From 2a1bb161c52c7fef2dde595852d3085e3e059ce5 Mon Sep 17 00:00:00 2001 From: jserv Date: Fri, 6 Aug 2021 10:10:16 +0000 Subject: [PATCH] deploy: 92b646ed64b8f08347f36b1e393431d94e204667 --- index.html | 6417 ++++++++++++++++++++++++++------------------------- lkmpg.css | 3498 ++++++++++++++-------------- lkmpg.html | 6417 ++++++++++++++++++++++++++------------------------- lkmpg0x.svg | 22 +- 4 files changed, 8208 insertions(+), 8146 deletions(-) diff --git a/index.html b/index.html index 81dbc96..8ba660c 100644 --- a/index.html +++ b/index.html @@ -17,7 +17,7 @@

The Linux Kernel Module Programming Guide

Peter Jay Salzman, Michael Burian, Ori Pomerantz, Bob Mottram, Jim Huang

-
August 5, 2021
+
August 6, 2021
@@ -49,51 +49,52 @@
  0.5.5 Code space
  0.5.6 Device Drivers
 0.6 Character Device drivers -
  0.6.1 The proc_ops Structure +
  0.6.1 The file_operations Structure
  0.6.2 The file structure
  0.6.3 Registering A Device
  0.6.4 Unregistering A Device
  0.6.5 chardev.c
  0.6.6 Writing Modules for Multiple Kernel Versions
 0.7 The /proc File System -
  0.7.1 Read and Write a /proc File -
  0.7.2 Manage /proc file with standard filesystem -
  0.7.3 Manage /proc file with seq_file -
 0.8 sysfs: Interacting with your module -
 0.9 Talking To Device Files -
 0.10 System Calls -
 0.11 Blocking Processes and threads -
  0.11.1 Sleep +
  0.7.1 The proc_ops Structure +
  0.7.2 Read and Write a /proc File +
  0.7.3 Manage /proc file with standard filesystem +
  0.7.4 Manage /proc file with seq_file +
 0.8 sysfs: Interacting with your module +
 0.9 Talking To Device Files +
 0.10 System Calls +
 0.11 Blocking Processes and threads -
  0.11.2 Completions -
 0.12 Avoiding Collisions and Deadlocks -
  0.12.1 Mutex -
  0.12.2 Spinlocks -
  0.12.3 Read and write locks -
  0.12.4 Atomic operations -
 0.13 Replacing Print Macros -
  0.13.1 Replacement -
  0.13.2 Flashing keyboard LEDs -
 0.14 Scheduling Tasks -
  0.14.1 Tasklets -
  0.14.2 Work queues -
 0.15 Interrupt Handlers -
  0.15.1 Interrupt Handlers -
  0.15.2 Detecting button presses -
  0.15.3 Bottom Half -
 0.16 Crypto -
  0.16.1 Hash functions -
  0.16.2 Symmetric key encryption -
 0.17 Standardising the interfaces: The Device Model -
 0.18 Optimizations -
  0.18.1 Likely and Unlikely conditions -
 0.19 Common Pitfalls -
  0.19.1 Using standard libraries -
  0.19.2 Disabling interrupts -
  0.19.3 Sticking your head inside a large carnivore -
 0.20 Where To Go From Here? +
  0.11.1 Sleep +
  0.11.2 Completions +
 0.12 Avoiding Collisions and Deadlocks +
  0.12.1 Mutex +
  0.12.2 Spinlocks +
  0.12.3 Read and write locks +
  0.12.4 Atomic operations +
 0.13 Replacing Print Macros +
  0.13.1 Replacement +
  0.13.2 Flashing keyboard LEDs +
 0.14 Scheduling Tasks +
  0.14.1 Tasklets +
  0.14.2 Work queues +
 0.15 Interrupt Handlers +
  0.15.1 Interrupt Handlers +
  0.15.2 Detecting button presses +
  0.15.3 Bottom Half +
 0.16 Crypto +
  0.16.1 Hash functions +
  0.16.2 Symmetric key encryption +
 0.17 Standardising the interfaces: The Device Model +
 0.18 Optimizations +
  0.18.1 Likely and Unlikely conditions +
 0.19 Common Pitfalls +
  0.19.1 Using standard libraries +
  0.19.2 Disabling interrupts +
  0.19.3 Sticking your head inside a large carnivore +
 0.20 Where To Go From Here?

0.1 Introduction

The Linux Kernel Module Programming Guide is a free book; you may reproduce @@ -111,10 +112,10 @@ electronic.

Derivative works and translations of this document must be placed under the Open Software License, and the original copyright notice must remain intact. If you have contributed new material to this book, you must make the material and source -code available for your revisions. Please make revisions and updates available directly +code available for your revisions. Please make revisions and updates available directly to the document maintainer, Peter Jay Salzman <p@dirac.org>. This will allow for the merging of updates and provide consistent revisions to the Linux community. @@ -153,10 +154,10 @@ 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 +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.

@@ -1135,129 +1136,141 @@ discussion can mean something very abstract.

0.6 Character Device drivers

-

0.6.1 The proc_ops Structure

-

The proc_ops structure is defined in /usr/include/linux/fs.h, and holds pointers -to functions defined by the driver that perform various operations on the device. -Each field of the structure corresponds to the address of some function defined by the -driver to handle a requested operation. +

0.6.1 The file_operations Structure

+

The file_operations structure is defined in /usr/include/linux/fs.h, and holds +pointers to functions defined by the driver that perform various operations on the +device. Each field of the structure corresponds to the address of some function +defined by the driver to handle a requested operation.

For example, every character driver needs to define a function that reads from the -device. The proc_ops structure holds the address of the module’s function that -performs that operation. Here is what the definition looks like for kernel -3.0: +device. The file_operations structure holds the address of the module’s function +that performs that operation. Here is what the definition looks like for kernel +5.4:

- 
1struct proc_ops { 
-2    struct module *owner; 
-3    loff_t (*llseek) (struct file *, loff_t, int); 
-4    ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 
-5    ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 
-6    ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t); 
-7    ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t); 
-8    int (*iterate) (struct file *, struct dir_context *); 
-9    unsigned int (*poll) (struct file *, struct poll_table_struct *); 
-10    long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 
-11    long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 
-12    int (*mmap) (struct file *, struct vm_area_struct *); 
-13    int (*open) (struct inode *, struct file *); 
-14    int (*flush) (struct file *, fl_owner_t id); 
-15    int (*release) (struct inode *, struct file *); 
-16    int (*fsync) (struct file *, loff_t, loff_t, int datasync); 
-17    int (*aio_fsync) (struct kiocb *, int datasync); 
-18    int (*fasync) (int, struct file *, int); 
-19    int (*lock) (struct file *, int, struct file_lock *); 
-20    ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); 
-21    unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 
-22    int (*check_flags)(int); 
-23    int (*flock) (struct file *, int, struct file_lock *); 
-24    ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 
-25    ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 
-26    int (*setlease)(struct file *, long, struct file_lock **); 
-27    long (*fallocate)(struct file *file, int mode, loff_t offset, 
-28                    loff_t len); 
-29    int (*show_fdinfo)(struct seq_file *m, struct file *f); 
-30};
-

Some operations are not implemented by a driver. For example, a driver that + 

1struct file_operations { 
+2    struct module *owner; 
+3    loff_t (*llseek) (struct file *, loff_t, int); 
+4    ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 
+5    ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 
+6    ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); 
+7    ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); 
+8    int (*iopoll)(struct kiocb *kiocb, bool spin); 
+9    int (*iterate) (struct file *, struct dir_context *); 
+10    int (*iterate_shared) (struct file *, struct dir_context *); 
+11    __poll_t (*poll) (struct file *, struct poll_table_struct *); 
+12    long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 
+13    long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 
+14    int (*mmap) (struct file *, struct vm_area_struct *); 
+15    unsigned long mmap_supported_flags; 
+16    int (*open) (struct inode *, struct file *); 
+17    int (*flush) (struct file *, fl_owner_t id); 
+18    int (*release) (struct inode *, struct file *); 
+19    int (*fsync) (struct file *, loff_t, loff_t, int datasync); 
+20    int (*fasync) (int, struct file *, int); 
+21    int (*lock) (struct file *, int, struct file_lock *); 
+22    ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); 
+23    unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 
+24    int (*check_flags)(int); 
+25    int (*flock) (struct file *, int, struct file_lock *); 
+26    ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 
+27    ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 
+28    int (*setlease)(struct file *, long, struct file_lock **, void **); 
+29    long (*fallocate)(struct file *file, int mode, loff_t offset, 
+30        loff_t len); 
+31    void (*show_fdinfo)(struct seq_file *m, struct file *f); 
+32    ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, 
+33        loff_t, size_t, unsigned int); 
+34    loff_t (*remap_file_range)(struct file *file_in, loff_t pos_in, 
+35             struct file *file_out, loff_t pos_out, 
+36             loff_t len, unsigned int remap_flags); 
+37    int (*fadvise)(struct file *, loff_t, loff_t, int); 
+38} __randomize_layout;
+

Some operations are not implemented by a driver. For example, a driver that handles a video card will not need to read from a directory structure. The -corresponding entries in the proc_ops structure should be set to NULL. -

There is a gcc extension that makes assigning to this structure more convenient. +corresponding entries in the file_operations structure should be set to +NULL. +

There is a gcc extension that makes assigning to this structure more convenient. You will see it in modern drivers, and may catch you by surprise. This is what the new way of assigning to the structure looks like:

-

- 
1struct proc_ops fops = { 
-2              proc_read: device_read, 
-3              proc_write: device_write, 
-4              proc_open: device_open, 
-5              proc_release: device_release 
-6};
-

However, there is also a C99 way of assigning to elements of a structure, +

+ 
1struct file_operations fops = { 
+2              read: device_read, 
+3              write: device_write, 
+4              open: device_open, 
+5              release: device_release 
+6};
+

However, there is also a C99 way of assigning to elements of a structure, designated initializers, and this is definitely preferred over using the GNU extension. You should use this syntax in case someone wants to port your driver. It will help with compatibility:

- 
1struct proc_ops fops = { 
-2              .proc_read = device_read, 
-3              .proc_write = device_write, 
-4              .proc_open = device_open, 
-5              .proc_release = device_release 
-6};
-

The meaning is clear, and you should be aware that any member of the + 

1struct file_operations fops = { 
+2              .read = device_read, 
+3              .write = device_write, 
+4              .open = device_open, 
+5              .release = device_release 
+6};
+

The meaning is clear, and you should be aware that any member of the structure which you do not explicitly assign will be initialized to NULL by gcc. -

An instance of struct proc_ops containing pointers to functions that are used to -implement read, write, open, … syscalls is commonly named fops. -

+

An instance of struct file_operations containing pointers to functions +that are used to implement read, write, open, … syscalls is commonly named +fops. +

Sin Linux v5.6, the proc_ops structure was introduced to replace the use of the +file_operations structure when registering proc handlers. +

0.6.2 The file structure

-

Each device is represented in the kernel by a file structure, which is defined in +

Each device is represented in the kernel by a file structure, which is defined in linux/fs.h. Be aware that a file is a kernel level structure and never appears in a user space program. It is not the same thing as a FILE, which is defined by glibc and would never appear in a kernel space function. Also, its name is a bit misleading; it represents an abstract open ‘file’, not a file on a disk, which is represented by a structure named inode. -

An instance of struct file is commonly named filp. You’ll also see it refered to as +

An instance of struct file is commonly named filp. You’ll also see it refered to as struct file file. Resist the temptation. -

Go ahead and look at the definition of file. Most of the entries you see, like struct +

Go ahead and look at the definition of file. Most of the entries you see, like struct dentry are not used by device drivers, and you can ignore them. This is because drivers do not fill file directly; they only use structures contained in file which are created elsewhere. -

+

0.6.3 Registering A Device

-

As discussed earlier, char devices are accessed through device files, usually located in +

As discussed earlier, char devices are accessed through device files, usually located in /dev. This is by convention. When writing a driver, it is OK to put the device file in your current directory. Just make sure you place it in /dev for a production driver. The major number tells you which driver handles which device file. The minor number is used only by the driver itself to differentiate -which device it is operating on, just in case the driver handles more than one -device. -

Adding a driver to your system means registering it with the kernel. -This is synonymous with assigning it a major number during the module’s -initialization. You do this by using the register_chrdev function, defined by +which device it is operating on, just in case the driver handles more than one +device. +

Adding a driver to your system means registering it with the kernel. +This is synonymous with assigning it a major number during the module’s +initialization. You do this by using the register_chrdev function, defined by linux/fs.h.

- 
1int register_chrdev(unsigned int major, const char *name, struct proc_ops *fops);
-

where unsigned int major is the major number you want to request, const char + 

1int register_chrdev(unsigned int major, const char *name, struct file_operations *fops);
+

where unsigned int major is the major number you want to request, const char *name is the name of the device as it will appear in /proc/devices and struct -proc_ops *fops is a pointer to the proc_ops table for your driver. A negative return -value means the registration failed. Note that we didn’t pass the minor number to -register_chrdev. That is because the kernel doesn’t care about the minor number; -only our driver uses it. -

Now the question is, how do you get a major number without hijacking one that’s +file_operations *fops is a pointer to the file_operations table for your driver. A +negative return value means the registration failed. Note that we didn’t pass the +minor number to register_chrdev. That is because the kernel doesn’t care about the +minor number; only our driver uses it. +

Now the question is, how do you get a major number without hijacking one that’s already in use? The easiest way would be to look through Documentation /devices.txt and pick an unused one. That is a bad way of doing things because you will never be sure if the number you picked will be assigned later. The answer is that you can ask the kernel to assign you a dynamic major number. -

If you pass a major number of 0 to register_chrdev, the return value will be the +

If you pass a major number of 0 to register_chrdev, the return value will be the dynamically allocated major number. The downside is that you can not make a device file in advance, since you don’t know what the major number will be. There are a couple of ways to do this. First, the driver itself can print the newly assigned @@ -1267,10 +1280,10 @@ or write a shell script to read the file in and make the device file. The third we can have our driver make the the device file using the device_create function after a successful registration and device_destroy during the call to cleanup_module. -

+

0.6.4 Unregistering A Device

-

We can not allow the kernel module to be rmmod’ed whenever root feels like it. If the +

We can not allow the kernel module to be rmmod’ed whenever root feels like it. If the device file is opened by a process and then we remove the kernel module, using the file would cause a call to the memory location where the appropriate function (read/write) used to be. If we are lucky, no other code was loaded there, and we’ll get @@ -1278,15 +1291,15 @@ an ugly error message. If we are unlucky, another kernel module was loaded into same location, which means a jump into the middle of another function within the kernel. The results of this would be impossible to predict, but they can not be very positive. -

Normally, when you don’t want to allow something, you return an error code + + + +

Normally, when you don’t want to allow something, you return an error code (a negative number) from the function which is supposed to do it. With cleanup_module that’s impossible because it is a void function. However, there is a counter which keeps track of how many processes are using your module. You can see what its value is by looking at the 3rd field of /proc/modules. If this number isn’t zero, rmmod will fail. Note that you don’t have to check the counter from within - - - cleanup_module because the check will be performed for you by the system call sys_delete_module, defined in linux/module.c. You should not use this counter directly, but there are functions defined in linux/module.h which let you increase, @@ -1296,19 +1309,19 @@ decrease and display this counter:

  • try_module_get(THIS_MODULE): Increment the use count.
  • module_put(THIS_MODULE): Decrement the use count.
    • -

    It is important to keep the counter accurate; if you ever do lose track of the +

    It is important to keep the counter accurate; if you ever do lose track of the correct usage count, you will never be able to unload the module; it’s now reboot time, boys and girls. This is bound to happen to you sooner or later during a module’s development. -

    +

    0.6.5 chardev.c

    -

    The next code sample creates a char driver named chardev. You can cat its device +

    The next code sample creates a char driver named chardev. You can cat its device file. 

    1cat /proc/devices
    -

    (or open the file with a program) and the driver will put the number of times the +

    (or open the file with a program) and the driver will put the number of times the device file has been read from into the file. We do not support writing to the file (like echo "hi" > /dev/hello), but catch these attempts and tell the user that the operation is not supported. Don’t worry if you don’t see what we @@ -1317,243 +1330,244 @@ simply read in the data and print a message acknowledging that we received it.

