output as well as input, so there has to be some mechanism for device drivers in the kernel to get the output to send to the device from processes. This is done by opening the device file for output and writing to it, just like writing to a file. In the following example, this is implemented by device_write.

This is not always enough. Imagine you had a serial port connected to a modem (even if you have an internal modem, it is still implemented from the CPU's perspective as a serial port connected to a modem, so you don't have to tax your imagination too hard). The natural thing to do would be to use the device file to write things to the modem (either modem commands or data to be sent through the phone line) and read things from the modem (either responses for commands or the data received through the phone line). However, this leaves open the question of what to do when you need to talk to the serial port itself, for example to send the rate at which data is sent and received.

The answer in Unix is to use a special function called ioctl (short for Input Output ConTroL). Every device can have its own ioctl commands, which can be read ioctl's (to send information from a process to the kernel), write ioctl's (to return information to a process), [9] both or neither. The ioctl function is called with three parameters: the file descriptor of the appropriate device file, the ioctl number, and a parameter, which is of type long so you can use a cast to use it to pass anything. [10]

The ioctl number encodes the major device number, the type of the ioctl, the command, and the type of the parameter. This ioctl number is usually created by a macro call (_IO, _IOR, _IOW or _IOWR --- depending on the type) in a header file. This header file should then be included both by the programs which will use ioctl (so they can generate the appropriate ioctl's) and by the kernel module (so it can understand it). In the example below, the header file is chardev.h and the program which uses it is ioctl.c.

If you want to use ioctls in your own kernel modules, it is best to receive an official ioctl assignment, so if you accidentally get somebody else's ioctls, or if they get yours, you'll know something is wrong. For more information, consult the kernel source tree at Documentation/ioctl-number.txt.

/*
 *  chardev.c - Create an input/output character device
 */

#include <linux/kernel.h>	/* We're doing kernel work */
#include <linux/module.h>	/* Specifically, a module */
#include <linux/fs.h>
#include <asm/uaccess.h>	/* for get_user and put_user */

#include "chardev.h"
#define SUCCESS 0
#define DEVICE_NAME "char_dev"
#define BUF_LEN 80

/* 
 * Is the device open right now? Used to prevent
 * concurent access into the same device 
 */
static int Device_Open = 0;

/* 
 * The message the device will give when asked 
 */
static char Message[BUF_LEN];

/* 
 * How far did the process reading the message get?
 * Useful if the message is larger than the size of the
 * buffer we get to fill in device_read. 
 */
static char *Message_Ptr;

/* 
 * This is called whenever a process attempts to open the device file 
 */
static int device_open(struct inode *inode, struct file *file)
{
#ifdef DEBUG
	printk("device_open(%p)\n", file);
#endif

	/* 
	 * We don't want to talk to two processes at the same time 
	 */
	if (Device_Open)
		return -EBUSY;

	Device_Open++;
	/*
	 * Initialize the message 
	 */
	Message_Ptr = Message;
	try_module_get(THIS_MODULE);
	return SUCCESS;
}

static int device_release(struct inode *inode, struct file *file)
{
#ifdef DEBUG
	printk("device_release(%p,%p)\n", inode, file);
#endif

	/* 
	 * We're now ready for our next caller 
	 */
	Device_Open--;

	module_put(THIS_MODULE);
	return SUCCESS;
}

/* 
 * This function is called whenever a process which has already opened the
 * device file attempts to read from it.
 */
static ssize_t device_read(struct file *file,	/* see include/linux/fs.h   */
			   char __user * buffer,	/* buffer to be
							 * filled with data */
			   size_t length,	/* length of the buffer     */
			   loff_t * offset)
{
	/* 
	 * Number of bytes actually written to the buffer 
	 */
	int bytes_read = 0;

#ifdef DEBUG
	printk("device_read(%p,%p,%d)\n", file, buffer, length);
#endif

	/* 
	 * If we're at the end of the message, return 0
	 * (which signifies end of file) 
	 */
	if (*Message_Ptr == 0)
		return 0;

	/* 
	 * Actually put the data into the buffer 
	 */
	while (length && *Message_Ptr) {

		/* 
		 * Because the buffer is in the user data segment,
		 * not the kernel data segment, assignment wouldn't
		 * work. Instead, we have to use put_user which
		 * copies data from the kernel data segment to the
		 * user data segment. 
		 */
		put_user(*(Message_Ptr++), buffer++);
		length--;
		bytes_read++;
	}

#ifdef DEBUG
	printk("Read %d bytes, %d left\n", bytes_read, length);
#endif

	/* 
	 * Read functions are supposed to return the number
	 * of bytes actually inserted into the buffer 
	 */
	return bytes_read;
}

/* 
 * This function is called when somebody tries to
 * write into our device file. 
 */
static ssize_t
device_write(struct file *file,
	     const char __user * buffer, size_t length, loff_t * offset)
{
	int i;

#ifdef DEBUG
	printk("device_write(%p,%s,%d)", file, buffer, length);
#endif

	for (i = 0; i < length && i < BUF_LEN; i++)
		get_user(Message[i], buffer + i);

	Message_Ptr = Message;

	/* 
	 * Again, return the number of input characters used 
	 */
	return i;
}

/* 
 * This function is called whenever a process tries to do an ioctl on our
 * device file. We get two extra parameters (additional to the inode and file
 * structures, which all device functions get): the number of the ioctl called
 * and the parameter given to the ioctl function.
 *
 * If the ioctl is write or read/write (meaning output is returned to the
 * calling process), the ioctl call returns the output of this function.
 *
 */
int device_ioctl(struct inode *inode,	/* see include/linux/fs.h */
		 struct file *file,	/* ditto */
		 unsigned int ioctl_num,	/* number and param for ioctl */
		 unsigned long ioctl_param)
{
	int i;
	char *temp;
	char ch;

	/* 
	 * Switch according to the ioctl called 
	 */
	switch (ioctl_num) {
	case IOCTL_SET_MSG:
		/* 
		 * Receive a pointer to a message (in user space) and set that
		 * to be the device's message.  Get the parameter given to 
		 * ioctl by the process. 
		 */
		temp = (char *)ioctl_param;

		/* 
		 * Find the length of the message 
		 */
		get_user(ch, temp);
		for (i = 0; ch && i < BUF_LEN; i++, temp++)
			get_user(ch, temp);

		device_write(file, (char *)ioctl_param, i, 0);
		break;

	case IOCTL_GET_MSG:
		/* 
		 * Give the current message to the calling process - 
		 * the parameter we got is a pointer, fill it. 
		 */
		i = device_read(file, (char *)ioctl_param, 99, 0);

		/* 
		 * Put a zero at the end of the buffer, so it will be 
		 * properly terminated 
		 */
		put_user('\0', (char *)ioctl_param + i);
		break;

	case IOCTL_GET_NTH_BYTE:
		/* 
		 * This ioctl is both input (ioctl_param) and 
		 * output (the return value of this function) 
		 */
		return Message[ioctl_param];
		break;
	}

	return SUCCESS;
}

/* Module Declarations */

/* 
 * This structure will hold the functions to be called
 * when a process does something to the device we
 * created. Since a pointer to this structure is kept in
 * the devices table, it can't be local to
 * init_module. NULL is for unimplemented functions. 
 */
struct file_operations Fops = {
	.read = device_read,
	.write = device_write,
	.ioctl = device_ioctl,
	.open = device_open,
	.release = device_release,	/* a.k.a. close */
};

/* 
 * Initialize the module - Register the character device 
 */
int init_module()
{
	int ret_val;
	/* 
	 * Register the character device (atleast try) 
	 */
	ret_val = register_chrdev(MAJOR_NUM, DEVICE_NAME, &Fops);

	/* 
	 * Negative values signify an error 
	 */
	if (ret_val < 0) {
		printk("%s failed with %d\n",
		       "Sorry, registering the character device ", ret_val);
		return ret_val;
	}

	printk("%s The major device number is %d.\n",
	       "Registeration is a success", MAJOR_NUM);
	printk("If you want to talk to the device driver,\n");
	printk("you'll have to create a device file. \n");
	printk("We suggest you use:\n");
	printk("mknod %s c %d 0\n", DEVICE_FILE_NAME, MAJOR_NUM);
	printk("The device file name is important, because\n");
	printk("the ioctl program assumes that's the\n");
	printk("file you'll use.\n");

	return 0;
}

/* 
 * Cleanup - unregister the appropriate file from /proc 
 */
void cleanup_module()
{
	int ret;

	/* 
	 * Unregister the device 
	 */
	ret = unregister_chrdev(MAJOR_NUM, DEVICE_NAME);

	/* 
	 * If there's an error, report it 
	 */
	if (ret < 0)
		printk("Error in module_unregister_chrdev: %d\n", ret);
}

/*
 *  chardev.h - the header file with the ioctl definitions.
 *
 *  The declarations here have to be in a header file, because
 *  they need to be known both to the kernel module
 *  (in chardev.c) and the process calling ioctl (ioctl.c)
 */

#ifndef CHARDEV_H
#define CHARDEV_H

#include <linux/ioctl.h>

/* 
 * The major device number. We can't rely on dynamic 
 * registration any more, because ioctls need to know 
 * it. 
 */
#define MAJOR_NUM 100

/* 
 * Set the message of the device driver 
 */
#define IOCTL_SET_MSG _IOR(MAJOR_NUM, 0, char *)
/*
 * _IOR means that we're creating an ioctl command 
 * number for passing information from a user process
 * to the kernel module. 
 *
 * The first arguments, MAJOR_NUM, is the major device 
 * number we're using.
 *
 * The second argument is the number of the command 
 * (there could be several with different meanings).
 *
 * The third argument is the type we want to get from 
 * the process to the kernel.
 */

/* 
 * Get the message of the device driver 
 */
#define IOCTL_GET_MSG _IOR(MAJOR_NUM, 1, char *)
/* 
 * This IOCTL is used for output, to get the message 
 * of the device driver. However, we still need the 
 * buffer to place the message in to be input, 
 * as it is allocated by the process.
 */

/* 
 * Get the n'th byte of the message 
 */
#define IOCTL_GET_NTH_BYTE _IOWR(MAJOR_NUM, 2, int)
/* 
 * The IOCTL is used for both input and output. It 
 * receives from the user a number, n, and returns 
 * Message[n]. 
 */

/* 
 * The name of the device file 
 */
#define DEVICE_FILE_NAME "char_dev"

#endif

/*
 *  ioctl.c - the process to use ioctl's to control the kernel module
 *
 *  Until now we could have used cat for input and output.  But now
 *  we need to do ioctl's, which require writing our own process.
 */

/* 
 * device specifics, such as ioctl numbers and the
 * major device file. 
 */
#include "chardev.h"

#include <fcntl.h>		/* open */
#include <unistd.h>		/* exit */
#include <sys/ioctl.h>		/* ioctl */

/* 
 * Functions for the ioctl calls 
 */

ioctl_set_msg(int file_desc, char *message)
{
	int ret_val;

	ret_val = ioctl(file_desc, IOCTL_SET_MSG, message);

	if (ret_val < 0) {
		printf("ioctl_set_msg failed:%d\n", ret_val);
		exit(-1);
	}
}

ioctl_get_msg(int file_desc)
{
	int ret_val;
	char message[100];