    - 
    1/* 
-2 *  chardev.c: Creates a read-only char device that says how many times 
-3 *  you've read from the dev file 
-4 */ 
+   
    1/* 
+2 *  chardev.c: Creates a read-only char device that says how many times 
+3 *  you've read from the dev file 
+4 */ 
 5 
-6#include <linux/cdev.h> 
-7#include <linux/delay.h> 
-8#include <linux/device.h> 
-9#include <linux/fs.h> 
-10#include <linux/init.h> 
-11#include <linux/irq.h> 
-12#include <linux/kernel.h> 
-13#include <linux/module.h> 
-14#include <linux/poll.h> 
+6#include <linux/cdev.h> 
+7#include <linux/delay.h> 
+8#include <linux/device.h> 
+9#include <linux/fs.h> 
+10#include <linux/init.h> 
+11#include <linux/irq.h> 
+12#include <linux/kernel.h> 
+13#include <linux/module.h> 
+14#include <linux/poll.h> 
 15 
-16/* 
-17 *  Prototypes - this would normally go in a .h file 
-18 */ 
-19int init_module(void); 
-20void cleanup_module(void); 
-21static int device_open(struct inode *, struct file *); 
-22static int device_release(struct inode *, struct file *); 
-23static ssize_t device_read(struct file *, char *, size_t, loff_t *); 
-24static ssize_t device_write(struct file *, const char *, size_t, loff_t *); 
+16/* 
+17 *  Prototypes - this would normally go in a .h file 
+18 */ 
+19int init_module(void); 
+20void cleanup_module(void); 
+21static int device_open(struct inode *, struct file *); 
+22static int device_release(struct inode *, struct file *); 
+23static ssize_t device_read(struct file *, char *, size_t, loff_t *); 
+24static ssize_t device_write(struct file *, const char *, size_t, loff_t *); 
 25 
-26#define SUCCESS 0 
-27#define DEVICE_NAME "chardev" /* Dev name as it appears in /proc/devices   */ 
-28#define BUF_LEN 80            /* Max length of the message from the device */ 
+26#define SUCCESS 0 
+27#define DEVICE_NAME "chardev" /* Dev name as it appears in /proc/devices   */ 
+28#define BUF_LEN 80            /* Max length of the message from the device */ 
 29 
-30/* 
-31 * Global variables are declared as static, so are global within the file. 
-32 */ 
+30/* 
+31 * Global variables are declared as static, so are global within the file. 
+32 */ 
 33 
-34static int Major;           /* Major number assigned to our device driver */ 
-35static int Device_Open = 0; /* Is device open? 
-36                             * Used to prevent multiple access to device */ 
-37static char msg[BUF_LEN];   /* The msg the device will give when asked */ 
-38static char *msg_Ptr; 
+34static int Major;           /* Major number assigned to our device driver */ 
+35static int Device_Open = 0; /* Is device open? 
+36                             * Used to prevent multiple access to device */ 
+37static char msg[BUF_LEN];   /* The msg the device will give when asked */ 
+38static char *msg_Ptr; 
 39 
-40static struct class *cls; 
+40static struct class *cls; 
 41 
-42static struct file_operations chardev_fops = {.read = device_read, 
+42static struct file_operations chardev_fops = {.read = device_read, 
 43                                              .write = device_write, 
 44                                              .open = device_open, 
 45                                              .release = device_release}; 
 46 
-47/* 
-48 * This function is called when the module is loaded 
-49 */ 
-50int init_module(void) 
+47/* 
+48 * This function is called when the module is loaded 
+49 */ 
+50int init_module(void) 
 51{ 
 52    Major = register_chrdev(0, DEVICE_NAME, &chardev_fops); 
 53 
-54    if (Major < 0) { 
-55        pr_alert("Registering char device failed with %d\n", Major); 
-56        return Major; 
+54    if (Major < 0) { 
+55        pr_alert("Registering char device failed with %d\n", Major); 
+56        return Major; 
 57    } 
 58 
-59    pr_info("I was assigned major number %d.\n", Major); 
+59    pr_info("I was assigned major number %d.\n", Major); 
 60 
 61    cls = class_create(THIS_MODULE, DEVICE_NAME); 
 62    device_create(cls, NULL, MKDEV(Major, 0), NULL, DEVICE_NAME); 
 63 
-64    pr_info("Device created on /dev/%s\n", DEVICE_NAME); 
+64    pr_info("Device created on /dev/%s\n", DEVICE_NAME); 
 65 
-66    return SUCCESS; 
+66    return SUCCESS; 
 67} 
 68 
-69/* 
-70 * This function is called when the module is unloaded 
-71 */ 
-72void cleanup_module(void) 
+69/* 
+70 * This function is called when the module is unloaded 
+71 */ 
+72void cleanup_module(void) 
 73{ 
 74    device_destroy(cls, MKDEV(Major, 0)); 
 75    class_destroy(cls); 
 76 
-77    /* 
-78     * Unregister the device 
-79     */ 
+77    /* 
+78     * Unregister the device 
+79     */ 
 80    unregister_chrdev(Major, DEVICE_NAME); 
 81} 
 82 
-83/* 
-84 * Methods 
-85 */ 
+83/* 
+84 * Methods 
+85 */ 
 86 
-87/* 
-88 * Called when a process tries to open the device file, like 
-89 * "sudo cat /dev/chardev" 
-90 */ 
-91static int device_open(struct inode *inode, struct file *file) 
+87/* 
+88 * Called when a process tries to open the device file, like 
+89 * "sudo cat /dev/chardev" 
+90 */ 
+91static int device_open(struct inode *inode, struct file *file) 
 92{ 
-93    static int counter = 0; 
+93    static int counter = 0; 
 94 
-95    if (Device_Open) 
-96        return -EBUSY; 
+95    if (Device_Open) 
+96        return -EBUSY; 
 97 
 98    Device_Open++; 
-99    sprintf(msg, "I already told you %d times Hello world!\n", counter++); 
+99    sprintf(msg, "I already told you %d times Hello world!\n", counter++); 
 100    msg_Ptr = msg; 
 101    try_module_get(THIS_MODULE); 
 102 
-103    return SUCCESS; 
+103    return SUCCESS; 
 104} 
 105 
-106/* 
-107 * Called when a process closes the device file. 
-108 */ 
-109static int device_release(struct inode *inode, struct file *file) 
+106/* 
+107 * Called when a process closes the device file. 
+108 */ 
+109static int device_release(struct inode *inode, struct file *file) 
 110{ 
-111    Device_Open--; /* We're now ready for our next caller */ 
+111    Device_Open--; /* We're now ready for our next caller */ 
 112 
-113    /* 
-114     * Decrement the usage count, or else once you opened the file, you'll 
-115     * never get get rid of the module. 
-116     */ 
+113    /* 
+114     * Decrement the usage count, or else once you opened the file, you'll 
+115     * never get get rid of the module. 
+116     */ 
 117    module_put(THIS_MODULE); 
 118 
-119    return SUCCESS; 
+119    return SUCCESS; 
 120} 
 121 
-122/* 
-123 * Called when a process, which already opened the dev file, attempts to 
-124 * read from it. 
-125 */ 
-126static ssize_t device_read(struct file *filp, /* see include/linux/fs.h   */ 
-127                           char *buffer,      /* buffer to fill with data */ 
-128                           size_t length,     /* length of the buffer     */ 
+122/* 
+123 * Called when a process, which already opened the dev file, attempts to 
+124 * read from it. 
+125 */ 
+126static ssize_t device_read(struct file *filp, /* see include/linux/fs.h   */ 
+127                           char *buffer,      /* buffer to fill with data */ 
+128                           size_t length,     /* length of the buffer     */ 
 129                           loff_t *offset) 
 130{ 
-131    /* 
-132     * Number of bytes actually written to the buffer 
-133     */ 
-134    int bytes_read = 0; 
+131    /* 
+132     * Number of bytes actually written to the buffer 
+133     */ 
+134    int bytes_read = 0; 
 135 
-136    /* 
-137     * If we're at the end of the message, 
-138     * return 0 signifying end of file 
-139     */ 
-140    if (*msg_Ptr == 0) 
-141        return 0; 
+136    /* 
+137     * If we're at the end of the message, 
+138     * return 0 signifying end of file 
+139     */ 
+140    if (*msg_Ptr == 0) 
+141        return 0; 
 142 
-143    /* 
-144     * Actually put the data into the buffer 
-145     */ 
-146    while (length && *msg_Ptr) { 
-147        /* 
-148         * The buffer is in the user data segment, not the kernel 
-149         * segment so "*" assignment won't work.  We have to use 
-150         * put_user which copies data from the kernel data segment to 
-151         * the user data segment. 
-152         */ 
+143    /* 
+144     * Actually put the data into the buffer 
+145     */ 
+146    while (length && *msg_Ptr) { 
+147        /* 
+148         * The buffer is in the user data segment, not the kernel 
+149         * segment so "*" assignment won't work.  We have to use 
+150         * put_user which copies data from the kernel data segment to 
+151         * the user data segment. 
+152         */ 
 153        put_user(*(msg_Ptr++), buffer++); 
 154 
 155        length--; 
 156        bytes_read++; 
 157    } 
 158 
-159    /* 
-160     * Most read functions return the number of bytes put into the buffer 
-161     */ 
-162    return bytes_read; 
+159    /* 
+160     * Most read functions return the number of bytes put into the buffer 
+161     */ 
+162    return bytes_read; 
 163} 
 164 
-165/* 
-166 * Called when a process writes to dev file: echo "hi" > /dev/hello 
-167 */ 
-168static ssize_t device_write(struct file *filp, 
-169                            const char *buff, 
-170                            size_t len, 
+165/* 
+166 * Called when a process writes to dev file: echo "hi" > /dev/hello 
+167 */ 
+168static ssize_t device_write(struct file *filp, 
+169                            const char *buff, 
+170                            size_t len, 
 171                            loff_t *off) 
 172{ 
-173    pr_alert("Sorry, this operation isn't supported.\n"); 
-174    return -EINVAL; 
+173    pr_alert("Sorry, this operation isn't supported.\n"); 
+174    return -EINVAL; 
 175} 
 176 
-177MODULE_LICENSE("GPL");
    
+177MODULE_LICENSE("GPL");
    -

    +

    0.6.6 Writing Modules for Multiple Kernel Versions

    -

    The system calls, which are the major interface the kernel shows to the processes, +

    The system calls, which are the major interface the kernel shows to the processes, generally stay the same across versions. A new system call may be added, but usually the old ones will behave exactly like they used to. This is necessary for backward compatibility – a new kernel version is not supposed to break regular processes. In most cases, the device files will also remain the same. On the other hand, the internal interfaces within the kernel can and do change between versions. -

    There are differences between different kernel versions, and if you want to support +

    There are differences between different kernel versions, and if you want to support multiple kernel versions, you will find yourself having to code conditional compilation directives. The way to do this to compare the macro LINUX_VERSION_CODE to the macro KERNEL_VERSION. In version a.b.c of the kernel, the value of this macro would -be 216a + 28b+ c . -

    While previous versions of this guide showed how you can write backward +be 16 8 +2 a + 2 b+ c . +

    While previous versions of this guide showed how you can write backward compatible code with such constructs in great detail, we decided to break with this tradition for the better. People interested in doing such might now use a LKMPG with a version matching to their kernel. -

    +

    0.7 The /proc File System

    -

    In Linux, there is an additional mechanism for the kernel and kernel modules to send +

    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 provides the list of modules and /proc/meminfo which stats memory usage statistics. -

    The method to use the proc file system is very similar to the one used with device +

    The method to use the proc file system is very similar to the one used with device drivers — a structure is created 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. -

    Normal file systems are located on a disk, rather than just in memory (which is +

    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 nowhere 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 +

    Because we don’t get called when the file is opened or closed, there’s nowhere 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. -

    Here a simple example showing how to use a /proc file. This is the HelloWorld +

    Here a simple example showing how to use a /proc file. This is the HelloWorld for the /proc filesystem. There are three parts: create the file /proc/helloworld in the function init_module, return a value (and a buffer) when the file /proc/helloworld is read in the callback function procfile_read, and delete the file /proc/helloworld in the function cleanup_module. -

    The /proc/helloworld is created when the module is loaded with the function +

    The /proc/helloworld is created when the module is loaded with the function proc_create. The return value is a struct proc_dir_entry , and it will be used to configure the file /proc/helloworld (for example, the owner of this file). A null return value means that the creation has failed. -

    Each time, everytime the file /proc/helloworld is read, the function +

    Each time, everytime the file /proc/helloworld is read, the function procfile_read is called. Two parameters of this function are very important: the buffer (the second parameter) and the offset (the fourth one). The content of the buffer will be returned to the application which read it (for example the @@ -1569,97 +1583,111 @@ endlessly. $ cat /proc/helloworld HelloWorld! -

    +

    - 
    1/* 
-2 *  procfs1.c 
-3 */ 
+   
    1/* 
+2 *  procfs1.c 
+3 */ 
 4 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/proc_fs.h> 
-8#include <linux/uaccess.h> 
-9#include <linux/version.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/proc_fs.h> 
+8#include <linux/uaccess.h> 
+9#include <linux/version.h> 
 10 
-11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-12#define HAVE_PROC_OPS 
-13#endif 
+11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+12#define HAVE_PROC_OPS 
+13#endif 
 14 
-15#define procfs_name "helloworld" 
+15#define procfs_name "helloworld" 
 16 
-17struct proc_dir_entry *Our_Proc_File; 
+17struct proc_dir_entry *Our_Proc_File; 
 18 
 19 
-20ssize_t procfile_read(struct file *filePointer, 
-21                      char *buffer, 
-22                      size_t buffer_length, 
+20ssize_t procfile_read(struct file *filePointer, 
+21                      char *buffer, 
+22                      size_t buffer_length, 
 23                      loff_t *offset) 
 24{ 
-25    char s[13] = "HelloWorld!\n"; 
-26    int len = sizeof(s); 
-27    ssize_t ret = len; 
+25    char s[13] = "HelloWorld!\n"; 
+26    int len = sizeof(s); 
+27    ssize_t ret = len; 
 28 
-29    if (*offset >= len || copy_to_user(buffer, s, len)) { 
-30                  pr_info("copy_to_user failed\n"); 
+29    if (*offset >= len || copy_to_user(buffer, s, len)) { 
+30                  pr_info("copy_to_user failed\n"); 
 31                  ret = 0; 
 32    } 
-33    else { 
-34        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
+33    else { 
+34        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
 35        *offset += len; 
 36    } 
 37 
-38    return ret; 
+38    return ret; 
 39} 
 40 
-41#ifdef HAVE_PROC_OPS 
-42static const struct proc_ops proc_file_fops = { 
+41#ifdef HAVE_PROC_OPS 
+42static const struct proc_ops proc_file_fops = { 
 43    .proc_read = procfile_read, 
 44}; 
-45#else 
-46static const struct file_operations proc_file_fops = { 
+45#else 
+46static const struct file_operations proc_file_fops = { 
 47    .read = procfile_read, 
 48}; 
-49#endif 
+49#endif 
 50 
-51int init_module() 
+51int init_module() 
 52{ 
 53    Our_Proc_File = proc_create(procfs_name, 0644, NULL, &proc_file_fops); 
-54    if (NULL == Our_Proc_File) { 
+54    if (NULL == Our_Proc_File) { 
 55        proc_remove(Our_Proc_File); 
-56        pr_alert("Error:Could not initialize /proc/%s\n", procfs_name); 
-57        return -ENOMEM; 
+56        pr_alert("Error:Could not initialize /proc/%s\n", procfs_name); 
+57        return -ENOMEM; 
 58    } 
 59 
-60    pr_info("/proc/%s created\n", procfs_name); 
-61    return 0; 
+60    pr_info("/proc/%s created\n", procfs_name); 
+61    return 0; 
 62} 
 63 
-64void cleanup_module() 
+64void cleanup_module() 
 65{ 
 66    proc_remove(Our_Proc_File); 
-67    pr_info("/proc/%s removed\n", procfs_name); 
+67    pr_info("/proc/%s removed\n", procfs_name); 
 68} 
 69 
-70MODULE_LICENSE("GPL");
    
-
    
+70MODULE_LICENSE("GPL");
    +

    -

    0.7.1 Read and Write a /proc File

    -

    We have seen a very simple example for a /proc file where we only read the +

    0.7.1 The proc_ops Structure

    +

    The proc_ops structure is defined in /usr/include/linux/proc_fs.h in Linux +v5.6+. In older kernels, it used file_operations for custom hooks in /proc file +system, but it contains some members that are unnecessary in VFS, and every time +VFS expands file_operations set, /proc code comes bloated. On the other hand, +not only the space, but also some operations were saved by this structure to improve +its performance. For example, the file which never disappears in /proc can set the +proc_flag as PROC_ENTRY_PERMANENT to save 2 atomic ops, 1 +allocation, 1 free in per open/read/close sequence. +

    +

    +

    0.7.2 Read and Write a /proc File

    +

    We have seen a very simple example for a /proc file where we only read the file /proc/helloworld. It is also possible to write in a /proc file. It works the same way as read, a function is called when the /proc file is written. But there is a little difference with read, data comes from user, so you have to import data from user space to kernel space (with copy_from_user or get_user) -

    The reason for copy_from_user or get_user is that Linux memory (on Intel +

    The reason for copy_from_user or get_user is that Linux memory (on 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 for each of the processes. -

    The only memory segment accessible to a process is its own, so when +

    The only memory segment accessible to a process is its own, so when writing regular programs to run as processes, there is 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 @@ -1672,148 +1700,145 @@ user space, but not for the read function because data is already in kernel space.

    - 
    1/* 
-2 *  procfs2.c -  create a "file" in /proc 
-3 */ 
-4 
-5#include <linux/kernel.h>  /* We're doing kernel work */ 
-6#include <linux/module.h>  /* Specifically, a module */ 
-7#include <linux/proc_fs.h> /* Necessary because we use the proc fs */ 
-8#include <linux/uaccess.h> /* for copy_from_user */ 
-9#include <linux/version.h> 
-10 
-11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-12#define HAVE_PROC_OPS 
-13#endif 
-14 
-15#define PROCFS_MAX_SIZE 1024 
-16#define PROCFS_NAME "buffer1k" 
-17 
-18/** 
-19 * This structure hold information about the /proc file 
-20 * 
-21 */ 
-22static struct proc_dir_entry *Our_Proc_File; 
-23 
-24/** 
-25 * The buffer used to store character for this module 
-26 * 
-27 */ 
-28static char procfs_buffer[PROCFS_MAX_SIZE]; 
-29 
-30/** 
-31 * The size of the buffer 
-32 * 
-33 */ 
-34static unsigned long procfs_buffer_size = 0; 
-35 
-36/** 
-37 * This function is called then the /proc file is read 
-38 * 
-39 */ 
-40ssize_t procfile_read(struct file *filePointer, 
-41                      char *buffer, 
-42                      size_t buffer_length, 
-43                      loff_t *offset) 
-44{ 
-45    char s[13] = "HelloWorld!\n"; 
-46    int len = sizeof(s); 
-47    ssize_t ret = len; 
-48 
-49    if (*offset >= len || copy_to_user(buffer, s, len)) { 
-50                  pr_info("copy_to_user failed\n"); 
-51                  ret = 0; 
-52    } 
-53    else { 
-54        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
-55        *offset += len; 
-56    } 
-57 
-58    return ret; 
-59} 
-60 
-61 
-62/** 
-63 * This function is called with the /proc file is written 
-64 * 
-65 */ 
-66static ssize_t procfile_write(struct file *file, 
-67                              const char *buff, 
-68                              size_t len, 
-69                              loff_t *off) 
-70{ 
-71    procfs_buffer_size = len; 
-72    if (procfs_buffer_size > PROCFS_MAX_SIZE) 
-73        procfs_buffer_size = PROCFS_MAX_SIZE; 
-74 
-75    if (copy_from_user(procfs_buffer, buff, procfs_buffer_size)) 
-76        return -EFAULT; 
-77 
-78    procfs_buffer[procfs_buffer_size] = '\0'; 
-79    return procfs_buffer_size; 
-80} 
-81 
-82#ifdef HAVE_PROC_OPS 
-83static const struct proc_ops proc_file_fops = { 
-84    .proc_read = procfile_read, 
-85    .proc_write = procfile_write, 
-86}; 
-87#else 
-88static const struct file_operations proc_file_fops = { 
-89    .read = procfile_read, 
-90    .write = procfile_write, 
-91}; 
-92#endif 
-93 
-94/** 
-95 *This function is called when the module is loaded 
-96 * 
-97 */ 
-98int init_module() 
-99{ 
-100    Our_Proc_File = proc_create(PROCFS_NAME, 0644, NULL, &proc_file_fops); 
-101    if (NULL == Our_Proc_File) { 
-102        proc_remove(Our_Proc_File); 
-103        pr_alert("Error:Could not initialize /proc/%s\n", PROCFS_NAME); 
-104        return -ENOMEM; 
-105    } 
-106 
-107    pr_info("/proc/%s created\n", PROCFS_NAME); 
-108    return 0; 
-109} 
-110 
-111/** 
-112 *This function is called when the module is unloaded 
-113 * 
-114 */ 
-115void cleanup_module() 
-116{ 
-117    proc_remove(Our_Proc_File); 
-118    pr_info("/proc/%s removed\n", PROCFS_NAME); 
-119} 
-120 
-121MODULE_LICENSE("GPL");
    - - - -