	/* 
	 * Warning - this is dangerous because we don't tell
	 * the kernel how far it's allowed to write, so it
	 * might overflow the buffer. In a real production
	 * program, we would have used two ioctls - one to tell
	 * the kernel the buffer length and another to give
	 * it the buffer to fill
	 */
	ret_val = ioctl(file_desc, IOCTL_GET_MSG, message);

	if (ret_val < 0) {
		printf("ioctl_get_msg failed:%d\n", ret_val);
		exit(-1);
	}

	printf("get_msg message:%s\n", message);
}

ioctl_get_nth_byte(int file_desc)
{
	int i;
	char c;

	printf("get_nth_byte message:");

	i = 0;
	while (c != 0) {
		c = ioctl(file_desc, IOCTL_GET_NTH_BYTE, i++);

		if (c < 0) {
			printf
			    ("ioctl_get_nth_byte failed at the %d'th byte:\n",
			     i);
			exit(-1);
		}

		putchar(c);
	}
	putchar('\n');
}

/* 
 * Main - Call the ioctl functions 
 */
main()
{
	int file_desc, ret_val;
	char *msg = "Message passed by ioctl\n";

	file_desc = open(DEVICE_FILE_NAME, 0);
	if (file_desc < 0) {
		printf("Can't open device file: %s\n", DEVICE_FILE_NAME);
		exit(-1);
	}

	ioctl_get_nth_byte(file_desc);
	ioctl_get_msg(file_desc);
	ioctl_set_msg(file_desc, msg);

	close(file_desc);
}

8.1. System Calls

So far, the only thing we've done was to use well defined kernel mechanisms to register /proc files and device handlers. This is fine if you want to do something the kernel programmers thought you'd want, such as write a device driver. But what if you want to do something unusual, to change the behavior of the system in some way? Then, you're mostly on your own.

This is where kernel programming gets dangerous. While writing the example below, I killed the open() system call. This meant I couldn't open any files, I couldn't run any programs, and I couldn't shutdown the computer. I had to pull the power switch. Luckily, no files died. To ensure you won't lose any files either, please run sync right before you do the insmod and the rmmod.

Forget about /proc files, forget about device files. They're just minor details. The real process to kernel communication mechanism, the one used by all processes, is system calls. When a process requests a service from the kernel (such as opening a file, forking to a new process, or requesting more memory), this is the mechanism used. If you want to change the behaviour of the kernel in interesting ways, this is the place to do it. By the way, if you want to see which system calls a program uses, run strace <arguments>.

In general, a process is not supposed to be able to access the kernel. It can't access kernel memory and it can't call kernel functions. The hardware of the CPU enforces this (that's the reason why it's called `protected mode').

System calls are an exception to this general rule. What happens is that the process fills the registers with the appropriate values and then calls a special instruction which jumps to a previously defined location in the kernel (of course, that location is readable by user processes, it is not writable by them). Under Intel CPUs, this is done by means of interrupt 0x80. The hardware knows that once you jump to this location, you are no longer running in restricted user mode, but as the operating system kernel --- and therefore you're allowed to do whatever you want.

The location in the kernel a process can jump to is called system_call. The procedure at that location checks the system call number, which tells the kernel what service the process requested. Then, it looks at the table of system calls (sys_call_table) to see the address of the kernel function to call. Then it calls the function, and after it returns, does a few system checks and then return back to the process (or to a different process, if the process time ran out). If you want to read this code, it's at the source file arch/$<$architecture$>$/kernel/entry.S, after the line ENTRY(system_call).

So, if we want to change the way a certain system call works, what we need to do is to write our own function to implement it (usually by adding a bit of our own code, and then calling the original function) and then change the pointer at sys_call_table to point to our function. Because we might be removed later and we don't want to leave the system in an unstable state, it's important for cleanup_module to restore the table to its original state.

The source code here is an example of such a kernel module. We want to `spy' on a certain user, and to printk() a message whenever that user opens a file. Towards this end, we replace the system call to open a file with our own function, called our_sys_open. This function checks the uid (user's id) of the current process, and if it's equal to the uid we spy on, it calls printk() to display the name of the file to be opened. Then, either way, it calls the original open() function with the same parameters, to actually open the file.

The init_module function replaces the appropriate location in sys_call_table and keeps the original pointer in a variable. The cleanup_module function uses that variable to restore everything back to normal. This approach is dangerous, because of the possibility of two kernel modules changing the same system call. Imagine we have two kernel modules, A and B. A's open system call will be A_open and B's will be B_open. Now, when A is inserted into the kernel, the system call is replaced with A_open, which will call the original sys_open when it's done. Next, B is inserted into the kernel, which replaces the system call with B_open, which will call what it thinks is the original system call, A_open, when it's done.

Now, if B is removed first, everything will be well---it will simply restore the system call to A_open, which calls the original. However, if A is removed and then B is removed, the system will crash. A's removal will restore the system call to the original, sys_open, cutting B out of the loop. Then, when B is removed, it will restore the system call to what it thinks is the original, A_open, which is no longer in memory. At first glance, it appears we could solve this particular problem by checking if the system call is equal to our open function and if so not changing it at all (so that B won't change the system call when it's removed), but that will cause an even worse problem. When A is removed, it sees that the system call was changed to B_open so that it is no longer pointing to A_open, so it won't restore it to sys_open before it is removed from memory. Unfortunately, B_open will still try to call A_open which is no longer there, so that even without removing B the system would crash.

Note that all the related problems make syscall stealing unfeasiable for production use. In order to keep people from doing potential harmful things sys_call_table is no longer exported. This means, if you want to do something more than a mere dry run of this example, you will have to patch your current kernel in order to have sys_call_table exported. In the example directory you will find a README and the patch. As you can imagine, such modifications are not to be taken lightly. Do not try this on valueable systems (ie systems that you do not own - or cannot restore easily). It should go O.K., but kernel people, choose to not support hacks like this in 2.6. and I'm sure they've had their reasons for that. For more details, see the README. If in doubt, saying N here and skipping this example is the safe choice.