    + 

    1/* 
+2 *  procfs2.c -  create a "file" in /proc 
+3 */ 
+4 
+5#include <linux/kernel.h>  /* We're doing kernel work */ 
+6#include <linux/module.h>  /* Specifically, a module */ 
+7#include <linux/proc_fs.h> /* Necessary because we use the proc fs */ 
+8#include <linux/uaccess.h> /* for copy_from_user */ 
+9#include <linux/version.h> 
+10 
+11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+12#define HAVE_PROC_OPS 
+13#endif 
+14 
+15#define PROCFS_MAX_SIZE 1024 
+16#define PROCFS_NAME "buffer1k" 
+17 
+18/** 
+19 * This structure hold information about the /proc file 
+20 * 
+21 */ 
+22static struct proc_dir_entry *Our_Proc_File; 
+23 
+24/** 
+25 * The buffer used to store character for this module 
+26 * 
+27 */ 
+28static char procfs_buffer[PROCFS_MAX_SIZE]; 
+29 
+30/** 
+31 * The size of the buffer 
+32 * 
+33 */ 
+34static unsigned long procfs_buffer_size = 0; 
+35 
+36/** 
+37 * This function is called then the /proc file is read 
+38 * 
+39 */ 
+40ssize_t procfile_read(struct file *filePointer, 
+41                      char *buffer, 
+42                      size_t buffer_length, 
+43                      loff_t *offset) 
+44{ 
+45    char s[13] = "HelloWorld!\n"; 
+46    int len = sizeof(s); 
+47    ssize_t ret = len; 
+48 
+49    if (*offset >= len || copy_to_user(buffer, s, len)) { 
+50                  pr_info("copy_to_user failed\n"); 
+51                  ret = 0; 
+52    } 
+53    else { 
+54        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
+55        *offset += len; 
+56    } 
+57 
+58    return ret; 
+59} 
+60 
+61 
+62/** 
+63 * This function is called with the /proc file is written 
+64 * 
+65 */ 
+66static ssize_t procfile_write(struct file *file, 
+67                              const char *buff, 
+68                              size_t len, 
+69                              loff_t *off) 
+70{ 
+71    procfs_buffer_size = len; 
+72    if (procfs_buffer_size > PROCFS_MAX_SIZE) 
+73        procfs_buffer_size = PROCFS_MAX_SIZE; 
+74 
+75    if (copy_from_user(procfs_buffer, buff, procfs_buffer_size)) 
+76        return -EFAULT; 
+77 
+78    procfs_buffer[procfs_buffer_size] = '\0'; 
+79    return procfs_buffer_size; 
+80} 
+81 
+82#ifdef HAVE_PROC_OPS 
+83static const struct proc_ops proc_file_fops = { 
+84    .proc_read = procfile_read, 
+85    .proc_write = procfile_write, 
+86}; 
+87#else 
+88static const struct file_operations proc_file_fops = { 
+89    .read = procfile_read, 
+90    .write = procfile_write, 
+91}; 
+92#endif 
+93 
+94/** 
+95 *This function is called when the module is loaded 
+96 * 
+97 */ 
+98int init_module() 
+99{ 
+100    Our_Proc_File = proc_create(PROCFS_NAME, 0644, NULL, &proc_file_fops); 
+101    if (NULL == Our_Proc_File) { 
+102        proc_remove(Our_Proc_File); 
+103        pr_alert("Error:Could not initialize /proc/%s\n", PROCFS_NAME); 
+104        return -ENOMEM; 
+105    } 
+106 
+107    pr_info("/proc/%s created\n", PROCFS_NAME); 
+108    return 0; 
+109} 
+110 
+111/** 
+112 *This function is called when the module is unloaded 
+113 * 
+114 */ 
+115void cleanup_module() 
+116{ 
+117    proc_remove(Our_Proc_File); 
+118    pr_info("/proc/%s removed\n", PROCFS_NAME); 
+119} 
+120 
+121MODULE_LICENSE("GPL");
    +

    -

    0.7.2 Manage /proc file with standard filesystem

    -

    We have seen how to read and write a /proc file with the /proc interface. But it is +

    0.7.3 Manage /proc file with standard filesystem

    +

    We have seen how to read and write a /proc file with the /proc interface. But it is also possible to manage /proc file with inodes. The main concern is to use advanced functions, like permissions. -

    In Linux, there is a standard mechanism for file system registration. Since every +

    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, there is a special structure to hold pointers to all those functions, struct inode_operations, which includes a pointer to struct proc_ops. -

    The difference between file and inode operations is that file operations deal with +

    The difference between file and inode operations is that file operations deal with the file itself whereas inode operations deal with ways of referencing the file, such as creating links to it. -

    In /proc, whenever we register a new file, we’re allowed to specify which struct +

    In /proc, whenever we register a new file, we’re allowed to specify which struct inode_operations will be used to access to it. This is the mechanism we use, a struct inode_operations which includes a pointer to a struct proc_ops which includes pointers to our procfs_read and procfs_write functions. -

    Another interesting point here is the module_permission function. This function +

    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 @@ -1821,139 +1846,139 @@ 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. -

    It is important to note that the standard roles of read and write are reversed in +

    It is 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. -

    -

    - 
    1/* 
-2 *  procfs3.c 
-3 */ 
-4 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/proc_fs.h> 
-8#include <linux/sched.h> 
-9#include <linux/uaccess.h> 
-10#include <linux/version.h> 
-11 
-12#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-13#define HAVE_PROC_OPS 
-14#endif 
-15 
-16#define PROCFS_MAX_SIZE 2048 
-17#define PROCFS_ENTRY_FILENAME "buffer2k" 
-18 
-19struct proc_dir_entry *Our_Proc_File; 
-20static char procfs_buffer[PROCFS_MAX_SIZE]; 
-21static unsigned long procfs_buffer_size = 0; 
-22 
-23static ssize_t procfs_read(struct file *filp, 
-24                           char *buffer, 
-25                           size_t length, 
-26                           loff_t *offset) 
-27{ 
-28    static int finished = 0; 
-29    if (finished) { 
-30        pr_debug("procfs_read: END\n"); 
-31        finished = 0; 
-32        return 0; 
-33    } 
-34    finished = 1; 
-35    if (copy_to_user(buffer, procfs_buffer, procfs_buffer_size)) 
-36        return -EFAULT; 
-37    pr_debug("procfs_read: read %lu bytes\n", procfs_buffer_size); 
-38    return procfs_buffer_size; 
-39} 
-40static ssize_t procfs_write(struct file *file, 
-41                            const char *buffer, 
-42                            size_t len, 
-43                            loff_t *off) 
-44{ 
-45    if (len > PROCFS_MAX_SIZE) 
-46        procfs_buffer_size = PROCFS_MAX_SIZE; 
-47    else 
-48        procfs_buffer_size = len; 
-49    if (copy_from_user(procfs_buffer, buffer, procfs_buffer_size)) 
-50        return -EFAULT; 
-51    pr_debug("procfs_write: write %lu bytes\n", procfs_buffer_size); 
-52    return procfs_buffer_size; 
-53} 
-54int procfs_open(struct inode *inode, struct file *file) 
-55{ 
-56    try_module_get(THIS_MODULE); 
-57    return 0; 
-58} 
-59int procfs_close(struct inode *inode, struct file *file) 
-60{ 
-61    module_put(THIS_MODULE); 
-62    return 0; 
-63} 
-64 
-65#ifdef HAVE_PROC_OPS 
-66static struct proc_ops File_Ops_4_Our_Proc_File = { 
-67    .proc_read = procfs_read, 
-68    .proc_write = procfs_write, 
-69    .proc_open = procfs_open, 
-70    .proc_release = procfs_close, 
-71}; 
-72#else 
-73static const struct file_operations File_Ops_4_Our_Proc_File = { 
-74    .read = procfs_read, 
-75    .write = procfs_write, 
-76    .open = procfs_open, 
-77    .release = procfs_close, 
-78}; 
-79#endif 
-80 
-81int init_module() 
-82{ 
-83    Our_Proc_File = proc_create(PROCFS_ENTRY_FILENAME, 0644, NULL, 
-84                                &File_Ops_4_Our_Proc_File); 
-85    if (Our_Proc_File == NULL) { 
-86        remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
-87        pr_debug("Error: Could not initialize /proc/%s\n", 
-88                 PROCFS_ENTRY_FILENAME); 
-89        return -ENOMEM; 
-90    } 
-91    proc_set_size(Our_Proc_File, 80); 
-92    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
-93 
-94    pr_debug("/proc/%s created\n", PROCFS_ENTRY_FILENAME); 
-95    return 0; 
-96} 
-97void cleanup_module() 
-98{ 
-99    remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
-100    pr_debug("/proc/%s removed\n", PROCFS_ENTRY_FILENAME); 
-101} 
-102 
-103MODULE_LICENSE("GPL");
    -

    Still hungry for procfs examples? Well, first of all keep in mind, there are rumors -around, claiming that procfs is on its way out, consider using sysfs instead. Consider -using this mechanism, in case you want to document something kernel related -yourself. -

    -

    -

    0.7.3 Manage /proc file with seq_file

    -

    As we have seen, writing a /proc file may be quite “complex”. So to help people -writting /proc file, there is an API named seq_file that helps formating a /proc file +input. +

    +

    + 
    1/* 
+2 *  procfs3.c 
+3 */ 
+4 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/proc_fs.h> 
+8#include <linux/sched.h> 
+9#include <linux/uaccess.h> 
+10#include <linux/version.h> 
+11 
+12#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+13#define HAVE_PROC_OPS 
+14#endif 
+15 
+16#define PROCFS_MAX_SIZE 2048 
+17#define PROCFS_ENTRY_FILENAME "buffer2k" 
+18 
+19struct proc_dir_entry *Our_Proc_File; 
+20static char procfs_buffer[PROCFS_MAX_SIZE]; 
+21static unsigned long procfs_buffer_size = 0; 
+22 
+23static ssize_t procfs_read(struct file *filp, 
+24                           char *buffer, 
+25                           size_t length, 
+26                           loff_t *offset) 
+27{ 
+28    static int finished = 0; 
+29    if (finished) { 
+30        pr_debug("procfs_read: END\n"); 
+31        finished = 0; 
+32        return 0; 
+33    } 
+34    finished = 1; 
+35    if (copy_to_user(buffer, procfs_buffer, procfs_buffer_size)) 
+36        return -EFAULT; 
+37    pr_debug("procfs_read: read %lu bytes\n", procfs_buffer_size); 
+38    return procfs_buffer_size; 
+39} 
+40static ssize_t procfs_write(struct file *file, 
+41                            const char *buffer, 
+42                            size_t len, 
+43                            loff_t *off) 
+44{ 
+45    if (len > PROCFS_MAX_SIZE) 
+46        procfs_buffer_size = PROCFS_MAX_SIZE; 
+47    else 
+48        procfs_buffer_size = len; 
+49    if (copy_from_user(procfs_buffer, buffer, procfs_buffer_size)) 
+50        return -EFAULT; 
+51    pr_debug("procfs_write: write %lu bytes\n", procfs_buffer_size); 
+52    return procfs_buffer_size; 
+53} 
+54int procfs_open(struct inode *inode, struct file *file) 
+55{ 
+56    try_module_get(THIS_MODULE); 
+57    return 0; 
+58} 
+59int procfs_close(struct inode *inode, struct file *file) 
+60{ 
+61    module_put(THIS_MODULE); 
+62    return 0; 
+63} 
+64 
+65#ifdef HAVE_PROC_OPS 
+66static struct proc_ops File_Ops_4_Our_Proc_File = { 
+67    .proc_read = procfs_read, 
+68    .proc_write = procfs_write, 
+69    .proc_open = procfs_open, 
+70    .proc_release = procfs_close, 
+71}; 
+72#else 
+73static const struct file_operations File_Ops_4_Our_Proc_File = { 
+74    .read = procfs_read, 
+75    .write = procfs_write, 
+76    .open = procfs_open, 
+77    .release = procfs_close, 
+78}; 
+79#endif 
+80 
+81int init_module() 
+82{ 
+83    Our_Proc_File = proc_create(PROCFS_ENTRY_FILENAME, 0644, NULL, 
+84                                &File_Ops_4_Our_Proc_File); 
+85    if (Our_Proc_File == NULL) { 
+86        remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
+87        pr_debug("Error: Could not initialize /proc/%s\n", 
+88                 PROCFS_ENTRY_FILENAME); 
+89        return -ENOMEM; 
+90    } 
+91    proc_set_size(Our_Proc_File, 80); 
+92    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
+93 
+94    pr_debug("/proc/%s created\n", PROCFS_ENTRY_FILENAME); 
+95    return 0; 
+96} 
+97void cleanup_module() 
+98{ 
+99    remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
+100    pr_debug("/proc/%s removed\n", PROCFS_ENTRY_FILENAME); 
+101} 
+102 
+103MODULE_LICENSE("GPL");
    +

    Still hungry for procfs examples? Well, first of all keep in mind, there are rumors +around, claiming that procfs is on its way out, consider using sysfs instead. Consider +using this mechanism, in case you want to document something kernel related +yourself. +

    +

    +

    0.7.4 Manage /proc file with seq_file

    +

    As we have seen, writing a /proc file may be quite “complex”. So to help people +writting /proc file, there is an API named seq_file that helps formating a /proc file for output. It is based on sequence, which is composed of 3 functions: start(), next(), and stop(). The seq_file API starts a sequence when a user read the /proc file. -

    A sequence begins with the call of the function start(). If the return is a non +

    A sequence begins with the call of the function start(). If the return is a non NULL value, the function next() is called. This function is an iterator, the goal is to go through all the data. Each time next() is called, the function show() is also called. It writes data values in the buffer read by the user. The function next() is called until it returns NULL. The sequence ends when next() returns NULL, then the function stop() is called. -

    BE CAREFUL: when a sequence is finished, another one starts. That means that +

    BE CAREFUL: when a sequence is finished, another one starts. That means that at the end of function stop(), the function start() is called again. This loop finishes when the function start() returns NULL. You can see a scheme of this in the Figure 1. @@ -1963,282 +1988,282 @@ Figure 1

    srYrsNNYtaeenetoooertusetupstrxr((ntn))( tis)istrr teeaNreNatUaUtmLtLmeLmLen?e?ntntt +

    srYrsNNYtaeenetoooertusetupstrxr((ntn))( tis)istrr teeaNreNatUaUtmLtLmeLmLen?e?ntntt

    -
    Figure 1:How seq_file works
    +
    Figure 1:How seq_file works
    -

    The seq_file provides basic functions for proc_ops, such as seq_read, +

    The seq_file provides basic functions for proc_ops, such as seq_read, seq_lseek, and some others. But nothing to write in the /proc file. Of course, you can still use the same way as in the previous example.

    - 
    1/* 
-2 *  procfs4.c -  create a "file" in /proc 
-3 *  This program uses the seq_file library to manage the /proc file. 
-4 * 
-5 */ 
-6 
-7#include <linux/kernel.h>   /* We're doing kernel work */ 
-8#include <linux/module.h>   /* Specifically, a module */ 
-9#include <linux/proc_fs.h>  /* Necessary because we use proc fs */ 
-10#include <linux/seq_file.h> /* for seq_file */ 
-11#include <linux/version.h> 
-12 
-13#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-14#define HAVE_PROC_OPS 
-15#endif 
-16 
-17#define PROC_NAME "iter" 
-18 
-19MODULE_LICENSE("GPL"); 
-20 
-21/** 
-22 * This function is called at the beginning of a sequence. 
-23 * ie, when: 
-24 *      - the /proc file is read (first time) 
-25 *      - after the function stop (end of sequence) 
-26 * 
-27 */ 
-28static void *my_seq_start(struct seq_file *s, loff_t *pos) 
-29{ 
-30    static unsigned long counter = 0; 
-31 
-32    /* beginning a new sequence ? */ 
-33    if (*pos == 0) { 
-34        /* yes => return a non null value to begin the sequence */ 
-35        return &counter; 
-36    } else { 
-37        /* no => it's the end of the sequence, return end to stop reading */ 
-38        *pos = 0; 
-39        return NULL; 
-40    } 
-41} 
-42 
-43/** 
-44 * This function is called after the beginning of a sequence. 
-45 * It's called untill the return is NULL (this ends the sequence). 
-46 * 
-47 */ 
-48static void *my_seq_next(struct seq_file *s, void *v, loff_t *pos) 
-49{ 
-50    unsigned long *tmp_v = (unsigned long *) v; 
-51    (*tmp_v)++; 
-52    (*pos)++; 
-53    return NULL; 
-54} 
-55 
-56/** 
-57 * This function is called at the end of a sequence 
-58 * 
-59 */ 
-60static void my_seq_stop(struct seq_file *s, void *v) 
-61{ 
-62    /* nothing to do, we use a static value in start() */ 
-63} 
-64 
-65/** 
-66 * This function is called for each "step" of a sequence 
-67 * 
-68 */ 
-69static int my_seq_show(struct seq_file *s, void *v) 
-70{ 
-71    loff_t *spos = (loff_t *) v; 
-72 
-73    seq_printf(s, "%Ld\n", *spos); 
-74    return 0; 
-75} 
-76 
-77/** 
-78 * This structure gather "function" to manage the sequence 
-79 * 
-80 */ 
-81static struct seq_operations my_seq_ops = {.start = my_seq_start, 
-82                                           .next = my_seq_next, 
-83                                           .stop = my_seq_stop, 
-84                                           .show = my_seq_show}; 
-85 
-86/** 
-87 * This function is called when the /proc file is open. 
-88 * 
-89 */ 
-90static int my_open(struct inode *inode, struct file *file) 
-91{ 
-92    return seq_open(file, &my_seq_ops); 
-93}; 
-94 
-95/** 
-96 * This structure gather "function" that manage the /proc file 
-97 * 
-98 */ 
-99#ifdef HAVE_PROC_OPS 
-100static const struct proc_ops my_file_ops = { 
-101    .proc_open = my_open, 
-102    .proc_read = seq_read, 
-103    .proc_lseek = seq_lseek, 
-104    .proc_release = seq_release, 
-105}; 
-106#else 
-107static const struct file_operations my_file_ops = { 
-108    .open = my_open, 
-109    .read = seq_read, 
-110    .llseek = seq_lseek, 
-111    .release = seq_release, 
-112}; 
-113#endif 
-114 
-115/** 
-116 * This function is called when the module is loaded 
-117 * 
-118 */ 
-119int init_module(void) 
-120{ 
-121    struct proc_dir_entry *entry; 
-122 
-123    entry = proc_create(PROC_NAME, 0, NULL, &my_file_ops); 
-124    if (entry == NULL) { 
-125        remove_proc_entry(PROC_NAME, NULL); 
-126        pr_debug("Error: Could not initialize /proc/%s\n", PROC_NAME); 
-127        return -ENOMEM; 
-128    } 
-129 
-130    return 0; 
-131} 
-132 
-133/** 
-134 * This function is called when the module is unloaded. 
-135 * 
-136 */ 
-137void cleanup_module(void) 
-138{ 
-139    remove_proc_entry(PROC_NAME, NULL); 
-140    pr_debug("/proc/%s removed\n", PROC_NAME); 
-141}
    -