/*
 *  syscall.c
 *
 *  System call "stealing" sample.
 */

/* 
 * 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, */
#include <linux/moduleparam.h>	/* which will have params */
#include <linux/unistd.h>	/* The list of system calls */

/* 
 * For the current (process) structure, we need
 * this to know who the current user is. 
 */
#include <linux/sched.h>
#include <asm/uaccess.h>

/* 
 * The system call table (a table of functions). We
 * just define this as external, and the kernel will
 * fill it up for us when we are insmod'ed
 *
 * sys_call_table is no longer exported in 2.6.x kernels.
 * If you really want to try this DANGEROUS module you will
 * have to apply the supplied patch against your current kernel
 * and recompile it.
 */
extern void *sys_call_table[];

/* 
 * UID we want to spy on - will be filled from the
 * command line 
 */
static int uid;
module_param(uid, int, 0644);

/* 
 * A pointer to the original system call. The reason
 * we keep this, rather than call the original function
 * (sys_open), is because somebody else might have
 * replaced the system call before us. Note that this
 * is not 100% safe, because if another module
 * replaced sys_open before us, then when we're inserted
 * we'll call the function in that module - and it
 * might be removed before we are.
 *
 * Another reason for this is that we can't get sys_open.
 * It's a static variable, so it is not exported. 
 */
asmlinkage int (*original_call) (const char *, int, int);

/* 
 * The function we'll replace sys_open (the function
 * called when you call the open system call) with. To
 * find the exact prototype, with the number and type
 * of arguments, we find the original function first
 * (it's at fs/open.c).
 *
 * In theory, this means that we're tied to the
 * current version of the kernel. In practice, the
 * system calls almost never change (it would wreck havoc
 * and require programs to be recompiled, since the system
 * calls are the interface between the kernel and the
 * processes).
 */
asmlinkage int our_sys_open(const char *filename, int flags, int mode)
{
	int i = 0;
	char ch;

	/* 
	 * Check if this is the user we're spying on 
	 */
	if (uid == current->uid) {
		/* 
		 * Report the file, if relevant 
		 */
		printk("Opened file by %d: ", uid);
		do {
			get_user(ch, filename + i);
			i++;
			printk("%c", ch);
		} while (ch != 0);
		printk("\n");
	}

	/* 
	 * Call the original sys_open - otherwise, we lose
	 * the ability to open files 
	 */
	return original_call(filename, flags, mode);
}

/* 
 * Initialize the module - replace the system call 
 */
int init_module()
{
	/* 
	 * Warning - too late for it now, but maybe for
	 * next time... 
	 */
	printk("I'm dangerous. I hope you did a ");
	printk("sync before you insmod'ed me.\n");
	printk("My counterpart, cleanup_module(), is even");
	printk("more dangerous. If\n");
	printk("you value your file system, it will ");
	printk("be \"sync; rmmod\" \n");
	printk("when you remove this module.\n");

	/* 
	 * Keep a pointer to the original function in
	 * original_call, and then replace the system call
	 * in the system call table with our_sys_open 
	 */
	original_call = sys_call_table[__NR_open];
	sys_call_table[__NR_open] = our_sys_open;

	/* 
	 * To get the address of the function for system
	 * call foo, go to sys_call_table[__NR_foo]. 
	 */

	printk("Spying on UID:%d\n", uid);

	return 0;
}

/* 
 * Cleanup - unregister the appropriate file from /proc 
 */
void cleanup_module()
{
	/* 
	 * Return the system call back to normal 
	 */
	if (sys_call_table[__NR_open] != our_sys_open) {
		printk("Somebody else also played with the ");
		printk("open system call\n");
		printk("The system may be left in ");
		printk("an unstable state.\n");
	}

	sys_call_table[__NR_open] = original_call;
}

9.1. Blocking Processes

/*
 *  sleep.c - create a /proc file, and if several processes try to open it at
 *  the same time, put all but one to sleep
 */

#include <linux/kernel.h>	/* We're doing kernel work */
#include <linux/module.h>	/* Specifically, a module */
#include <linux/proc_fs.h>	/* Necessary because we use proc fs */
#include <linux/sched.h>	/* For putting processes to sleep and 
				   waking them up */
#include <asm/uaccess.h>	/* for get_user and put_user */

/* 
 * The module's file functions 
 */

/* 
 * Here we keep the last message received, to prove that we can process our
 * input
 */
#define MESSAGE_LENGTH 80
static char Message[MESSAGE_LENGTH];

static struct proc_dir_entry *Our_Proc_File;
#define PROC_ENTRY_FILENAME "sleep"

/* 
 * Since we use the file operations struct, we can't use the special proc
 * output provisions - we have to use a standard read function, which is this
 * function
 */
static ssize_t module_output(struct file *file,	/* see include/linux/fs.h   */
			     char *buf,	/* The buffer to put data to 
					   (in the user segment)    */
			     size_t len,	/* The length of the buffer */
			     loff_t * offset)
{
	static int finished = 0;
	int i;
	char message[MESSAGE_LENGTH + 30];

	/* 
	 * Return 0 to signify end of file - that we have nothing 
	 * more to say at this point.
	 */
	if (finished) {
		finished = 0;
		return 0;
	}

	/* 
	 * If you don't understand this by now, you're hopeless as a kernel
	 * programmer.
	 */
	sprintf(message, "Last input:%s\n", Message);
	for (i = 0; i < len && message[i]; i++)
		put_user(message[i], buf + i);

	finished = 1;
	return i;		/* Return the number of bytes "read" */
}

/* 
 * This function receives input from the user when the user writes to the /proc
 * file.
 */
static ssize_t module_input(struct file *file,	/* The file itself */
			    const char *buf,	/* The buffer with input */
			    size_t length,	/* The buffer's length */
			    loff_t * offset)
{				/* offset to file - ignore */
	int i;

	/* 
	 * Put the input into Message, where module_output will later be 
	 * able to use it
	 */
	for (i = 0; i < MESSAGE_LENGTH - 1 && i < length; i++)
		get_user(Message[i], buf + i);
	/* 
	 * we want a standard, zero terminated string 
	 */
	Message[i] = '\0';

	/* 
	 * We need to return the number of input characters used 
	 */
	return i;
}

/* 
 * 1 if the file is currently open by somebody 
 */
int Already_Open = 0;

/* 
 * Queue of processes who want our file 
 */
DECLARE_WAIT_QUEUE_HEAD(WaitQ);
/* 
 * Called when the /proc file is opened 
 */
static int module_open(struct inode *inode, struct file *file)
{
	/* 
	 * If the file's flags include O_NONBLOCK, it means the process doesn't
	 * want to wait for the file.  In this case, if the file is already 
	 * open, we should fail with -EAGAIN, meaning "you'll have to try 
	 * again", instead of blocking a process which would rather stay awake.
	 */
	if ((file->f_flags & O_NONBLOCK) && Already_Open)
		return -EAGAIN;