    If you want more information, you can read this web page: + 

    1/* 
+2 *  procfs4.c -  create a "file" in /proc 
+3 *  This program uses the seq_file library to manage the /proc file. 
+4 * 
+5 */ 
+6 
+7#include <linux/kernel.h>   /* We're doing kernel work */ 
+8#include <linux/module.h>   /* Specifically, a module */ 
+9#include <linux/proc_fs.h>  /* Necessary because we use proc fs */ 
+10#include <linux/seq_file.h> /* for seq_file */ 
+11#include <linux/version.h> 
+12 
+13#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+14#define HAVE_PROC_OPS 
+15#endif 
+16 
+17#define PROC_NAME "iter" 
+18 
+19MODULE_LICENSE("GPL"); 
+20 
+21/** 
+22 * This function is called at the beginning of a sequence. 
+23 * ie, when: 
+24 *      - the /proc file is read (first time) 
+25 *      - after the function stop (end of sequence) 
+26 * 
+27 */ 
+28static void *my_seq_start(struct seq_file *s, loff_t *pos) 
+29{ 
+30    static unsigned long counter = 0; 
+31 
+32    /* beginning a new sequence ? */ 
+33    if (*pos == 0) { 
+34        /* yes => return a non null value to begin the sequence */ 
+35        return &counter; 
+36    } else { 
+37        /* no => it's the end of the sequence, return end to stop reading */ 
+38        *pos = 0; 
+39        return NULL; 
+40    } 
+41} 
+42 
+43/** 
+44 * This function is called after the beginning of a sequence. 
+45 * It's called untill the return is NULL (this ends the sequence). 
+46 * 
+47 */ 
+48static void *my_seq_next(struct seq_file *s, void *v, loff_t *pos) 
+49{ 
+50    unsigned long *tmp_v = (unsigned long *) v; 
+51    (*tmp_v)++; 
+52    (*pos)++; 
+53    return NULL; 
+54} 
+55 
+56/** 
+57 * This function is called at the end of a sequence 
+58 * 
+59 */ 
+60static void my_seq_stop(struct seq_file *s, void *v) 
+61{ 
+62    /* nothing to do, we use a static value in start() */ 
+63} 
+64 
+65/** 
+66 * This function is called for each "step" of a sequence 
+67 * 
+68 */ 
+69static int my_seq_show(struct seq_file *s, void *v) 
+70{ 
+71    loff_t *spos = (loff_t *) v; 
+72 
+73    seq_printf(s, "%Ld\n", *spos); 
+74    return 0; 
+75} 
+76 
+77/** 
+78 * This structure gather "function" to manage the sequence 
+79 * 
+80 */ 
+81static struct seq_operations my_seq_ops = {.start = my_seq_start, 
+82                                           .next = my_seq_next, 
+83                                           .stop = my_seq_stop, 
+84                                           .show = my_seq_show}; 
+85 
+86/** 
+87 * This function is called when the /proc file is open. 
+88 * 
+89 */ 
+90static int my_open(struct inode *inode, struct file *file) 
+91{ 
+92    return seq_open(file, &my_seq_ops); 
+93}; 
+94 
+95/** 
+96 * This structure gather "function" that manage the /proc file 
+97 * 
+98 */ 
+99#ifdef HAVE_PROC_OPS 
+100static const struct proc_ops my_file_ops = { 
+101    .proc_open = my_open, 
+102    .proc_read = seq_read, 
+103    .proc_lseek = seq_lseek, 
+104    .proc_release = seq_release, 
+105}; 
+106#else 
+107static const struct file_operations my_file_ops = { 
+108    .open = my_open, 
+109    .read = seq_read, 
+110    .llseek = seq_lseek, 
+111    .release = seq_release, 
+112}; 
+113#endif 
+114 
+115/** 
+116 * This function is called when the module is loaded 
+117 * 
+118 */ 
+119int init_module(void) 
+120{ 
+121    struct proc_dir_entry *entry; 
+122 
+123    entry = proc_create(PROC_NAME, 0, NULL, &my_file_ops); 
+124    if (entry == NULL) { 
+125        remove_proc_entry(PROC_NAME, NULL); 
+126        pr_debug("Error: Could not initialize /proc/%s\n", PROC_NAME); 
+127        return -ENOMEM; 
+128    } 
+129 
+130    return 0; 
+131} 
+132 
+133/** 
+134 * This function is called when the module is unloaded. 
+135 * 
+136 */ 
+137void cleanup_module(void) 
+138{ 
+139    remove_proc_entry(PROC_NAME, NULL); 
+140    pr_debug("/proc/%s removed\n", PROC_NAME); 
+141}
    +

    If you want more information, you can read this web page:

    -

    You can also read the code of fs/seq_file.c in the linux kernel. +

    You can also read the code of fs/seq_file.c in the linux kernel.

    -

    0.8 sysfs: Interacting with your module

    -

    sysfs allows you to interact with the running kernel from userspace by reading or +

    0.8 sysfs: Interacting with your module

    +

    sysfs allows you to interact with the running kernel from userspace by reading or setting variables inside of modules. This can be useful for debugging purposes, or just as an interface for applications or scripts. You can find sysfs directories and files under the sys directory on your system.

    - 
    1ls -l /sys
    -

    An example of a hello world module which includes the creation of a variable + 

    1ls -l /sys
    +

    An example of a hello world module which includes the creation of a variable accessible via sysfs is given below.

    - 
    1/* 
-2 *  hello-sysfs.c sysfs example 
-3 */ 
-4#include <linux/fs.h> 
-5#include <linux/init.h> 
-6#include <linux/kobject.h> 
-7#include <linux/module.h> 
-8#include <linux/string.h> 
-9#include <linux/sysfs.h> 
-10 
-11MODULE_LICENSE("GPL"); 
-12 
-13static struct kobject *mymodule; 
-14 
-15/* the variable you want to be able to change */ 
-16static int myvariable = 0; 
-17 
-18static ssize_t myvariable_show(struct kobject *kobj, 
-19                               struct kobj_attribute *attr, 
-20                               char *buf) 
-21{ 
-22    return sprintf(buf, "%d\n", myvariable); 
-23} 
-24 
-25static ssize_t myvariable_store(struct kobject *kobj, 
-26                                struct kobj_attribute *attr, 
-27                                char *buf, 
-28                                size_t count) 
-29{ 
-30    sscanf(buf, "%du", &myvariable); 
-31    return count; 
-32} 
-33 
-34 
-35static struct kobj_attribute myvariable_attribute = 
-36    __ATTR(myvariable, 0660, myvariable_show, (void *) myvariable_store); 
-37 
-38static int __init mymodule_init(void) 
-39{ 
-40    int error = 0; 
-41 
-42    pr_info("mymodule: initialised\n"); 
-43 
-44    mymodule = kobject_create_and_add("mymodule", kernel_kobj); 
-45    if (!mymodule) 
-46        return -ENOMEM; 
-47 
-48    error = sysfs_create_file(mymodule, &myvariable_attribute.attr); 
-49    if (error) { 
-50        pr_info( 
-51            "failed to create the myvariable file " 
-52            "in /sys/kernel/mymodule\n"); 
-53    } 
-54 
-55    return error; 
-56} 
-57 
-58static void __exit mymodule_exit(void) 
-59{ 
-60    pr_info("mymodule: Exit success\n"); 
-61    kobject_put(mymodule); 
-62} 
-63 
-64module_init(mymodule_init); 
-65module_exit(mymodule_exit);
    -

    Make and install the module: + 

    1/* 
+2 *  hello-sysfs.c sysfs example 
+3 */ 
+4#include <linux/fs.h> 
+5#include <linux/init.h> 
+6#include <linux/kobject.h> 
+7#include <linux/module.h> 
+8#include <linux/string.h> 
+9#include <linux/sysfs.h> 
+10 
+11MODULE_LICENSE("GPL"); 
+12 
+13static struct kobject *mymodule; 
+14 
+15/* the variable you want to be able to change */ 
+16static int myvariable = 0; 
+17 
+18static ssize_t myvariable_show(struct kobject *kobj, 
+19                               struct kobj_attribute *attr, 
+20                               char *buf) 
+21{ 
+22    return sprintf(buf, "%d\n", myvariable); 
+23} 
+24 
+25static ssize_t myvariable_store(struct kobject *kobj, 
+26                                struct kobj_attribute *attr, 
+27                                char *buf, 
+28                                size_t count) 
+29{ 
+30    sscanf(buf, "%du", &myvariable); 
+31    return count; 
+32} 
+33 
+34 
+35static struct kobj_attribute myvariable_attribute = 
+36    __ATTR(myvariable, 0660, myvariable_show, (void *) myvariable_store); 
+37 
+38static int __init mymodule_init(void) 
+39{ 
+40    int error = 0; 
+41 
+42    pr_info("mymodule: initialised\n"); 
+43 
+44    mymodule = kobject_create_and_add("mymodule", kernel_kobj); 
+45    if (!mymodule) 
+46        return -ENOMEM; 
+47 
+48    error = sysfs_create_file(mymodule, &myvariable_attribute.attr); 
+49    if (error) { 
+50        pr_info( 
+51            "failed to create the myvariable file " 
+52            "in /sys/kernel/mymodule\n"); 
+53    } 
+54 
+55    return error; 
+56} 
+57 
+58static void __exit mymodule_exit(void) 
+59{ 
+60    pr_info("mymodule: Exit success\n"); 
+61    kobject_put(mymodule); 
+62} 
+63 
+64module_init(mymodule_init); 
+65module_exit(mymodule_exit);
    +

    Make and install the module:

    - 
    1make 
-2sudo insmod hello-sysfs.ko
    -

    Check that it exists: + 

    1make 
+2sudo insmod hello-sysfs.ko
    +

    Check that it exists:

    - 
    1sudo lsmod | grep hello_sysfs
    -

    What is the current value of myvariable ? + 

    1sudo lsmod | grep hello_sysfs
    +

    What is the current value of myvariable ?

    - 
    1cat /sys/kernel/mymodule/myvariable
    -

    Set the value of myvariable and check that it changed. + 

    1cat /sys/kernel/mymodule/myvariable
    +

    Set the value of myvariable and check that it changed.

    - 
    1echo "32" > /sys/kernel/mymodule/myvariable 
-2cat /sys/kernel/mymodule/myvariable
    -

    Finally, remove the test module: + 

    1echo "32" > /sys/kernel/mymodule/myvariable 
+2cat /sys/kernel/mymodule/myvariable
    +

    Finally, remove the test module:

    - 
    1sudo rmmod hello_sysfs
    -

    + 

    1sudo rmmod hello_sysfs
    +

    -

    0.9 Talking To Device Files

    -

    Device files are supposed to represent physical devices. Most physical devices are +

    0.9 Talking To Device Files

    +

    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 +

    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 @@ -2248,7 +2273,7 @@ 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 +

    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), both or neither. Notice @@ -2258,600 +2283,600 @@ write ioctl’s (to return information to a process), both or neither. Notice here the roles of read and write are reversed again, so in ioctl’s read is to send information to the kernel and write is to receive information from the kernel. -

    The ioctl function is called with three parameters: the file descriptor of the +

    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. You will not be able to pass a structure this way, but you will be able to pass a pointer to the structure. -

    The ioctl number encodes the major device number, the type of the ioctl, the +

    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 +

    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.

    - 
    1/* 
-2 *  chardev2.c - Create an input/output character device 
-3 */ 
-4 
-5#include <linux/cdev.h> 
-6#include <linux/delay.h> 
-7#include <linux/device.h> 
-8#include <linux/fs.h> 
-9#include <linux/init.h> 
-10#include <linux/irq.h> 
-11#include <linux/kernel.h> /* We're doing kernel work */ 
-12#include <linux/module.h> /* Specifically, a module */ 
-13#include <linux/poll.h> 
-14 
-15#include "chardev.h" 
-16#define SUCCESS 0 
-17#define DEVICE_NAME "char_dev" 
-18#define BUF_LEN 80 
-19 
-20/* 
-21 * Is the device open right now? Used to prevent 
-22 * concurent access into the same device 
-23 */ 
-24static int Device_Open = 0; 
-25 
-26/* 
-27 * The message the device will give when asked 
-28 */ 
-29static char Message[BUF_LEN]; 
-30 
-31/* 
-32 * How far did the process reading the message get? 
-33 * Useful if the message is larger than the size of the 
-34 * buffer we get to fill in device_read. 
-35 */ 
-36static char *Message_Ptr; 
-37 
-38static int Major; /* Major number assigned to our device driver */ 
-39static struct class *cls; 
-40 
-41/* 
-42 * This is called whenever a process attempts to open the device file 
-43 */ 
-44static int device_open(struct inode *inode, struct file *file) 
-45{ 
-46    pr_info("device_open(%p)\n", file); 
-47 
-48    /* 
-49     * We don't want to talk to two processes at the same time 
-50     */ 
-51    if (Device_Open) 
-52        return -EBUSY; 
-53 
-54    Device_Open++; 
-55    /* 
-56     * Initialize the message 
-57     */ 
-58    Message_Ptr = Message; 
-59    try_module_get(THIS_MODULE); 
-60    return SUCCESS; 
-61} 
-62 
-63static int device_release(struct inode *inode, struct file *file) 
-64{ 
-65    pr_info("device_release(%p,%p)\n", inode, file); 
-66 
-67    /* 
-68     * We're now ready for our next caller 
-69     */ 
-70    Device_Open--; 
-71 
-72    module_put(THIS_MODULE); 
-73    return SUCCESS; 
-74} 
-75 
-76/* 
-77 * This function is called whenever a process which has already opened the 
-78 * device file attempts to read from it. 
-79 */ 
-80static ssize_t device_read(struct file *file,   /* see include/linux/fs.h   */ 
-81                           char __user *buffer, /* buffer to be 
-82                                                 * filled with data */ 
-83                           size_t length,       /* length of the buffer     */ 
-84                           loff_t *offset) 
-85{ 
-86    /* 
-87     * Number of bytes actually written to the buffer 
-88     */ 
-89    int bytes_read = 0; 
-90 
-91    pr_info("device_read(%p,%p,%ld)\n", file, buffer, length); 
-92 
-93    /* 
-94     * If we're at the end of the message, return 0 
-95     * (which signifies end of file) 
-96     */ 
-97    if (*Message_Ptr == 0) 
-98        return 0; 
-99 
-100    /* 
-101     * Actually put the data into the buffer 
-102     */ 
-103    while (length && *Message_Ptr) { 
-104        /* 
-105         * Because the buffer is in the user data segment, 
-106         * not the kernel data segment, assignment wouldn't 
-107         * work. Instead, we have to use put_user which 
-108         * copies data from the kernel data segment to the 
-109         * user data segment. 
-110         */ 
-111        put_user(*(Message_Ptr++), buffer++); 
-112        length--; 
-113        bytes_read++; 
-114    } 
-115 
-116    pr_info("Read %d bytes, %ld left\n", bytes_read, length); 
-117 
-118    /* 
-119     * Read functions are supposed to return the number 
-120     * of bytes actually inserted into the buffer 
-121     */ 
-122    return bytes_read; 
-123} 
-124 
-125/* 
-126 * This function is called when somebody tries to 
-127 * write into our device file. 
-128 */ 
-129static ssize_t device_write(struct file *file, 
-130                            const char __user *buffer, 
-131                            size_t length, 
-132                            loff_t *offset) 
-133{ 
-134    int i; 
-135 
-136    pr_info("device_write(%p,%s,%ld)", file, buffer, length); 
-137 
-138    for (i = 0; i < length && i < BUF_LEN; i++) 
-139        get_user(Message[i], buffer + i); 
-140 
-141    Message_Ptr = Message; 
-142 
-143    /* 
-144     * Again, return the number of input characters used 
-145     */ 
-146    return i; 
-147} 
-148 
-149/* 
-150 * This function is called whenever a process tries to do an ioctl on our 
-151 * device file. We get two extra parameters (additional to the inode and file 
-152 * structures, which all device functions get): the number of the ioctl called 
-153 * and the parameter given to the ioctl function. 
-154 * 
-155 * If the ioctl is write or read/write (meaning output is returned to the 
-156 * calling process), the ioctl call returns the output of this function. 
-157 * 
-158 */ 
-159long device_ioctl(struct file *file,      /* ditto */ 
-160                  unsigned int ioctl_num, /* number and param for ioctl */ 
-161                  unsigned long ioctl_param) 
-162{ 
-163    int i; 
-164    char *temp; 
-165    char ch; 
-166 
-167    /* 
-168     * Switch according to the ioctl called 
-169     */ 
-170    switch (ioctl_num) { 
-171    case IOCTL_SET_MSG: 
-172        /* 
-173         * Receive a pointer to a message (in user space) and set that 
-174         * to be the device's message.  Get the parameter given to 
-175         * ioctl by the process. 
-176         */ 
-177        temp = (char *) ioctl_param; 
-178 
-179        /* 
-180         * Find the length of the message 
-181         */ 
-182        get_user(ch, temp); 
-183        for (i = 0; ch && i < BUF_LEN; i++, temp++) 
-184            get_user(ch, temp); 
-185 
-186        device_write(file, (char *) ioctl_param, i, 0); 
-187        break; 
-188 
-189    case IOCTL_GET_MSG: 
-190        /* 
-191         * Give the current message to the calling process - 
-192         * the parameter we got is a pointer, fill it. 
-193         */ 
-194        i = device_read(file, (char *) ioctl_param, 99, 0); 
-195 
-196        /* 
-197         * Put a zero at the end of the buffer, so it will be 
-198         * properly terminated 
-199         */ 
-200        put_user('\0', (char *) ioctl_param + i); 
-201        break; 
-202 
-203    case IOCTL_GET_NTH_BYTE: 
-204        /* 
-205         * This ioctl is both input (ioctl_param) and 
-206         * output (the return value of this function) 
-207         */ 
-208        return Message[ioctl_param]; 
-209        break; 
-210    } 
-211 
-212    return SUCCESS; 
-213} 
-214 
-215/* Module Declarations */ 
-216 
-217/* 
-218 * This structure will hold the functions to be called 
-219 * when a process does something to the device we 
-220 * created. Since a pointer to this structure is kept in 
-221 * the devices table, it can't be local to 
-222 * init_module. NULL is for unimplemented functions. 
-223 */ 
-224struct file_operations Fops = { 
-225    .read = device_read, 
-226    .write = device_write, 
-227    .unlocked_ioctl = device_ioctl, 
-228    .open = device_open, 
-229    .release = device_release, /* a.k.a. close */ 
-230}; 
-231 
-232/* 
-233 * Initialize the module - Register the character device 
-234 */ 
-235int init_module() 
-236{ 
-237    int ret_val; 
-238    /* 
-239     * Register the character device (atleast try) 
-240     */ 
-241    ret_val = register_chrdev(MAJOR_NUM, DEVICE_NAME, &Fops); 
-242 
-243    /* 
-244     * Negative values signify an error 
-245     */ 
-246    if (ret_val < 0) { 
-247        pr_alert("%s failed with %d\n", 
-248                 "Sorry, registering the character device ", ret_val); 
-249        return ret_val; 
-250    } 
-251 
-252    Major = ret_val; 
-253 
-254    cls = class_create(THIS_MODULE, DEVICE_FILE_NAME); 
-255    device_create(cls, NULL, MKDEV(Major, MAJOR_NUM), NULL, DEVICE_FILE_NAME); 
-256 
-257    pr_info("Device created on /dev/%s\n", DEVICE_FILE_NAME); 
-258 
-259    return 0; 
-260} 
-261 
-262/* 
-263 * Cleanup - unregister the appropriate file from /proc 
-264 */ 
-265void cleanup_module() 
-266{ 
-267    device_destroy(cls, MKDEV(Major, 0)); 
-268    class_destroy(cls); 
-269 
-270    /* 
-271     * Unregister the device 
-272     */ 
-273    unregister_chrdev(Major, DEVICE_NAME); 
-274} 
-275 
-276MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  chardev2.c - Create an input/output character device 
+3 */ 
+4 
+5#include <linux/cdev.h> 
+6#include <linux/delay.h> 
+7#include <linux/device.h> 
+8#include <linux/fs.h> 
+9#include <linux/init.h> 
+10#include <linux/irq.h> 
+11#include <linux/kernel.h> /* We're doing kernel work */ 
+12#include <linux/module.h> /* Specifically, a module */ 
+13#include <linux/poll.h> 
+14 
+15#include "chardev.h" 
+16#define SUCCESS 0 
+17#define DEVICE_NAME "char_dev" 
+18#define BUF_LEN 80 
+19 
+20/* 
+21 * Is the device open right now? Used to prevent 
+22 * concurent access into the same device 
+23 */ 
+24static int Device_Open = 0; 
+25 
+26/* 
+27 * The message the device will give when asked 
+28 */ 
+29static char Message[BUF_LEN]; 
+30 
+31/* 
+32 * How far did the process reading the message get? 
+33 * Useful if the message is larger than the size of the 
+34 * buffer we get to fill in device_read. 
+35 */ 
+36static char *Message_Ptr; 
+37 
+38static int Major; /* Major number assigned to our device driver */ 
+39static struct class *cls; 
+40 
+41/* 
+42 * This is called whenever a process attempts to open the device file 
+43 */ 
+44static int device_open(struct inode *inode, struct file *file) 
+45{ 
+46    pr_info("device_open(%p)\n", file); 
+47 
+48    /* 
+49     * We don't want to talk to two processes at the same time 
+50     */ 
+51    if (Device_Open) 
+52        return -EBUSY; 
+53 
+54    Device_Open++; 
+55    /* 
+56     * Initialize the message 
+57     */ 
+58    Message_Ptr = Message; 
+59    try_module_get(THIS_MODULE); 
+60    return SUCCESS; 
+61} 
+62 
+63static int device_release(struct inode *inode, struct file *file) 
+64{ 
+65    pr_info("device_release(%p,%p)\n", inode, file); 
+66 
+67    /* 
+68     * We're now ready for our next caller 
+69     */ 
+70    Device_Open--; 
+71 
+72    module_put(THIS_MODULE); 
+73    return SUCCESS; 
+74} 
+75 
+76/* 
+77 * This function is called whenever a process which has already opened the 
+78 * device file attempts to read from it. 
+79 */ 
+80static ssize_t device_read(struct file *file,   /* see include/linux/fs.h   */ 
+81                           char __user *buffer, /* buffer to be 
+82                                                 * filled with data */ 
+83                           size_t length,       /* length of the buffer     */ 
+84                           loff_t *offset) 
+85{ 
+86    /* 
+87     * Number of bytes actually written to the buffer 
+88     */ 
+89    int bytes_read = 0; 
+90 
+91    pr_info("device_read(%p,%p,%ld)\n", file, buffer, length); 
+92 
+93    /* 
+94     * If we're at the end of the message, return 0 
+95     * (which signifies end of file) 
+96     */ 
+97    if (*Message_Ptr == 0) 
+98        return 0; 
+99 
+100    /* 
+101     * Actually put the data into the buffer 
+102     */ 
+103    while (length && *Message_Ptr) { 
+104        /* 
+105         * Because the buffer is in the user data segment, 
+106         * not the kernel data segment, assignment wouldn't 
+107         * work. Instead, we have to use put_user which 
+108         * copies data from the kernel data segment to the 
+109         * user data segment. 
+110         */ 
+111        put_user(*(Message_Ptr++), buffer++); 
+112        length--; 
+113        bytes_read++; 
+114    } 
+115 
+116    pr_info("Read %d bytes, %ld left\n", bytes_read, length); 
+117 
+118    /* 
+119     * Read functions are supposed to return the number 
+120     * of bytes actually inserted into the buffer 
+121     */ 
+122    return bytes_read; 
+123} 
+124 
+125/* 
+126 * This function is called when somebody tries to 
+127 * write into our device file. 
+128 */ 
+129static ssize_t device_write(struct file *file, 
+130                            const char __user *buffer, 
+131                            size_t length, 
+132                            loff_t *offset) 
+133{ 
+134    int i; 
+135 
+136    pr_info("device_write(%p,%s,%ld)", file, buffer, length); 
+137 
+138    for (i = 0; i < length && i < BUF_LEN; i++) 
+139        get_user(Message[i], buffer + i); 
+140 
+141    Message_Ptr = Message; 
+142 
+143    /* 
+144     * Again, return the number of input characters used 
+145     */ 
+146    return i; 
+147} 
+148 
+149/* 
+150 * This function is called whenever a process tries to do an ioctl on our 
+151 * device file. We get two extra parameters (additional to the inode and file 
+152 * structures, which all device functions get): the number of the ioctl called 
+153 * and the parameter given to the ioctl function. 
+154 * 
+155 * If the ioctl is write or read/write (meaning output is returned to the 
+156 * calling process), the ioctl call returns the output of this function. 
+157 * 
+158 */ 
+159long device_ioctl(struct file *file,      /* ditto */ 
+160                  unsigned int ioctl_num, /* number and param for ioctl */ 
+161                  unsigned long ioctl_param) 
+162{ 
+163    int i; 
+164    char *temp; 
+165    char ch; 
+166 
+167    /* 
+168     * Switch according to the ioctl called 
+169     */ 
+170    switch (ioctl_num) { 
+171    case IOCTL_SET_MSG: 
+172        /* 
+173         * Receive a pointer to a message (in user space) and set that 
+174         * to be the device's message.  Get the parameter given to 
+175         * ioctl by the process. 
+176         */ 
+177        temp = (char *) ioctl_param; 
+178 
+179        /* 
+180         * Find the length of the message 
+181         */ 
+182        get_user(ch, temp); 
+183        for (i = 0; ch && i < BUF_LEN; i++, temp++) 
+184            get_user(ch, temp); 
+185 
+186        device_write(file, (char *) ioctl_param, i, 0); 
+187        break; 
+188 
+189    case IOCTL_GET_MSG: 
+190        /* 
+191         * Give the current message to the calling process - 
+192         * the parameter we got is a pointer, fill it. 
+193         */ 
+194        i = device_read(file, (char *) ioctl_param, 99, 0); 
+195 
+196        /* 
+197         * Put a zero at the end of the buffer, so it will be 
+198         * properly terminated 
+199         */ 
+200        put_user('\0', (char *) ioctl_param + i); 
+201        break; 
+202 
+203    case IOCTL_GET_NTH_BYTE: 
+204        /* 
+205         * This ioctl is both input (ioctl_param) and 
+206         * output (the return value of this function) 
+207         */ 
+208        return Message[ioctl_param]; 
+209        break; 
+210    } 
+211 
+212    return SUCCESS; 
+213} 
+214 
+215/* Module Declarations */ 
+216 
+217/* 
+218 * This structure will hold the functions to be called 
+219 * when a process does something to the device we 
+220 * created. Since a pointer to this structure is kept in 
+221 * the devices table, it can't be local to 
+222 * init_module. NULL is for unimplemented functions. 
+223 */ 
+224struct file_operations Fops = { 
+225    .read = device_read, 
+226    .write = device_write, 
+227    .unlocked_ioctl = device_ioctl, 
+228    .open = device_open, 
+229    .release = device_release, /* a.k.a. close */ 
+230}; 
+231 
+232/* 
+233 * Initialize the module - Register the character device 
+234 */ 
+235int init_module() 
+236{ 
+237    int ret_val; 
+238    /* 
+239     * Register the character device (atleast try) 
+240     */ 
+241    ret_val = register_chrdev(MAJOR_NUM, DEVICE_NAME, &Fops); 
+242 
+243    /* 
+244     * Negative values signify an error 
+245     */ 
+246    if (ret_val < 0) { 
+247        pr_alert("%s failed with %d\n", 
+248                 "Sorry, registering the character device ", ret_val); 
+249        return ret_val; 
+250    } 
+251 
+252    Major = ret_val; 
+253 
+254    cls = class_create(THIS_MODULE, DEVICE_FILE_NAME); 
+255    device_create(cls, NULL, MKDEV(Major, MAJOR_NUM), NULL, DEVICE_FILE_NAME); 
+256 
+257    pr_info("Device created on /dev/%s\n", DEVICE_FILE_NAME); 
+258 
+259    return 0; 
+260} 
+261 
+262/* 
+263 * Cleanup - unregister the appropriate file from /proc 
+264 */ 
+265void cleanup_module() 
+266{ 
+267    device_destroy(cls, MKDEV(Major, 0)); 
+268    class_destroy(cls); 
+269 
+270    /* 
+271     * Unregister the device 
+272     */ 
+273    unregister_chrdev(Major, DEVICE_NAME); 
+274} 
+275 
+276MODULE_LICENSE("GPL");

    - 
    1/* 
-2 *  chardev.h - the header file with the ioctl definitions. 
-3 * 
-4 *  The declarations here have to be in a header file, because 
-5 *  they need to be known both to the kernel module 
-6 *  (in chardev.c) and the process calling ioctl (ioctl.c) 
-7 */ 
-8 
-9#ifndef CHARDEV_H 
-10#define CHARDEV_H 
-11 
-12#include <linux/ioctl.h> 
-13 
-14/* 
-15 * The major device number. We can't rely on dynamic 
-16 * registration any more, because ioctls need to know 
-17 * it. 
-18 */ 
-19#define MAJOR_NUM 100 
-20 
-21/* 
-22 * Set the message of the device driver 
-23 */ 
-24#define IOCTL_SET_MSG _IOW(MAJOR_NUM, 0, char *) 
-25/* 
-26 * _IOW means that we're creating an ioctl command 
-27 * number for passing information from a user process 
-28 * to the kernel module. 
-29 * 
-30 * The first arguments, MAJOR_NUM, is the major device 
-31 * number we're using. 
-32 * 
-33 * The second argument is the number of the command 
-34 * (there could be several with different meanings). 
-35 * 
-36 * The third argument is the type we want to get from 
-37 * the process to the kernel. 
-38 */ 
-39 
-40/* 
-41 * Get the message of the device driver 
-42 */ 
-43#define IOCTL_GET_MSG _IOR(MAJOR_NUM, 1, char *) 
-44/* 
-45 * This IOCTL is used for output, to get the message 
-46 * of the device driver. However, we still need the 
-47 * buffer to place the message in to be input, 
-48 * as it is allocated by the process. 
-49 */ 
-50 
-51/* 
-52 * Get the n'th byte of the message 
-53 */ 
-54#define IOCTL_GET_NTH_BYTE _IOWR(MAJOR_NUM, 2, int) 
-55/* 
-56 * The IOCTL is used for both input and output. It 
-57 * receives from the user a number, n, and returns 
-58 * Message[n]. 
-59 */ 
-60 
-61/* 
-62 * The name of the device file 
-63 */ 
-64#define DEVICE_FILE_NAME "char_dev" 
-65 
-66#endif
    + 
    1/* 
+2 *  chardev.h - the header file with the ioctl definitions. 
+3 * 
+4 *  The declarations here have to be in a header file, because 
+5 *  they need to be known both to the kernel module 
+6 *  (in chardev.c) and the process calling ioctl (ioctl.c) 
+7 */ 
+8 
+9#ifndef CHARDEV_H 
+10#define CHARDEV_H 
+11 
+12#include <linux/ioctl.h> 
+13 
+14/* 
+15 * The major device number. We can't rely on dynamic 
+16 * registration any more, because ioctls need to know 
+17 * it. 
+18 */ 
+19#define MAJOR_NUM 100 
+20 
+21/* 
+22 * Set the message of the device driver 
+23 */ 
+24#define IOCTL_SET_MSG _IOW(MAJOR_NUM, 0, char *) 
+25/* 
+26 * _IOW means that we're creating an ioctl command 
+27 * number for passing information from a user process 
+28 * to the kernel module. 
+29 * 
+30 * The first arguments, MAJOR_NUM, is the major device 
+31 * number we're using. 
+32 * 
+33 * The second argument is the number of the command 
+34 * (there could be several with different meanings). 
+35 * 
+36 * The third argument is the type we want to get from 
+37 * the process to the kernel. 
+38 */ 
+39 
+40/* 
+41 * Get the message of the device driver 
+42 */ 
+43#define IOCTL_GET_MSG _IOR(MAJOR_NUM, 1, char *) 
+44/* 
+45 * This IOCTL is used for output, to get the message 
+46 * of the device driver. However, we still need the 
+47 * buffer to place the message in to be input, 
+48 * as it is allocated by the process. 
+49 */ 
+50 
+51/* 
+52 * Get the n'th byte of the message 
+53 */ 
+54#define IOCTL_GET_NTH_BYTE _IOWR(MAJOR_NUM, 2, int) 
+55/* 
+56 * The IOCTL is used for both input and output. It 
+57 * receives from the user a number, n, and returns 
+58 * Message[n]. 
+59 */ 
+60 
+61/* 
+62 * The name of the device file 
+63 */ 
+64#define DEVICE_FILE_NAME "char_dev" 
+65 
+66#endif

    - 
    1/* 
-2 *  ioctl.c 
-3 */ 
-4#include <linux/cdev.h> 
-5#include <linux/fs.h> 
-6#include <linux/init.h> 
-7#include <linux/ioctl.h> 
-8#include <linux/module.h> 
-9#include <linux/slab.h> 
-10#include <linux/uaccess.h> 
-11 
-12struct ioctl_arg { 
-13    unsigned int reg; 
-14    unsigned int val; 
-15}; 
-16 
-17/* Documentation/ioctl/ioctl-number.txt */ 
-18#define IOC_MAGIC '\x66' 
-19 
-20#define IOCTL_VALSET _IOW(IOC_MAGIC, 0, struct ioctl_arg) 
-21#define IOCTL_VALGET _IOR(IOC_MAGIC, 1, struct ioctl_arg) 
-22#define IOCTL_VALGET_NUM _IOR(IOC_MAGIC, 2, int) 
-23#define IOCTL_VALSET_NUM _IOW(IOC_MAGIC, 3, int) 
-24 
-25#define IOCTL_VAL_MAXNR 3 
-26#define DRIVER_NAME "ioctltest" 
-27 
-28static unsigned int test_ioctl_major = 0; 
-29static unsigned int num_of_dev = 1; 
-30static struct cdev test_ioctl_cdev; 
-31static int ioctl_num = 0; 
-32 
-33struct test_ioctl_data { 
-34    unsigned char val; 
-35    rwlock_t lock; 
-36}; 
-37 
-38static long test_ioctl_ioctl(struct file *filp, 
-39                             unsigned int cmd, 
-40                             unsigned long arg) 
-41{ 
-42    struct test_ioctl_data *ioctl_data = filp->private_data; 
-43    int retval = 0; 
-44    unsigned char val; 
-45    struct ioctl_arg data; 
-46    memset(&data, 0, sizeof(data)); 
-47 
-48    switch (cmd) { 
-49    case IOCTL_VALSET: 
-50 
-51        /* 
-52        if (!capable(CAP_SYS_ADMIN)) { 
-53         retval = -EPERM; 
-54         goto done; 
-55        } 
-56        if (!access_ok(VERIFY_READ, (void __user *)arg, _IOC_SIZE(cmd))) { 
-57         retval = -EFAULT; 
-58         goto done; 
-59        } 
-60        */ 
-61        if (copy_from_user(&data, (int __user *) arg, sizeof(data))) { 
-62            retval = -EFAULT; 
-63            goto done; 
-64        } 
-65 
-66        pr_alert("IOCTL set val:%x .\n", data.val); 
-67        write_lock(&ioctl_data->lock); 
-68        ioctl_data->val = data.val; 
-69        write_unlock(&ioctl_data->lock); 
-70        break; 
-71 
-72    case IOCTL_VALGET: 
-73        /* 
-74        if (!access_ok(VERIFY_WRITE, (void __user *)arg, _IOC_SIZE(cmd))) { 
-75                                     retval = -EFAULT; 
-76                                     goto done; 
-77                             } 
-78        */ 
-79        read_lock(&ioctl_data->lock); 
-80        val = ioctl_data->val; 
-81        read_unlock(&ioctl_data->lock); 
-82        data.val = val; 
-83 
-84        if (copy_to_user((int __user *) arg, &data, sizeof(data))) { 
-85            retval = -EFAULT; 
-86            goto done; 
-87        } 
-88 
-89        break; 
-90 
-91    case IOCTL_VALGET_NUM: 
-92        retval = __put_user(ioctl_num, (int __user *) arg); 
-93        break; 
-94 
-95    case IOCTL_VALSET_NUM: 
-96        /* 
-97        if (!capable(CAP_SYS_ADMIN)) 
-98         return -EPERM; 
-99        */ 
-100        ioctl_num = arg; 
-101        break; 
-102 
-103    default: 
-104        retval = -ENOTTY; 
-105    } 
-106 
-107done: 
-108    return retval; 
-109} 
-110 
-111ssize_t test_ioctl_read(struct file *filp, 
-112                        char __user *buf, 
-113                        size_t count, 
-114                        loff_t *f_pos) 
-115{ 
-116    struct test_ioctl_data *ioctl_data = filp->private_data; 
-117    unsigned char val; 
-118    int retval; 
-119    int i = 0; 
-120    read_lock(&ioctl_data->lock); 
-121    val = ioctl_data->val; 
-122    read_unlock(&ioctl_data->lock); 
-123 
-124    for (; i < count; i++) { 
-125        if (copy_to_user(&buf[i], &val, 1)) { 
-126            retval = -EFAULT; 
-127            goto out; 
-128        } 
-129    } 
-130 
-131    retval = count; 
-132out: 
-133    return retval; 
-134} 
-135 
-136static int test_ioctl_close(struct inode *inode, struct file *filp) 
-137{ 
-138    pr_alert("%s call.\n", __func__); 
-139 
-140    if (filp->private_data) { 
-141        kfree(filp->private_data); 
-142        filp->private_data = NULL; 
-143    } 
-144 
-145    return 0; 
-146} 
-147 
-148static int test_ioctl_open(struct inode *inode, struct file *filp) 
-149{ 
-150    struct test_ioctl_data *ioctl_data; 
-151    pr_alert("%s call.\n", __func__); 
-152    ioctl_data = kmalloc(sizeof(struct test_ioctl_data), GFP_KERNEL); 
-153 
-154    if (ioctl_data == NULL) { 
-155        return -ENOMEM; 
-156    } 
-157 
-158    rwlock_init(&ioctl_data->lock); 
-159    ioctl_data->val = 0xFF; 
-160    filp->private_data = ioctl_data; 
-161    return 0; 
-162} 
-163 
-164struct file_operations fops = { 
-165    .owner = THIS_MODULE, 
-166    .open = test_ioctl_open, 
-167    .release = test_ioctl_close, 
-168    .read = test_ioctl_read, 
-169    .unlocked_ioctl = test_ioctl_ioctl, 
-170}; 
-171 
-172static int ioctl_init(void) 
-173{ 
-174    dev_t dev = MKDEV(test_ioctl_major, 0); 
-175    int alloc_ret = 0; 
-176    int cdev_ret = 0; 
-177    alloc_ret = alloc_chrdev_region(&dev, 0, num_of_dev, DRIVER_NAME); 
-178 
-179    if (alloc_ret) { 
-180        goto error; 
-181    } 
-182 
-183    test_ioctl_major = MAJOR(dev); 
-184    cdev_init(&test_ioctl_cdev, &fops); 
-185    cdev_ret = cdev_add(&test_ioctl_cdev, dev, num_of_dev); 
-186 
-187    if (cdev_ret) { 
-188        goto error; 
-189    } 
-190 
-191    pr_alert("%s driver(major: %d) installed.\n", DRIVER_NAME, 
-192             test_ioctl_major); 
-193    return 0; 
-194error: 
-195 
-196    if (cdev_ret == 0) { 
-197        cdev_del(&test_ioctl_cdev); 
-198    } 
-199 
-200    if (alloc_ret == 0) { 
-201        unregister_chrdev_region(dev, num_of_dev); 
-202    } 
-203 
-204    return -1; 
-205} 
-206 
-207static void ioctl_exit(void) 
-208{ 
-209    dev_t dev = MKDEV(test_ioctl_major, 0); 
-210    cdev_del(&test_ioctl_cdev); 
-211    unregister_chrdev_region(dev, num_of_dev); 
-212    pr_alert("%s driver removed.\n", DRIVER_NAME); 
-213} 
-214 
-215module_init(ioctl_init); 
-216module_exit(ioctl_exit); 
-217 
-218MODULE_LICENSE("GPL"); 
-219MODULE_DESCRIPTION("This is test_ioctl module");
    -