	/* 
	 * This is the correct place for try_module_get(THIS_MODULE) because 
	 * if a process is in the loop, which is within the kernel module,
	 * the kernel module must not be removed.
	 */
	try_module_get(THIS_MODULE);

	/* 
	 * If the file is already open, wait until it isn't 
	 */

	while (Already_Open) {
		int i, is_sig = 0;

		/* 
		 * This function puts the current process, including any system
		 * calls, such as us, to sleep.  Execution will be resumed right
		 * after the function call, either because somebody called 
		 * wake_up(&WaitQ) (only module_close does that, when the file 
		 * is closed) or when a signal, such as Ctrl-C, is sent 
		 * to the process
		 */
		wait_event_interruptible(WaitQ, !Already_Open);

		/* 
		 * If we woke up because we got a signal we're not blocking, 
		 * return -EINTR (fail the system call).  This allows processes
		 * to be killed or stopped.
		 */

/*
 * Emmanuel Papirakis:
 *
 * This is a little update to work with 2.2.*.  Signals now are contained in
 * two words (64 bits) and are stored in a structure that contains an array of
 * two unsigned longs.  We now have to make 2 checks in our if.
 *
 * Ori Pomerantz:
 *
 * Nobody promised me they'll never use more than 64 bits, or that this book
 * won't be used for a version of Linux with a word size of 16 bits.  This code
 * would work in any case.
 */
		for (i = 0; i < _NSIG_WORDS && !is_sig; i++)
			is_sig =
			    current->pending.signal.sig[i] & ~current->
			    blocked.sig[i];

		if (is_sig) {
			/* 
			 * It's important to put module_put(THIS_MODULE) here,
			 * because for processes where the open is interrupted
			 * there will never be a corresponding close. If we 
			 * don't decrement the usage count here, we will be 
			 * left with a positive usage count which we'll have no
			 * way to bring down to zero, giving us an immortal 
			 * module, which can only be killed by rebooting 
			 * the machine.
			 */
			module_put(THIS_MODULE);
			return -EINTR;
		}
	}

	/* 
	 * If we got here, Already_Open must be zero 
	 */

	/* 
	 * Open the file 
	 */
	Already_Open = 1;
	return 0;		/* Allow the access */
}

/* 
 * Called when the /proc file is closed 
 */
int module_close(struct inode *inode, struct file *file)
{
	/* 
	 * Set Already_Open to zero, so one of the processes in the WaitQ will
	 * be able to set Already_Open back to one and to open the file. All 
	 * the other processes will be called when Already_Open is back to one,
	 * so they'll go back to sleep.
	 */
	Already_Open = 0;

	/* 
	 * Wake up all the processes in WaitQ, so if anybody is waiting for the
	 * file, they can have it.
	 */
	wake_up(&WaitQ);

	module_put(THIS_MODULE);

	return 0;		/* success */
}

/*
 * This function decides whether to allow an operation (return zero) or not
 * allow it (return a non-zero which indicates why it is not allowed).
 *
 * The operation can be one of the following values:
 * 0 - Execute (run the "file" - meaningless in our case)
 * 2 - Write (input to the kernel module)
 * 4 - Read (output from the kernel module)
 *
 * This is the real function that checks file permissions. The permissions
 * returned by ls -l are for reference only, and can be overridden here.
 */
static int module_permission(struct inode *inode, int op, struct nameidata *nd)
{
	/* 
	 * We allow everybody to read from our module, but only root (uid 0)
	 * may write to it
	 */
	if (op == 4 || (op == 2 && current->euid == 0))
		return 0;

	/* 
	 * If it's anything else, access is denied 
	 */
	return -EACCES;
}

/* 
 * Structures to register as the /proc file, with pointers to all the relevant
 * functions.
 */

/* 
 * File operations for our proc file. This is where we place pointers to all
 * the functions called when somebody tries to do something to our file. NULL
 * means we don't want to deal with something.
 */
static struct file_operations File_Ops_4_Our_Proc_File = {
	.read = module_output,	/* "read" from the file */
	.write = module_input,	/* "write" to the file */
	.open = module_open,	/* called when the /proc file is opened */
	.release = module_close,	/* called when it's closed */
};

/* 
 * Inode operations for our proc file.  We need it so we'll have somewhere to
 * specify the file operations structure we want to use, and the function we
 * use for permissions. It's also possible to specify functions to be called
 * for anything else which could be done to an inode (although we don't bother,
 * we just put NULL).
 */

static struct inode_operations Inode_Ops_4_Our_Proc_File = {
	.permission = module_permission,	/* check for permissions */
};

/* 
 * Module initialization and cleanup 
 */

/* 
 * Initialize the module - register the proc file 
 */

int init_module()
{
	int rv = 0;
	Our_Proc_File = create_proc_entry(PROC_ENTRY_FILENAME, 0644, NULL);
	Our_Proc_File->owner = THIS_MODULE;
	Our_Proc_File->proc_iops = &Inode_Ops_4_Our_Proc_File;
	Our_Proc_File->proc_fops = &File_Ops_4_Our_Proc_File;
	Our_Proc_File->mode = S_IFREG | S_IRUGO | S_IWUSR;
	Our_Proc_File->uid = 0;
	Our_Proc_File->gid = 0;
	Our_Proc_File->size = 80;

	if (Our_Proc_File == NULL) {
		rv = -ENOMEM;
		remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
		printk(KERN_INFO "Error: Could not initialize /proc/test\n");
	}

	return rv;
}

/* 
 * Cleanup - unregister our file from /proc.  This could get dangerous if
 * there are still processes waiting in WaitQ, because they are inside our
 * open function, which will get unloaded. I'll explain how to avoid removal
 * of a kernel module in such a case in chapter 10.
 */
void cleanup_module()
{
	remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
}

10.1. Replacing printk

In Section 1.2.1.2, I said that X and kernel module programming don't mix. That's true for developing kernel modules, but in actual use, you want to be able to send messages to whichever tty[14] the command to load the module came from.

The way this is done is by using current, a pointer to the currently running task, to get the current task's tty structure. Then, we look inside that tty structure to find a pointer to a string write function, which we use to write a string to the tty.

/* 
 *  print_string.c - Send output to the tty we're running on, regardless if it's
 *  through X11, telnet, etc.  We do this by printing the string to the tty
 *  associated with the current task.
 */
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/sched.h>	/* For current */
#include <linux/tty.h>		/* For the tty declarations */
#include <linux/version.h>	/* For LINUX_VERSION_CODE */

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Peter Jay Salzman");

static void print_string(char *str)
{
	struct tty_struct *my_tty;

	/* 
	 * tty struct went into signal struct in 2.6.6 
	 */
#if ( LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,5) )
	/* 
	 * The tty for the current task 
	 */
	my_tty = current->tty;
#else
	/* 
	 * The tty for the current task, for 2.6.6+ kernels 
	 */
	my_tty = current->signal->tty;
#endif

	/* 
	 * If my_tty is NULL, the current task has no tty you can print to 
	 * (ie, if it's a daemon).  If so, there's nothing we can do.
	 */
	if (my_tty != NULL) {

		/* 
		 * my_tty->driver is a struct which holds the tty's functions,
		 * one of which (write) is used to write strings to the tty. 
		 * It can be used to take a string either from the user's or 
		 * kernel's memory segment.
		 *
		 * The function's 1st parameter is the tty to write to,
		 * because the same function would normally be used for all 
		 * tty's of a certain type.  The 2nd parameter controls whether
		 * the function receives a string from kernel memory (false, 0)
		 * or from user memory (true, non zero).  The 3rd parameter is
		 * a pointer to a string.  The 4th parameter is the length of
		 * the string.
		 */
		((my_tty->driver)->write) (my_tty,	/* The tty itself */
					   0,	/* Don't take the string 
						   from user space        */
					   str,	/* String                 */
					   strlen(str));	/* Length */

		/* 
		 * ttys were originally hardware devices, which (usually) 
		 * strictly followed the ASCII standard.  In ASCII, to move to
		 * a new line you need two characters, a carriage return and a
		 * line feed.  On Unix, the ASCII line feed is used for both 
		 * purposes - so we can't just use \n, because it wouldn't have
		 * a carriage return and the next line will start at the
		 * column right after the line feed.
		 *
		 * This is why text files are different between Unix and 
		 * MS Windows.  In CP/M and derivatives, like MS-DOS and 
		 * MS Windows, the ASCII standard was strictly adhered to,
		 * and therefore a newline requirs both a LF and a CR.
		 */
		((my_tty->driver)->write) (my_tty, 0, "\015\012", 2);
	}
}

static int __init print_string_init(void)
{
	print_string("The module has been inserted.  Hello world!");
	return 0;
}

static void __exit print_string_exit(void)
{
	print_string("The module has been removed.  Farewell world!");
}

module_init(print_string_init);
module_exit(print_string_exit);

10.2. Flashing keyboard LEDs

In certain conditions, you may desire for your module a simpler and more direct way to communicate to the external world that he's running. Flashing keyboard LEDs can be such a solution: It is an immediate way to attract attention or to display a status condition. Keyboard LEDs are present on every hardware, they are always visible, they do not need any setup, and their use is rather simple and non-intrusive, if compared to writing to a tty or a file.

The following source code illustrates a minimal kernel module which, when loaded, starts blinking the keyboard LEDs until it is unloaded.

/* 
 *  kbleds.c - Blink keyboard leds until the module is unloaded.
 */

#include <linux/module.h>
#include <linux/config.h>
#include <linux/init.h>
#include <linux/tty.h>		/* For fg_console, MAX_NR_CONSOLES */
#include <linux/kd.h>		/* For KDSETLED */
#include <linux/console_struct.h>	/* For vc_cons */

MODULE_DESCRIPTION("Example module illustrating the use of Keyboard LEDs.");
MODULE_AUTHOR("Daniele Paolo Scarpazza");
MODULE_LICENSE("GPL");

struct timer_list my_timer;
struct tty_driver *my_driver;
char kbledstatus = 0;

#define BLINK_DELAY   HZ/5
#define ALL_LEDS_ON   0x07
#define RESTORE_LEDS  0xFF

/*
 * Function my_timer_func blinks the keyboard LEDs periodically by invoking
 * command KDSETLED of ioctl() on the keyboard driver. To learn more on virtual 
 * terminal ioctl operations, please see file:
 *     /usr/src/linux/drivers/char/vt_ioctl.c, function vt_ioctl().
 *
 * The argument to KDSETLED is alternatively set to 7 (thus causing the led 
 * mode to be set to LED_SHOW_IOCTL, and all the leds are lit) and to 0xFF
 * (any value above 7 switches back the led mode to LED_SHOW_FLAGS, thus
 * the LEDs reflect the actual keyboard status).  To learn more on this, 
 * please see file:
 *     /usr/src/linux/drivers/char/keyboard.c, function setledstate().
 * 
 */

static void my_timer_func(unsigned long ptr)
{
	int *pstatus = (int *)ptr;

	if (*pstatus == ALL_LEDS_ON)
		*pstatus = RESTORE_LEDS;
	else
		*pstatus = ALL_LEDS_ON;

	(my_driver->ioctl) (vc_cons[fg_console].d->vc_tty, NULL, KDSETLED,
			    *pstatus);

	my_timer.expires = jiffies + BLINK_DELAY;
	add_timer(&my_timer);
}

static int __init kbleds_init(void)
{
	int i;

	printk(KERN_INFO "kbleds: loading\n");
	printk(KERN_INFO "kbleds: fgconsole is %x\n", fg_console);
	for (i = 0; i < MAX_NR_CONSOLES; i++) {
		if (!vc_cons[i].d)
			break;
		printk(KERN_INFO "poet_atkm: console[%i/%i] #%i, tty %lx\n", i,
		       MAX_NR_CONSOLES, vc_cons[i].d->vc_num,
		       (unsigned long)vc_cons[i].d->vc_tty);
	}
	printk(KERN_INFO "kbleds: finished scanning consoles\n");

	my_driver = vc_cons[fg_console].d->vc_tty->driver;
	printk(KERN_INFO "kbleds: tty driver magic %x\n", my_driver->magic);