    + 

    1/* 
+2 *  ioctl.c 
+3 */ 
+4#include <linux/cdev.h> 
+5#include <linux/fs.h> 
+6#include <linux/init.h> 
+7#include <linux/ioctl.h> 
+8#include <linux/module.h> 
+9#include <linux/slab.h> 
+10#include <linux/uaccess.h> 
+11 
+12struct ioctl_arg { 
+13    unsigned int reg; 
+14    unsigned int val; 
+15}; 
+16 
+17/* Documentation/ioctl/ioctl-number.txt */ 
+18#define IOC_MAGIC '\x66' 
+19 
+20#define IOCTL_VALSET _IOW(IOC_MAGIC, 0, struct ioctl_arg) 
+21#define IOCTL_VALGET _IOR(IOC_MAGIC, 1, struct ioctl_arg) 
+22#define IOCTL_VALGET_NUM _IOR(IOC_MAGIC, 2, int) 
+23#define IOCTL_VALSET_NUM _IOW(IOC_MAGIC, 3, int) 
+24 
+25#define IOCTL_VAL_MAXNR 3 
+26#define DRIVER_NAME "ioctltest" 
+27 
+28static unsigned int test_ioctl_major = 0; 
+29static unsigned int num_of_dev = 1; 
+30static struct cdev test_ioctl_cdev; 
+31static int ioctl_num = 0; 
+32 
+33struct test_ioctl_data { 
+34    unsigned char val; 
+35    rwlock_t lock; 
+36}; 
+37 
+38static long test_ioctl_ioctl(struct file *filp, 
+39                             unsigned int cmd, 
+40                             unsigned long arg) 
+41{ 
+42    struct test_ioctl_data *ioctl_data = filp->private_data; 
+43    int retval = 0; 
+44    unsigned char val; 
+45    struct ioctl_arg data; 
+46    memset(&data, 0, sizeof(data)); 
+47 
+48    switch (cmd) { 
+49    case IOCTL_VALSET: 
+50 
+51        /* 
+52        if (!capable(CAP_SYS_ADMIN)) { 
+53         retval = -EPERM; 
+54         goto done; 
+55        } 
+56        if (!access_ok(VERIFY_READ, (void __user *)arg, _IOC_SIZE(cmd))) { 
+57         retval = -EFAULT; 
+58         goto done; 
+59        } 
+60        */ 
+61        if (copy_from_user(&data, (int __user *) arg, sizeof(data))) { 
+62            retval = -EFAULT; 
+63            goto done; 
+64        } 
+65 
+66        pr_alert("IOCTL set val:%x .\n", data.val); 
+67        write_lock(&ioctl_data->lock); 
+68        ioctl_data->val = data.val; 
+69        write_unlock(&ioctl_data->lock); 
+70        break; 
+71 
+72    case IOCTL_VALGET: 
+73        /* 
+74        if (!access_ok(VERIFY_WRITE, (void __user *)arg, _IOC_SIZE(cmd))) { 
+75                                     retval = -EFAULT; 
+76                                     goto done; 
+77                             } 
+78        */ 
+79        read_lock(&ioctl_data->lock); 
+80        val = ioctl_data->val; 
+81        read_unlock(&ioctl_data->lock); 
+82        data.val = val; 
+83 
+84        if (copy_to_user((int __user *) arg, &data, sizeof(data))) { 
+85            retval = -EFAULT; 
+86            goto done; 
+87        } 
+88 
+89        break; 
+90 
+91    case IOCTL_VALGET_NUM: 
+92        retval = __put_user(ioctl_num, (int __user *) arg); 
+93        break; 
+94 
+95    case IOCTL_VALSET_NUM: 
+96        /* 
+97        if (!capable(CAP_SYS_ADMIN)) 
+98         return -EPERM; 
+99        */ 
+100        ioctl_num = arg; 
+101        break; 
+102 
+103    default: 
+104        retval = -ENOTTY; 
+105    } 
+106 
+107done: 
+108    return retval; 
+109} 
+110 
+111ssize_t test_ioctl_read(struct file *filp, 
+112                        char __user *buf, 
+113                        size_t count, 
+114                        loff_t *f_pos) 
+115{ 
+116    struct test_ioctl_data *ioctl_data = filp->private_data; 
+117    unsigned char val; 
+118    int retval; 
+119    int i = 0; 
+120    read_lock(&ioctl_data->lock); 
+121    val = ioctl_data->val; 
+122    read_unlock(&ioctl_data->lock); 
+123 
+124    for (; i < count; i++) { 
+125        if (copy_to_user(&buf[i], &val, 1)) { 
+126            retval = -EFAULT; 
+127            goto out; 
+128        } 
+129    } 
+130 
+131    retval = count; 
+132out: 
+133    return retval; 
+134} 
+135 
+136static int test_ioctl_close(struct inode *inode, struct file *filp) 
+137{ 
+138    pr_alert("%s call.\n", __func__); 
+139 
+140    if (filp->private_data) { 
+141        kfree(filp->private_data); 
+142        filp->private_data = NULL; 
+143    } 
+144 
+145    return 0; 
+146} 
+147 
+148static int test_ioctl_open(struct inode *inode, struct file *filp) 
+149{ 
+150    struct test_ioctl_data *ioctl_data; 
+151    pr_alert("%s call.\n", __func__); 
+152    ioctl_data = kmalloc(sizeof(struct test_ioctl_data), GFP_KERNEL); 
+153 
+154    if (ioctl_data == NULL) { 
+155        return -ENOMEM; 
+156    } 
+157 
+158    rwlock_init(&ioctl_data->lock); 
+159    ioctl_data->val = 0xFF; 
+160    filp->private_data = ioctl_data; 
+161    return 0; 
+162} 
+163 
+164struct file_operations fops = { 
+165    .owner = THIS_MODULE, 
+166    .open = test_ioctl_open, 
+167    .release = test_ioctl_close, 
+168    .read = test_ioctl_read, 
+169    .unlocked_ioctl = test_ioctl_ioctl, 
+170}; 
+171 
+172static int ioctl_init(void) 
+173{ 
+174    dev_t dev = MKDEV(test_ioctl_major, 0); 
+175    int alloc_ret = 0; 
+176    int cdev_ret = 0; 
+177    alloc_ret = alloc_chrdev_region(&dev, 0, num_of_dev, DRIVER_NAME); 
+178 
+179    if (alloc_ret) { 
+180        goto error; 
+181    } 
+182 
+183    test_ioctl_major = MAJOR(dev); 
+184    cdev_init(&test_ioctl_cdev, &fops); 
+185    cdev_ret = cdev_add(&test_ioctl_cdev, dev, num_of_dev); 
+186 
+187    if (cdev_ret) { 
+188        goto error; 
+189    } 
+190 
+191    pr_alert("%s driver(major: %d) installed.\n", DRIVER_NAME, 
+192             test_ioctl_major); 
+193    return 0; 
+194error: 
+195 
+196    if (cdev_ret == 0) { 
+197        cdev_del(&test_ioctl_cdev); 
+198    } 
+199 
+200    if (alloc_ret == 0) { 
+201        unregister_chrdev_region(dev, num_of_dev); 
+202    } 
+203 
+204    return -1; 
+205} 
+206 
+207static void ioctl_exit(void) 
+208{ 
+209    dev_t dev = MKDEV(test_ioctl_major, 0); 
+210    cdev_del(&test_ioctl_cdev); 
+211    unregister_chrdev_region(dev, num_of_dev); 
+212    pr_alert("%s driver removed.\n", DRIVER_NAME); 
+213} 
+214 
+215module_init(ioctl_init); 
+216module_exit(ioctl_exit); 
+217 
+218MODULE_LICENSE("GPL"); 
+219MODULE_DESCRIPTION("This is test_ioctl module");
    +

    -

    0.10 System Calls

    -

    So far, the only thing we’ve done was to use well defined kernel mechanisms to +

    0.10 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 are mostly on your own. -

    If you are not being sensible and using a virtual machine then this is where kernel +

    If you are not being sensible and using a virtual machine then this is where kernel programming can become hazardous. While writing the example below, I killed the open() system call. This meant I could not open any files, I could not run any programs, and I could not shutdown the system. I had @@ -2863,7 +2888,7 @@ was doing this on some live mission critical system then that could have been a possible outcome. To ensure you do not lose any files, even within a test environment, please run sync right before you do the insmod and the rmmod. -

    Forget about /proc files, forget about device files. They are just minor details. +

    Forget about /proc files, forget about device files. They are just minor details. Minutiae in the vast expanse of the universe. 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 @@ -2871,11 +2896,11 @@ 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 not +

    In general, a process is not supposed to be able to access the kernel. It can not access kernel memory and it can’t call kernel functions. The hardware of the CPU enforces this (that is the reason why it is called “protected mode” or “page protection”). -

    System calls are an exception to this general rule. What happens is that the +

    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, @@ -2883,7 +2908,7 @@ this is done by means of interrupt 0x80. The hardware knows that once you jump t 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 +

    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 @@ -2891,13 +2916,13 @@ function, and after it returns, does a few system checks and then return back to process (or to a different process, if the process time ran out). If you want to read this code, it is 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 +

    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 +

    The source code here is an example of such a kernel module. We want to “spy” on a certain user, and to pr_info() 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 @@ -2908,7 +2933,7 @@ file. -

    The init_module function replaces the appropriate location in sys_call_table +

    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 @@ -2918,7 +2943,7 @@ A_open, which will call the original sys_open when it is 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 +

    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 @@ -2932,7 +2957,7 @@ pointing to A_open, so it will not restore it to 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 +

    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 @@ -2945,191 +2970,191 @@ Depending on your kernel version, you might even need to hand apply the patch.

    - 
    1/* 
-2 *  syscall.c 
-3 * 
-4 *  System call "stealing" sample. 
-5 * 
-6 *  Disables page protection at a processor level by 
-7 *  changing the 16th bit in the cr0 register (could be Intel specific) 
-8 * 
-9 *  Based on example by Peter Jay Salzman and 
-10 *  https://bbs.archlinux.org/viewtopic.php?id=139406 
-11 */ 
-12 
-13#include <linux/delay.h> 
-14#include <linux/kernel.h> 
-15#include <linux/module.h> 
-16#include <linux/moduleparam.h> /* which will have params */ 
-17#include <linux/syscalls.h> 
-18#include <linux/unistd.h> /* The list of system calls */ 
-19 
-20/* 
-21 * For the current (process) structure, we need 
-22 * this to know who the current user is. 
-23 */ 
-24#include <linux/sched.h> 
-25#include <linux/uaccess.h> 
-26 
-27unsigned long **sys_call_table; 
-28unsigned long original_cr0; 
-29 
-30/* 
-31 * UID we want to spy on - will be filled from the 
-32 * command line 
-33 */ 
-34static int uid; 
-35module_param(uid, int, 0644); 
-36 
-37/* 
-38 * A pointer to the original system call. The reason 
-39 * we keep this, rather than call the original function 
-40 * (sys_open), is because somebody else might have 
-41 * replaced the system call before us. Note that this 
-42 * is not 100% safe, because if another module 
-43 * replaced sys_open before us, then when we're inserted 
-44 * we'll call the function in that module - and it 
-45 * might be removed before we are. 
-46 * 
-47 * Another reason for this is that we can't get sys_open. 
-48 * It's a static variable, so it is not exported. 
-49 */ 
-50asmlinkage int (*original_call)(const char *, int, int); 
-51 
-52/* 
-53 * The function we'll replace sys_open (the function 
-54 * called when you call the open system call) with. To 
-55 * find the exact prototype, with the number and type 
-56 * of arguments, we find the original function first 
-57 * (it's at fs/open.c). 
-58 * 
-59 * In theory, this means that we're tied to the 
-60 * current version of the kernel. In practice, the 
-61 * system calls almost never change (it would wreck havoc 
-62 * and require programs to be recompiled, since the system 
-63 * calls are the interface between the kernel and the 
-64 * processes). 
-65 */ 
-66asmlinkage int our_sys_open(const char *filename, int flags, int mode) 
-67{ 
-68    int i = 0; 
-69    char ch; 
-70 
-71    /* 
-72     * Report the file, if relevant 
-73     */ 
-74    pr_info("Opened file by %d: ", uid); 
-75    do { 
-76        get_user(ch, filename + i); 
-77        i++; 
-78        pr_info("%c", ch); 
-79    } while (ch != 0); 
-80    pr_info("\n"); 
-81 
-82    /* 
-83     * Call the original sys_open - otherwise, we lose 
-84     * the ability to open files 
-85     */ 
-86    return original_call(filename, flags, mode); 
-87} 
-88 
-89static unsigned long **aquire_sys_call_table(void) 
-90{ 
-91    unsigned long int offset = PAGE_OFFSET; 
-92    unsigned long **sct; 
-93 
-94    while (offset < ULLONG_MAX) { 
-95        sct = (unsigned long **) offset; 
-96 
-97        if (sct[__NR_close] == (unsigned long *) ksys_close) 
-98            return sct; 
-99 
-100        offset += sizeof(void *); 
-101    } 
-102 
-103    return NULL; 
-104} 
-105 
-106static int __init syscall_start(void) 
-107{ 
-108    if (!(sys_call_table = aquire_sys_call_table())) 
-109        return -1; 
-110 
-111    original_cr0 = read_cr0(); 
-112 
-113    write_cr0(original_cr0 & ~0x00010000); 
-114 
-115    /* keep track of the original open function */ 
-116    original_call = (void *) sys_call_table[__NR_open]; 
-117 
-118    /* use our open function instead */ 
-119    sys_call_table[__NR_open] = (unsigned long *) our_sys_open; 
-120 
-121    write_cr0(original_cr0); 
-122 
-123    pr_info("Spying on UID:%d\n", uid); 
-124 
-125    return 0; 
-126} 
-127 
-128static void __exit syscall_end(void) 
-129{ 
-130    if (!sys_call_table) { 
-131        return; 
-132    } 
-133 
-134    /* 
-135     * Return the system call back to normal 
-136     */ 
-137    if (sys_call_table[__NR_open] != (unsigned long *) our_sys_open) { 
-138        pr_alert("Somebody else also played with the "); 
-139        pr_alert("open system call\n"); 
-140        pr_alert("The system may be left in "); 
-141        pr_alert("an unstable state.\n"); 
-142    } 
-143 
-144    write_cr0(original_cr0 & ~0x00010000); 
-145    sys_call_table[__NR_open] = (unsigned long *) original_call; 
-146    write_cr0(original_cr0); 
-147 
-148    msleep(2000); 
-149} 
-150 
-151module_init(syscall_start); 
-152module_exit(syscall_end); 
-153 
-154MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  syscall.c 
+3 * 
+4 *  System call "stealing" sample. 
+5 * 
+6 *  Disables page protection at a processor level by 
+7 *  changing the 16th bit in the cr0 register (could be Intel specific) 
+8 * 
+9 *  Based on example by Peter Jay Salzman and 
+10 *  https://bbs.archlinux.org/viewtopic.php?id=139406 
+11 */ 
+12 
+13#include <linux/delay.h> 
+14#include <linux/kernel.h> 
+15#include <linux/module.h> 
+16#include <linux/moduleparam.h> /* which will have params */ 
+17#include <linux/syscalls.h> 
+18#include <linux/unistd.h> /* The list of system calls */ 
+19 
+20/* 
+21 * For the current (process) structure, we need 
+22 * this to know who the current user is. 
+23 */ 
+24#include <linux/sched.h> 
+25#include <linux/uaccess.h> 
+26 
+27unsigned long **sys_call_table; 
+28unsigned long original_cr0; 
+29 
+30/* 
+31 * UID we want to spy on - will be filled from the 
+32 * command line 
+33 */ 
+34static int uid; 
+35module_param(uid, int, 0644); 
+36 
+37/* 
+38 * A pointer to the original system call. The reason 
+39 * we keep this, rather than call the original function 
+40 * (sys_open), is because somebody else might have 
+41 * replaced the system call before us. Note that this 
+42 * is not 100% safe, because if another module 
+43 * replaced sys_open before us, then when we're inserted 
+44 * we'll call the function in that module - and it 
+45 * might be removed before we are. 
+46 * 
+47 * Another reason for this is that we can't get sys_open. 
+48 * It's a static variable, so it is not exported. 
+49 */ 
+50asmlinkage int (*original_call)(const char *, int, int); 
+51 
+52/* 
+53 * The function we'll replace sys_open (the function 
+54 * called when you call the open system call) with. To 
+55 * find the exact prototype, with the number and type 
+56 * of arguments, we find the original function first 
+57 * (it's at fs/open.c). 
+58 * 
+59 * In theory, this means that we're tied to the 
+60 * current version of the kernel. In practice, the 
+61 * system calls almost never change (it would wreck havoc 
+62 * and require programs to be recompiled, since the system 
+63 * calls are the interface between the kernel and the 
+64 * processes). 
+65 */ 
+66asmlinkage int our_sys_open(const char *filename, int flags, int mode) 
+67{ 
+68    int i = 0; 
+69    char ch; 
+70 
+71    /* 
+72     * Report the file, if relevant 
+73     */ 
+74    pr_info("Opened file by %d: ", uid); 
+75    do { 
+76        get_user(ch, filename + i); 
+77        i++; 
+78        pr_info("%c", ch); 
+79    } while (ch != 0); 
+80    pr_info("\n"); 
+81 
+82    /* 
+83     * Call the original sys_open - otherwise, we lose 
+84     * the ability to open files 
+85     */ 
+86    return original_call(filename, flags, mode); 
+87} 
+88 
+89static unsigned long **aquire_sys_call_table(void) 
+90{ 
+91    unsigned long int offset = PAGE_OFFSET; 
+92    unsigned long **sct; 
+93 
+94    while (offset < ULLONG_MAX) { 
+95        sct = (unsigned long **) offset; 
+96 
+97        if (sct[__NR_close] == (unsigned long *) ksys_close) 
+98            return sct; 
+99 
+100        offset += sizeof(void *); 
+101    } 
+102 
+103    return NULL; 
+104} 
+105 
+106static int __init syscall_start(void) 
+107{ 
+108    if (!(sys_call_table = aquire_sys_call_table())) 
+109        return -1; 
+110 
+111    original_cr0 = read_cr0(); 
+112 
+113    write_cr0(original_cr0 & ~0x00010000); 
+114 
+115    /* keep track of the original open function */ 
+116    original_call = (void *) sys_call_table[__NR_open]; 
+117 
+118    /* use our open function instead */ 
+119    sys_call_table[__NR_open] = (unsigned long *) our_sys_open; 
+120 
+121    write_cr0(original_cr0); 
+122 
+123    pr_info("Spying on UID:%d\n", uid); 
+124 
+125    return 0; 
+126} 
+127 
+128static void __exit syscall_end(void) 
+129{ 
+130    if (!sys_call_table) { 
+131        return; 
+132    } 
+133 
+134    /* 
+135     * Return the system call back to normal 
+136     */ 
+137    if (sys_call_table[__NR_open] != (unsigned long *) our_sys_open) { 
+138        pr_alert("Somebody else also played with the "); 
+139        pr_alert("open system call\n"); 
+140        pr_alert("The system may be left in "); 
+141        pr_alert("an unstable state.\n"); 
+142    } 
+143 
+144    write_cr0(original_cr0 & ~0x00010000); 
+145    sys_call_table[__NR_open] = (unsigned long *) original_call; 
+146    write_cr0(original_cr0); 
+147 
+148    msleep(2000); 
+149} 
+150 
+151module_init(syscall_start); 
+152module_exit(syscall_end); 
+153 
+154MODULE_LICENSE("GPL");
    +