	/*
	 * Set up the LED blink timer the first time
	 */
	init_timer(&my_timer);
	my_timer.function = my_timer_func;
	my_timer.data = (unsigned long)&kbledstatus;
	my_timer.expires = jiffies + BLINK_DELAY;
	add_timer(&my_timer);

	return 0;
}

static void __exit kbleds_cleanup(void)
{
	printk(KERN_INFO "kbleds: unloading...\n");
	del_timer(&my_timer);
	(my_driver->ioctl) (vc_cons[fg_console].d->vc_tty, NULL, KDSETLED,
			    RESTORE_LEDS);
}

module_init(kbleds_init);
module_exit(kbleds_cleanup);

If none of the examples in this chapter fit your debugging needs there might yet be some other tricks to try. Ever wondered what CONFIG_LL_DEBUG in make menuconfig is good for? If you activate that you get low level access to the serial port. While this might not sound very powerful by itself, you can patch kernel/printk.c or any other essential syscall to use printascii, thus makeing it possible to trace virtually everything what your code does over a serial line. If your architecture does not support this and is equipped with a serial port, this might be one of the first things that should be implemented. Logging over a netconsole might also be worth a try.

While you have seen lots of stuff that can be used to aid debugging here, there are some things to be aware of. Debugging is almost always intrusive. Adding debug code can change the situation enough to make the bug seem to dissappear. Thus you should try to keep debug code to a minimum and make shure it does not show up in production code.

11.1. Scheduling Tasks

Very often, we have "housekeeping" tasks which have to be done at a certain time, or every so often. If the task is to be done by a process, we do it by putting it in the crontab file. If the task is to be done by a kernel module, we have two possibilities. The first is to put a process in the crontab file which will wake up the module by a system call when necessary, for example by opening a file. This is terribly inefficient, however -- we run a new process off of crontab, read a new executable to memory, and all this just to wake up a kernel module which is in memory anyway.

Instead of doing that, we can create a function that will be called once for every timer interrupt. The way we do this is we create a task, held in a tq_struct structure, which will hold a pointer to the function. Then, we use queue_task to put that task on a task list called tq_timer, which is the list of tasks to be executed on the next timer interrupt. Because we want the function to keep on being executed, we need to put it back on tq_timer whenever it is called, for the next timer interrupt.

There's one more point we need to remember here. When a module is removed by rmmod, first its reference count is checked. If it is zero, module_cleanup is called. Then, the module is removed from memory with all its functions. Nobody checks to see if the timer's task list happens to contain a pointer to one of those functions, which will no longer be available. Ages later (from the computer's perspective, from a human perspective it's nothing, less than a hundredth of a second), the kernel has a timer interrupt and tries to call the function on the task list. Unfortunately, the function is no longer there. In most cases, the memory page where it sat is unused, and you get an ugly error message. But if some other code is now sitting at the same memory location, things could get very ugly. Unfortunately, we don't have an easy way to unregister a task from a task list.

Since cleanup_module can't return with an error code (it's a void function), the solution is to not let it return at all. Instead, it calls sleep_on or module_sleep_on[15] to put the rmmod process to sleep. Before that, it informs the function called on the timer interrupt to stop attaching itself by setting a global variable. Then, on the next timer interrupt, the rmmod process will be woken up, when our function is no longer in the queue and it's safe to remove the module.

/*
 *  sched.c - scheduale a function to be called on every timer interrupt.
 *
 *  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 */
#include <linux/proc_fs.h>	/* Necessary because we use the proc fs */
#include <linux/workqueue.h>	/* We scheduale tasks here */
#include <linux/sched.h>	/* We need to put ourselves to sleep 
				   and wake up later */
#include <linux/init.h>		/* For __init and __exit */
#include <linux/interrupt.h>	/* For irqreturn_t */

struct proc_dir_entry *Our_Proc_File;
#define PROC_ENTRY_FILENAME "sched"
#define MY_WORK_QUEUE_NAME "WQsched.c"

/* 
 * The number of times the timer interrupt has been called so far 
 */
static int TimerIntrpt = 0;

static void intrpt_routine(void *);

static int die = 0;		/* set this to 1 for shutdown */

/* 
 * The work queue structure for this task, from workqueue.h 
 */
static struct workqueue_struct *my_workqueue;

static struct work_struct Task;
static DECLARE_WORK(Task, intrpt_routine, NULL);

/* 
 * This function will be called on every timer interrupt. Notice the void*
 * pointer - task functions can be used for more than one purpose, each time
 * getting a different parameter.
 */
static void intrpt_routine(void *irrelevant)
{
	/* 
	 * Increment the counter 
	 */
	TimerIntrpt++;

	/* 
	 * If cleanup wants us to die
	 */
	if (die == 0)
		queue_delayed_work(my_workqueue, &Task, 100);
}

/* 
 * Put data into the proc fs file. 
 */
ssize_t
procfile_read(char *buffer,
	      char **buffer_location,
	      off_t offset, int buffer_length, int *eof, void *data)
{
	int len;		/* The number of bytes actually used */

	/* 
	 * It's 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 anybody asks us
	 * if we have more information the answer should always be no.
	 */
	if (offset > 0)
		return 0;

	/* 
	 * Fill the buffer and get its length 
	 */
	len = sprintf(my_buffer, "Timer called %d times so far\n", TimerIntrpt);
	count++;

	/* 
	 * Tell the function which called us where the buffer is 
	 */
	*buffer_location = my_buffer;

	/* 
	 * Return the length 
	 */
	return len;
}

/* 
 * Initialize the module - register the proc file 
 */
int __init init_module()
{
	int rv = 0;
	/* 
	 * Put the task in the work_timer task queue, so it will be executed at
	 * next timer interrupt
	 */
	my_workqueue = create_workqueue(MY_WORK_QUEUE_NAME);
	queue_delayed_work(my_workqueue, &Task, 100);