    -

    0.11 Blocking Processes and threads

    +

    0.11 Blocking Processes and threads

    -

    +

    -

    0.11.1 Sleep

    -

    What do you do when somebody asks you for something you can’t do right away? If +

    0.11.1 Sleep

    +

    What do you do when somebody asks you for something you can’t do right away? If you’re a human being and you’re bothered by a human being, the only thing you can say is: "Not right now, I’m busy. Go away!". But if you’re a kernel module and you’re bothered by a process, you have another possibility. You can put the process to sleep until you can service it. After all, processes are being put to sleep by the kernel and woken up all the time (that’s the way multiple processes appear to run on the same time on a single CPU). -

    This kernel module is an example of this. The file (called /proc/sleep) can only +

    This kernel module is an example of this. The file (called /proc/sleep) can only be opened by a single process at a time. If the file is already open, the kernel module calls wait_event_interruptible. The easiest way to keep a file open is to open it with:

    - 
    1tail -f
    -

    This function changes the status of the task (a task is the kernel data + 

    1tail -f
    +

    This function changes the status of the task (a task is the kernel data structure which holds information about a process and the system call it’s in, if any) to TASK_INTERRUPTIBLE, which means that the task will not run until it is woken up somehow, and adds it to WaitQ, the queue of tasks waiting to access the file. Then, the function calls the scheduler to context switch to a different process, one which has some use for the CPU. -

    When a process is done with the file, it closes it, and module_close is called. That +

    When a process is done with the file, it closes it, and module_close is called. That function wakes up all the processes in the queue (there’s no mechanism to only wake up one of them). It then returns and the process which just closed the file can continue to run. In time, the scheduler decides that that @@ -3137,14 +3162,14 @@ process has had enough and gives control of the CPU to another process. Eventually, one of the processes which was in the queue will be given control of the CPU by the scheduler. It starts at the point right after the call to module_interruptible_sleep_on. -

    This means that the process is still in kernel mode - as far as the process is +

    This means that the process is still in kernel mode - as far as the process is concerned, it issued the open system call and the system call hasn’t returned yet. The process doesn’t know somebody else used the CPU for most of the time between the moment it issued the call and the moment it returned. -

    It can then proceed to set a global variable to tell all the other processes that the +

    It can then proceed to set a global variable to tell all the other processes that the file is still open and go on with its life. When the other processes get a piece of the CPU, they’ll see that global variable and go back to sleep. -

    So we’ll use tail -f to keep the file open in the background, while trying to +

    So we’ll use tail -f to keep the file open in the background, while trying to access it with another process (again in the background, so that we need not switch to a different vt). As soon as the first background process is killed with kill %1 , the second is woken up, is able to access the file and finally @@ -3152,15 +3177,15 @@ terminates. -

    To make our life more interesting, module_close doesn’t have a monopoly +

    To make our life more interesting, module_close doesn’t have a monopoly on waking up the processes which wait to access the file. A signal, such as Ctrl +c (SIGINT) can also wake up a process. This is because we used module_interruptible_sleep_on. We could have used module_sleep_on instead, but that would have resulted in extremely angry users whose Ctrl+c’s are ignored. -

    In that case, we want to return with -EINTR immediately. This is important so +

    In that case, we want to return with -EINTR immediately. This is important so users can, for example, kill the process before it receives the file. -

    There is one more point to remember. Some times processes don’t want to sleep, +

    There is one more point to remember. Some times processes don’t want to sleep, they want either to get what they want immediately, or to be told it cannot be done. Such processes use the O_NONBLOCK flag when opening the file. The kernel is supposed to respond by returning with the error code -EAGAIN from operations @@ -3193,510 +3218,510 @@ $ cat_nonblock /proc/sleep Last input: $ -

    +

    - 
    1/* 
-2 *  sleep.c - create a /proc file, and if several processes try to open it at 
-3 *  the same time, put all but one to sleep 
-4 */ 
-5 
-6#include <linux/kernel.h>  /* We're doing kernel work */ 
-7#include <linux/module.h>  /* Specifically, a module */ 
-8#include <linux/proc_fs.h> /* Necessary because we use proc fs */ 
-9#include <linux/sched.h>   /* For putting processes to sleep and 
-10                                   waking them up */  
-11#include <linux/uaccess.h> /* for get_user and put_user */ 
-12#include <linux/version.h> 
-13 
-14#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-15#define HAVE_PROC_OPS 
-16#endif 
-17 
-18/* 
-19 * The module's file functions 
-20 */ 
-21 
-22/* 
-23 * Here we keep the last message received, to prove that we can process our 
-24 * input 
-25 */ 
-26#define MESSAGE_LENGTH 80 
-27static char Message[MESSAGE_LENGTH]; 
-28 
-29static struct proc_dir_entry *Our_Proc_File; 
-30#define PROC_ENTRY_FILENAME "sleep" 
-31 
-32/* 
-33 * Since we use the file operations struct, we can't use the special proc 
-34 * output provisions - we have to use a standard read function, which is this 
-35 * function 
-36 */ 
-37static ssize_t module_output(struct file *file, /* see include/linux/fs.h   */ 
-38                             char *buf,         /* The buffer to put data to 
-39                                                   (in the user segment)    */  
-40                             size_t len,        /* The length of the buffer */ 
-41                             loff_t *offset) 
-42{ 
-43    static int finished = 0; 
-44    int i; 
-45    char message[MESSAGE_LENGTH + 30]; 
-46 
-47    /* 
-48     * Return 0 to signify end of file - that we have nothing 
-49     * more to say at this point. 
-50     */ 
-51    if (finished) { 
-52        finished = 0; 
-53        return 0; 
-54    } 
-55 
-56    /* 
-57     * If you don't understand this by now, you're hopeless as a kernel 
-58     * programmer. 
-59     */ 
-60    sprintf(message, "Last input:%s\n", Message); 
-61    for (i = 0; i < len && message[i]; i++) 
-62        put_user(message[i], buf + i); 
-63 
-64    finished = 1; 
-65    return i; /* Return the number of bytes "read" */ 
-66} 
-67 
-68/* 
-69 * This function receives input from the user when the user writes to the /proc 
-70 * file. 
-71 */ 
-72static ssize_t module_input(struct file *file, /* The file itself */ 
-73                            const char *buf,   /* The buffer with input */ 
-74                            size_t length,     /* The buffer's length */ 
-75                            loff_t *offset)    /* offset to file - ignore */ 
-76{ 
-77    int i; 
-78 
-79    /* 
-80     * Put the input into Message, where module_output will later be 
-81     * able to use it 
-82     */ 
-83    for (i = 0; i < MESSAGE_LENGTH - 1 && i < length; i++) 
-84        get_user(Message[i], buf + i); 
-85    /* 
-86     * we want a standard, zero terminated string 
-87     */ 
-88    Message[i] = '\0'; 
-89 
-90    /* 
-91     * We need to return the number of input characters used 
-92     */ 
-93    return i; 
-94} 
-95 
-96/* 
-97 * 1 if the file is currently open by somebody 
-98 */ 
-99int Already_Open = 0; 
-100 
-101/* 
-102 * Queue of processes who want our file 
-103 */ 
-104DECLARE_WAIT_QUEUE_HEAD(WaitQ); 
-105/* 
-106 * Called when the /proc file is opened 
-107 */ 
-108static int module_open(struct inode *inode, struct file *file) 
-109{ 
-110    /* 
-111     * If the file's flags include O_NONBLOCK, it means the process doesn't 
-112     * want to wait for the file.  In this case, if the file is already 
-113     * open, we should fail with -EAGAIN, meaning "you'll have to try 
-114     * again", instead of blocking a process which would rather stay awake. 
-115     */ 
-116    if ((file->f_flags & O_NONBLOCK) && Already_Open) 
-117        return -EAGAIN; 
-118 
-119    /* 
-120     * This is the correct place for try_module_get(THIS_MODULE) because 
-121     * if a process is in the loop, which is within the kernel module, 
-122     * the kernel module must not be removed. 
-123     */ 
-124    try_module_get(THIS_MODULE); 
-125 
-126    /* 
-127     * If the file is already open, wait until it isn't 
-128     */ 
-129 
-130    while (Already_Open) { 
-131        int i, is_sig = 0; 
-132 
-133        /* 
-134         * This function puts the current process, including any system 
-135         * calls, such as us, to sleep.  Execution will be resumed right 
-136         * after the function call, either because somebody called 
-137         * wake_up(&WaitQ) (only module_close does that, when the file 
-138         * is closed) or when a signal, such as Ctrl-C, is sent 
-139         * to the process 
-140         */ 
-141        wait_event_interruptible(WaitQ, !Already_Open); 
-142 
-143        /* 
-144         * If we woke up because we got a signal we're not blocking, 
-145         * return -EINTR (fail the system call).  This allows processes 
-146         * to be killed or stopped. 
-147         */ 
-148 
-149        /* 
-150         * Emmanuel Papirakis: 
-151         * 
-152         * This is a little update to work with 2.2.*.  Signals now are 
-153         * contained in two words (64 bits) and are stored in a structure that 
-154         * contains an array of two unsigned longs.  We now have to make 2 
-155         * checks in our if. 
-156         * 
-157         * Ori Pomerantz: 
-158         * 
-159         * Nobody promised me they'll never use more than 64 bits, or that this 
-160         * book won't be used for a version of Linux with a word size of 16 
-161         * bits.  This code would work in any case. 
-162         */ 
-163        for (i = 0; i < _NSIG_WORDS && !is_sig; i++) 
-164            is_sig = current->pending.signal.sig[i] & ~current->blocked.sig[i]; 
-165 
-166        if (is_sig) { 
-167            /* 
-168             * It's important to put module_put(THIS_MODULE) here, 
-169             * because for processes where the open is interrupted 
-170             * there will never be a corresponding close. If we 
-171             * don't decrement the usage count here, we will be 
-172             * left with a positive usage count which we'll have no 
-173             * way to bring down to zero, giving us an immortal 
-174             * module, which can only be killed by rebooting 
-175             * the machine. 
-176             */ 
-177            module_put(THIS_MODULE); 
-178            return -EINTR; 
-179        } 
-180    } 
-181 
-182    /* 
-183     * If we got here, Already_Open must be zero 
-184     */ 
-185 
-186    /* 
-187     * Open the file 
-188     */ 
-189    Already_Open = 1; 
-190    return 0; /* Allow the access */ 
-191} 
-192 
-193/* 
-194 * Called when the /proc file is closed 
-195 */ 
-196int module_close(struct inode *inode, struct file *file) 
-197{ 
-198    /* 
-199     * Set Already_Open to zero, so one of the processes in the WaitQ will 
-200     * be able to set Already_Open back to one and to open the file. All 
-201     * the other processes will be called when Already_Open is back to one, 
-202     * so they'll go back to sleep. 
-203     */ 
-204    Already_Open = 0; 
-205 
-206    /* 
-207     * Wake up all the processes in WaitQ, so if anybody is waiting for the 
-208     * file, they can have it. 
-209     */ 
-210    wake_up(&WaitQ); 
-211 
-212    module_put(THIS_MODULE); 
-213 
-214    return 0; /* success */ 
-215} 
-216 
-217/* 
-218 * Structures to register as the /proc file, with pointers to all the relevant 
-219 * functions. 
-220 */ 
-221 
-222/* 
-223 * File operations for our proc file. This is where we place pointers to all 
-224 * the functions called when somebody tries to do something to our file. NULL 
-225 * means we don't want to deal with something. 
-226 */ 
-227#ifdef HAVE_PROC_OPS 
-228static const struct proc_ops File_Ops_4_Our_Proc_File = { 
-229    .proc_read = module_output,   /* "read" from the file */ 
-230    .proc_write = module_input,   /* "write" to the file */ 
-231    .proc_open = module_open,     /* called when the /proc file is opened */ 
-232    .proc_release = module_close, /* called when it's closed */ 
-233}; 
-234#else 
-235static const struct file_operations File_Ops_4_Our_Proc_File = { 
-236    .read = module_output, 
-237    .write = module_input, 
-238    .open = module_open, 
-239    .release = module_close, 
-240}; 
-241#endif 
-242 
-243/* 
-244 * Module initialization and cleanup 
-245 */ 
-246 
-247/* 
-248 * Initialize the module - register the proc file 
-249 */ 
-250 
-251int init_module() 
-252{ 
-253    Our_Proc_File = 
-254        proc_create(PROC_ENTRY_FILENAME, 0644, NULL, &File_Ops_4_Our_Proc_File); 
-255    if (Our_Proc_File == NULL) { 
-256        remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
-257        pr_debug("Error: Could not initialize /proc/%s\n", PROC_ENTRY_FILENAME); 
-258        return -ENOMEM; 
-259    } 
-260    proc_set_size(Our_Proc_File, 80); 
-261    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
-262 
-263    pr_info("/proc/test created\n"); 
-264 
-265    return 0; 
-266} 
-267 
-268/* 
-269 * Cleanup - unregister our file from /proc.  This could get dangerous if 
-270 * there are still processes waiting in WaitQ, because they are inside our 
-271 * open function, which will get unloaded. I'll explain how to avoid removal 
-272 * of a kernel module in such a case in chapter 10. 
-273 */ 
-274void cleanup_module() 
-275{ 
-276    remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
-277    pr_debug("/proc/%s removed\n", PROC_ENTRY_FILENAME); 
-278} 
-279 
-280MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  sleep.c - create a /proc file, and if several processes try to open it at 
+3 *  the same time, put all but one to sleep 
+4 */ 
+5 
+6#include <linux/kernel.h>  /* We're doing kernel work */ 
+7#include <linux/module.h>  /* Specifically, a module */ 
+8#include <linux/proc_fs.h> /* Necessary because we use proc fs */ 
+9#include <linux/sched.h>   /* For putting processes to sleep and 
+10                                   waking them up */  
+11#include <linux/uaccess.h> /* for get_user and put_user */ 
+12#include <linux/version.h> 
+13 
+14#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+15#define HAVE_PROC_OPS 
+16#endif 
+17 
+18/* 
+19 * The module's file functions 
+20 */ 
+21 
+22/* 
+23 * Here we keep the last message received, to prove that we can process our 
+24 * input 
+25 */ 
+26#define MESSAGE_LENGTH 80 
+27static char Message[MESSAGE_LENGTH]; 
+28 
+29static struct proc_dir_entry *Our_Proc_File; 
+30#define PROC_ENTRY_FILENAME "sleep" 
+31 
+32/* 
+33 * Since we use the file operations struct, we can't use the special proc 
+34 * output provisions - we have to use a standard read function, which is this 
+35 * function 
+36 */ 
+37static ssize_t module_output(struct file *file, /* see include/linux/fs.h   */ 
+38                             char *buf,         /* The buffer to put data to 
+39                                                   (in the user segment)    */  
+40                             size_t len,        /* The length of the buffer */ 
+41                             loff_t *offset) 
+42{ 
+43    static int finished = 0; 
+44    int i; 
+45    char message[MESSAGE_LENGTH + 30]; 
+46 
+47    /* 
+48     * Return 0 to signify end of file - that we have nothing 
+49     * more to say at this point. 
+50     */ 
+51    if (finished) { 
+52        finished = 0; 
+53        return 0; 
+54    } 
+55 
+56    /* 
+57     * If you don't understand this by now, you're hopeless as a kernel 
+58     * programmer. 
+59     */ 
+60    sprintf(message, "Last input:%s\n", Message); 
+61    for (i = 0; i < len && message[i]; i++) 
+62        put_user(message[i], buf + i); 
+63 
+64    finished = 1; 
+65    return i; /* Return the number of bytes "read" */ 
+66} 
+67 
+68/* 
+69 * This function receives input from the user when the user writes to the /proc 
+70 * file. 
+71 */ 
+72static ssize_t module_input(struct file *file, /* The file itself */ 
+73                            const char *buf,   /* The buffer with input */ 
+74                            size_t length,     /* The buffer's length */ 
+75                            loff_t *offset)    /* offset to file - ignore */ 
+76{ 
+77    int i; 
+78 
+79    /* 
+80     * Put the input into Message, where module_output will later be 
+81     * able to use it 
+82     */ 
+83    for (i = 0; i < MESSAGE_LENGTH - 1 && i < length; i++) 
+84        get_user(Message[i], buf + i); 
+85    /* 
+86     * we want a standard, zero terminated string 
+87     */ 
+88    Message[i] = '\0'; 
+89 
+90    /* 
+91     * We need to return the number of input characters used 
+92     */ 
+93    return i; 
+94} 
+95 
+96/* 
+97 * 1 if the file is currently open by somebody 
+98 */ 
+99int Already_Open = 0; 
+100 
+101/* 
+102 * Queue of processes who want our file 
+103 */ 
+104DECLARE_WAIT_QUEUE_HEAD(WaitQ); 
+105/* 
+106 * Called when the /proc file is opened 
+107 */ 
+108static int module_open(struct inode *inode, struct file *file) 
+109{ 
+110    /* 
+111     * If the file's flags include O_NONBLOCK, it means the process doesn't 
+112     * want to wait for the file.  In this case, if the file is already 
+113     * open, we should fail with -EAGAIN, meaning "you'll have to try 
+114     * again", instead of blocking a process which would rather stay awake. 
+115     */ 
+116    if ((file->f_flags & O_NONBLOCK) && Already_Open) 
+117        return -EAGAIN; 
+118 
+119    /* 
+120     * This is the correct place for try_module_get(THIS_MODULE) because 
+121     * if a process is in the loop, which is within the kernel module, 
+122     * the kernel module must not be removed. 
+123     */ 
+124    try_module_get(THIS_MODULE); 
+125 
+126    /* 
+127     * If the file is already open, wait until it isn't 
+128     */ 
+129 
+130    while (Already_Open) { 
+131        int i, is_sig = 0; 
+132 
+133        /* 
+134         * This function puts the current process, including any system 
+135         * calls, such as us, to sleep.  Execution will be resumed right 
+136         * after the function call, either because somebody called 
+137         * wake_up(&WaitQ) (only module_close does that, when the file 
+138         * is closed) or when a signal, such as Ctrl-C, is sent 
+139         * to the process 
+140         */ 
+141        wait_event_interruptible(WaitQ, !Already_Open); 
+142 
+143        /* 
+144         * If we woke up because we got a signal we're not blocking, 
+145         * return -EINTR (fail the system call).  This allows processes 
+146         * to be killed or stopped. 
+147         */ 
+148 
+149        /* 
+150         * Emmanuel Papirakis: 
+151         * 
+152         * This is a little update to work with 2.2.*.  Signals now are 
+153         * contained in two words (64 bits) and are stored in a structure that 
+154         * contains an array of two unsigned longs.  We now have to make 2 
+155         * checks in our if. 
+156         * 
+157         * Ori Pomerantz: 
+158         * 
+159         * Nobody promised me they'll never use more than 64 bits, or that this 
+160         * book won't be used for a version of Linux with a word size of 16 
+161         * bits.  This code would work in any case. 
+162         */ 
+163        for (i = 0; i < _NSIG_WORDS && !is_sig; i++) 
+164            is_sig = current->pending.signal.sig[i] & ~current->blocked.sig[i]; 
+165 
+166        if (is_sig) { 
+167            /* 
+168             * It's important to put module_put(THIS_MODULE) here, 
+169             * because for processes where the open is interrupted 
+170             * there will never be a corresponding close. If we 
+171             * don't decrement the usage count here, we will be 
+172             * left with a positive usage count which we'll have no 
+173             * way to bring down to zero, giving us an immortal 
+174             * module, which can only be killed by rebooting 
+175             * the machine. 
+176             */ 
+177            module_put(THIS_MODULE); 
+178            return -EINTR; 
+179        } 
+180    } 
+181 
+182    /* 
+183     * If we got here, Already_Open must be zero 
+184     */ 
+185 
+186    /* 
+187     * Open the file 
+188     */ 
+189    Already_Open = 1; 
+190    return 0; /* Allow the access */ 
+191} 
+192 
+193/* 
+194 * Called when the /proc file is closed 
+195 */ 
+196int module_close(struct inode *inode, struct file *file) 
+197{ 
+198    /* 
+199     * Set Already_Open to zero, so one of the processes in the WaitQ will 
+200     * be able to set Already_Open back to one and to open the file. All 
+201     * the other processes will be called when Already_Open is back to one, 
+202     * so they'll go back to sleep. 
+203     */ 
+204    Already_Open = 0; 
+205 
+206    /* 
+207     * Wake up all the processes in WaitQ, so if anybody is waiting for the 
+208     * file, they can have it. 
+209     */ 
+210    wake_up(&WaitQ); 
+211 
+212    module_put(THIS_MODULE); 
+213 
+214    return 0; /* success */ 
+215} 
+216 
+217/* 
+218 * Structures to register as the /proc file, with pointers to all the relevant 
+219 * functions. 
+220 */ 
+221 
+222/* 
+223 * File operations for our proc file. This is where we place pointers to all 
+224 * the functions called when somebody tries to do something to our file. NULL 
+225 * means we don't want to deal with something. 
+226 */ 
+227#ifdef HAVE_PROC_OPS 
+228static const struct proc_ops File_Ops_4_Our_Proc_File = { 
+229    .proc_read = module_output,   /* "read" from the file */ 
+230    .proc_write = module_input,   /* "write" to the file */ 
+231    .proc_open = module_open,     /* called when the /proc file is opened */ 
+232    .proc_release = module_close, /* called when it's closed */ 
+233}; 
+234#else 
+235static const struct file_operations File_Ops_4_Our_Proc_File = { 
+236    .read = module_output, 
+237    .write = module_input, 
+238    .open = module_open, 
+239    .release = module_close, 
+240}; 
+241#endif 
+242 
+243/* 
+244 * Module initialization and cleanup 
+245 */ 
+246 
+247/* 
+248 * Initialize the module - register the proc file 
+249 */ 
+250 
+251int init_module() 
+252{ 
+253    Our_Proc_File = 
+254        proc_create(PROC_ENTRY_FILENAME, 0644, NULL, &File_Ops_4_Our_Proc_File); 
+255    if (Our_Proc_File == NULL) { 
+256        remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
+257        pr_debug("Error: Could not initialize /proc/%s\n", PROC_ENTRY_FILENAME); 
+258        return -ENOMEM; 
+259    } 
+260    proc_set_size(Our_Proc_File, 80); 
+261    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
+262 
+263    pr_info("/proc/test created\n"); 
+264 
+265    return 0; 
+266} 
+267 
+268/* 
+269 * Cleanup - unregister our file from /proc.  This could get dangerous if 
+270 * there are still processes waiting in WaitQ, because they are inside our 
+271 * open function, which will get unloaded. I'll explain how to avoid removal 
+272 * of a kernel module in such a case in chapter 10. 
+273 */ 
+274void cleanup_module() 
+275{ 
+276    remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
+277    pr_debug("/proc/%s removed\n", PROC_ENTRY_FILENAME); 
+278} 
+279 
+280MODULE_LICENSE("GPL");