	Our_Proc_File = create_proc_entry(PROC_ENTRY_FILENAME, 0644, NULL);
	Our_Proc_File->read_proc = procfile_read;
	Our_Proc_File->owner = THIS_MODULE;
	Our_Proc_File->mode = S_IFREG | S_IRUGO;
	Our_Proc_File->uid = 0;
	Our_Proc_File->gid = 0;
	Our_Proc_File->size = 80;

	if (Our_Proc_File == NULL) {
		rv = -ENOMEM;
		remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
		printk(KERN_INFO "Error: Could not initialize /proc/%s\n",
		       PROC_ENTRY_FILENAME);
	}

	return rv;
}

/* 
 * Cleanup
 */
void __exit cleanup_module()
{
	/* 
	 * Unregister our /proc file 
	 */
	remove_proc_entry(PROC_ENTRY_FILENAME, &proc_root);
	printk(KERN_INFO "/proc/%s removed\n", PROC_ENTRY_FILENAME);

	die = 1;		/* keep intrp_routine from queueing itself */
	cancel_delayed_work(&Task);	/* no "new ones" */
	flush_workqueue(my_workqueue);	/* wait till all "old ones" finished */
	destroy_workqueue(my_workqueue);

	/* 
	 * Sleep until intrpt_routine is called one last time. This is 
	 * necessary, because otherwise we'll deallocate the memory holding 
	 * intrpt_routine and Task while work_timer still references them.
	 * Notice that here we don't allow signals to interrupt us.
	 *
	 * Since WaitQ is now not NULL, this automatically tells the interrupt
	 * routine it's time to die.
	 */

}

/* 
 * some work_queue related functions
 * are just available to GPL licensed Modules 
 */
MODULE_LICENSE("GPL");

12.1. Interrupt Handlers

/*
 *  intrpt.c - An interrupt handler.
 *
 *  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 */
#include <linux/sched.h>
#include <linux/workqueue.h>
#include <linux/interrupt.h>	/* We want an interrupt */
#include <asm/io.h>

#define MY_WORK_QUEUE_NAME "WQsched.c"

static struct workqueue_struct *my_workqueue;

/* 
 * This will get called by the kernel as soon as it's safe
 * to do everything normally allowed by kernel modules.
 */
static void got_char(void *scancode)
{
	printk("Scan Code %x %s.\n",
	       (int)*((char *)scancode) & 0x7F,
	       *((char *)scancode) & 0x80 ? "Released" : "Pressed");
}

/* 
 * This function services keyboard interrupts. It reads the relevant
 * information from the keyboard and then puts the non time critical
 * part into the work queue. This will be run when the kernel considers it safe.
 */
irqreturn_t irq_handler(int irq, void *dev_id, struct pt_regs *regs)
{
	/* 
	 * This variables are static because they need to be
	 * accessible (through pointers) to the bottom half routine.
	 */
	static int initialised = 0;
	static unsigned char scancode;
	static struct work_struct task;
	unsigned char status;

	/* 
	 * Read keyboard status
	 */
	status = inb(0x64);
	scancode = inb(0x60);

	if (initialised == 0) {
		INIT_WORK(&task, got_char, &scancode);
		initialised = 1;
	} else {
		PREPARE_WORK(&task, got_char, &scancode);
	}

	queue_work(my_workqueue, &task);

	return IRQ_HANDLED;
}

/* 
 * Initialize the module - register the IRQ handler 
 */
int init_module()
{
	my_workqueue = create_workqueue(MY_WORK_QUEUE_NAME);

	/* 
	 * Since the keyboard handler won't co-exist with another handler,
	 * such as us, we have to disable it (free its IRQ) before we do
	 * anything.  Since we don't know where it is, there's no way to
	 * reinstate it later - so the computer will have to be rebooted
	 * when we're done.
	 */
	free_irq(1, NULL);

	/* 
	 * Request IRQ 1, the keyboard IRQ, to go to our irq_handler.
	 * SA_SHIRQ means we're willing to have othe handlers on this IRQ.
	 * SA_INTERRUPT can be used to make the handler into a fast interrupt.
	 */
	return request_irq(1,	/* The number of the keyboard IRQ on PCs */
			   irq_handler,	/* our handler */
			   SA_SHIRQ, "test_keyboard_irq_handler",
			   (void *)(irq_handler));
}

/* 
 * Cleanup 
 */
void cleanup_module()
{
	/* 
	 * This is only here for completeness. It's totally irrelevant, since
	 * we don't have a way to restore the normal keyboard interrupt so the
	 * computer is completely useless and has to be rebooted.
	 */
	free_irq(1, NULL);
}

/* 
 * some work_queue related functions are just available to GPL licensed Modules
 */
MODULE_LICENSE("GPL");

13.1. Symmetrical Multi-Processing

One of the easiest and cheapest ways to improve hardware performance is to put more than one CPU on the board. This can be done either making the different CPU's take on different jobs (asymmetrical multi-processing) or by making them all run in parallel, doing the same job (symmetrical multi-processing, a.k.a. SMP). Doing asymmetrical multi-processing effectively requires specialized knowledge about the tasks the computer should do, which is unavailable in a general purpose operating system such as Linux. On the other hand, symmetrical multi-processing is relatively easy to implement.

By relatively easy, I mean exactly that: not that it's really easy. In a symmetrical multi-processing environment, the CPU's share the same memory, and as a result code running in one CPU can affect the memory used by another. You can no longer be certain that a variable you've set to a certain value in the previous line still has that value; the other CPU might have played with it while you weren't looking. Obviously, it's impossible to program like this.

In the case of process programming this normally isn't an issue, because a process will normally only run on one CPU at a time[18]. The kernel, on the other hand, could be called by different processes running on different CPU's.

In version 2.0.x, this isn't a problem because the entire kernel is in one big spinlock. This means that if one CPU is in the kernel and another CPU wants to get in, for example because of a system call, it has to wait until the first CPU is done. This makes Linux SMP safe[19], but inefficient.

In version 2.2.x, several CPU's can be in the kernel at the same time. This is something module writers need to be aware of.

14.1. Common Pitfalls

Before I send you on your way to go out into the world and write kernel modules, there are a few things I need to warn you about. If I fail to warn you and something bad happens, please report the problem to me for a full refund of the amount I was paid for your copy of the book.