    - 
    1/* 
-2 *  cat_nonblock.c - open a file and display its contents, but exit rather than 
-3 *  wait for input. 
-4 */ 
-5#include <errno.h>  /* for errno */ 
-6#include <fcntl.h>  /* for open */ 
-7#include <stdio.h>  /* standard I/O */ 
-8#include <stdlib.h> /* for exit */ 
-9#include <unistd.h> /* for read */ 
-10 
-11#define MAX_BYTES 1024 * 4 
-12 
-13int main(int argc, char *argv[]) 
-14{ 
-15    int fd;                 /* The file descriptor for the file to read */ 
-16    size_t bytes;           /* The number of bytes read */ 
-17    char buffer[MAX_BYTES]; /* The buffer for the bytes */ 
-18 
-19    /* Usage */ 
-20    if (argc != 2) { 
-21        printf("Usage: %s <filename>\n", argv[0]); 
-22        puts("Reads the content of a file, but doesn't wait for input"); 
-23        exit(-1); 
-24    } 
-25 
-26    /* Open the file for reading in non blocking mode */ 
-27    fd = open(argv[1], O_RDONLY | O_NONBLOCK); 
-28 
-29    /* If open failed */ 
-30    if (fd == -1) { 
-31        puts(errno == EAGAIN ? "Open would block" : "Open failed"); 
-32        exit(-1); 
-33    } 
-34 
-35    /* Read the file and output its contents */ 
-36    do { 
-37        /* Read characters from the file */ 
-38        bytes = read(fd, buffer, MAX_BYTES); 
-39 
-40        /* If there's an error, report it and die */ 
-41        if (bytes == -1) { 
-42            if (errno = EAGAIN) 
-43                puts("Normally I'd block, but you told me not to"); 
-44            else 
-45                puts("Another read error"); 
-46            exit(-1); 
-47        } 
-48 
-49        /* Print the characters */ 
-50        if (bytes > 0) { 
-51            for (int i = 0; i < bytes; i++) 
-52                putchar(buffer[i]); 
-53        } 
-54 
-55        /* While there are no errors and the file isn't over */ 
-56    } while (bytes > 0); 
-57 
-58    return 0; 
-59}
    -

    + 

    1/* 
+2 *  cat_nonblock.c - open a file and display its contents, but exit rather than 
+3 *  wait for input. 
+4 */ 
+5#include <errno.h>  /* for errno */ 
+6#include <fcntl.h>  /* for open */ 
+7#include <stdio.h>  /* standard I/O */ 
+8#include <stdlib.h> /* for exit */ 
+9#include <unistd.h> /* for read */ 
+10 
+11#define MAX_BYTES 1024 * 4 
+12 
+13int main(int argc, char *argv[]) 
+14{ 
+15    int fd;                 /* The file descriptor for the file to read */ 
+16    size_t bytes;           /* The number of bytes read */ 
+17    char buffer[MAX_BYTES]; /* The buffer for the bytes */ 
+18 
+19    /* Usage */ 
+20    if (argc != 2) { 
+21        printf("Usage: %s <filename>\n", argv[0]); 
+22        puts("Reads the content of a file, but doesn't wait for input"); 
+23        exit(-1); 
+24    } 
+25 
+26    /* Open the file for reading in non blocking mode */ 
+27    fd = open(argv[1], O_RDONLY | O_NONBLOCK); 
+28 
+29    /* If open failed */ 
+30    if (fd == -1) { 
+31        puts(errno == EAGAIN ? "Open would block" : "Open failed"); 
+32        exit(-1); 
+33    } 
+34 
+35    /* Read the file and output its contents */ 
+36    do { 
+37        /* Read characters from the file */ 
+38        bytes = read(fd, buffer, MAX_BYTES); 
+39 
+40        /* If there's an error, report it and die */ 
+41        if (bytes == -1) { 
+42            if (errno = EAGAIN) 
+43                puts("Normally I'd block, but you told me not to"); 
+44            else 
+45                puts("Another read error"); 
+46            exit(-1); 
+47        } 
+48 
+49        /* Print the characters */ 
+50        if (bytes > 0) { 
+51            for (int i = 0; i < bytes; i++) 
+52                putchar(buffer[i]); 
+53        } 
+54 
+55        /* While there are no errors and the file isn't over */ 
+56    } while (bytes > 0); 
+57 
+58    return 0; 
+59}
    +

    -

    0.11.2 Completions

    -

    Sometimes one thing should happen before another within a module having multiple +

    0.11.2 Completions

    +

    Sometimes one thing should happen before another within a module having multiple threads. Rather than using /proc/sleep commands the kernel has another way to do this which allows timeouts or interrupts to also happen. -

    In the following example two threads are started, but one needs to start before +

    In the following example two threads are started, but one needs to start before another.

    - 
    1/* 
-2 *  completions.c 
-3 */ 
-4#include <linux/completion.h> 
-5#include <linux/init.h> 
-6#include <linux/kernel.h> 
-7#include <linux/kthread.h> 
-8#include <linux/module.h> 
-9 
-10static struct { 
-11    struct completion crank_comp; 
-12    struct completion flywheel_comp; 
-13} machine; 
-14 
-15static int machine_crank_thread(void *arg) 
-16{ 
-17    pr_info("Turn the crank\n"); 
-18 
-19    complete_all(&machine.crank_comp); 
-20    complete_and_exit(&machine.crank_comp, 0); 
-21} 
-22 
-23static int machine_flywheel_spinup_thread(void *arg) 
-24{ 
-25    wait_for_completion(&machine.crank_comp); 
-26 
-27    pr_info("Flywheel spins up\n"); 
-28 
-29    complete_all(&machine.flywheel_comp); 
-30    complete_and_exit(&machine.flywheel_comp, 0); 
-31} 
-32 
-33static int completions_init(void) 
-34{ 
-35    struct task_struct *crank_thread; 
-36    struct task_struct *flywheel_thread; 
-37 
-38    pr_info("completions example\n"); 
-39 
-40    init_completion(&machine.crank_comp); 
-41    init_completion(&machine.flywheel_comp); 
-42 
-43    crank_thread = kthread_create(machine_crank_thread, NULL, "KThread Crank"); 
-44    if (IS_ERR(crank_thread)) 
-45        goto ERROR_THREAD_1; 
-46 
-47    flywheel_thread = kthread_create(machine_flywheel_spinup_thread, NULL, 
-48                                     "KThread Flywheel"); 
-49    if (IS_ERR(flywheel_thread)) 
-50        goto ERROR_THREAD_2; 
-51 
-52    wake_up_process(flywheel_thread); 
-53    wake_up_process(crank_thread); 
-54 
-55    return 0; 
-56 
-57ERROR_THREAD_2: 
-58    kthread_stop(crank_thread); 
-59ERROR_THREAD_1: 
-60 
-61    return -1; 
-62} 
-63 
-64void completions_exit(void) 
-65{ 
-66    wait_for_completion(&machine.crank_comp); 
-67    wait_for_completion(&machine.flywheel_comp); 
-68 
-69    pr_info("completions exit\n"); 
-70} 
-71 
-72module_init(completions_init); 
-73module_exit(completions_exit); 
-74 
-75MODULE_DESCRIPTION("Completions example"); 
-76MODULE_LICENSE("GPL");
    -

    The machine structure stores the completion states for the two threads. At the + 

    1/* 
+2 *  completions.c 
+3 */ 
+4#include <linux/completion.h> 
+5#include <linux/init.h> 
+6#include <linux/kernel.h> 
+7#include <linux/kthread.h> 
+8#include <linux/module.h> 
+9 
+10static struct { 
+11    struct completion crank_comp; 
+12    struct completion flywheel_comp; 
+13} machine; 
+14 
+15static int machine_crank_thread(void *arg) 
+16{ 
+17    pr_info("Turn the crank\n"); 
+18 
+19    complete_all(&machine.crank_comp); 
+20    complete_and_exit(&machine.crank_comp, 0); 
+21} 
+22 
+23static int machine_flywheel_spinup_thread(void *arg) 
+24{ 
+25    wait_for_completion(&machine.crank_comp); 
+26 
+27    pr_info("Flywheel spins up\n"); 
+28 
+29    complete_all(&machine.flywheel_comp); 
+30    complete_and_exit(&machine.flywheel_comp, 0); 
+31} 
+32 
+33static int completions_init(void) 
+34{ 
+35    struct task_struct *crank_thread; 
+36    struct task_struct *flywheel_thread; 
+37 
+38    pr_info("completions example\n"); 
+39 
+40    init_completion(&machine.crank_comp); 
+41    init_completion(&machine.flywheel_comp); 
+42 
+43    crank_thread = kthread_create(machine_crank_thread, NULL, "KThread Crank"); 
+44    if (IS_ERR(crank_thread)) 
+45        goto ERROR_THREAD_1; 
+46 
+47    flywheel_thread = kthread_create(machine_flywheel_spinup_thread, NULL, 
+48                                     "KThread Flywheel"); 
+49    if (IS_ERR(flywheel_thread)) 
+50        goto ERROR_THREAD_2; 
+51 
+52    wake_up_process(flywheel_thread); 
+53    wake_up_process(crank_thread); 
+54 
+55    return 0; 
+56 
+57ERROR_THREAD_2: 
+58    kthread_stop(crank_thread); 
+59ERROR_THREAD_1: 
+60 
+61    return -1; 
+62} 
+63 
+64void completions_exit(void) 
+65{ 
+66    wait_for_completion(&machine.crank_comp); 
+67    wait_for_completion(&machine.flywheel_comp); 
+68 
+69    pr_info("completions exit\n"); 
+70} 
+71 
+72module_init(completions_init); 
+73module_exit(completions_exit); 
+74 
+75MODULE_DESCRIPTION("Completions example"); 
+76MODULE_LICENSE("GPL");
    +

    The machine structure stores the completion states for the two threads. At the exit point of each thread the respective completion state is updated, and wait_for_completion is used by the flywheel thread to ensure that it doesn’t begin prematurely. -

    So even though flywheel_thread is started first you should notice if you load this +

    So even though flywheel_thread is started first you should notice if you load this module and run dmesg that turning the crank always happens first because the flywheel thread waits for it to complete. -

    There are other variations upon the wait_for_completion function, which include +

    There are other variations upon the wait_for_completion function, which include timeouts or being interrupted, but this basic mechanism is enough for many common situations without adding a lot of complexity. -

    +

    -

    0.12 Avoiding Collisions and Deadlocks

    -

    If processes running on different CPUs or in different threads try to access the same +

    0.12 Avoiding Collisions and Deadlocks

    +

    If processes running on different CPUs or in different threads try to access the same memory then it’s possible that strange things can happen or your system can lock up. To avoid this various types of mutual exclusion kernel functions are available. These indicate if a section of code is "locked" or "unlocked" so that simultaneous attempts to run it can’t happen.

    -

    0.12.1 Mutex

    -

    You can use kernel mutexes (mutual exclusions) in much the same manner that you +

    0.12.1 Mutex

    +

    You can use kernel mutexes (mutual exclusions) in much the same manner that you might deploy them in userland. This may be all that’s needed to avoid collisions in most cases.

    - 
    1/* 
-2 *  example_mutex.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/mutex.h> 
-8 
-9DEFINE_MUTEX(mymutex); 
-10 
-11static int example_mutex_init(void) 
-12{ 
-13    int ret; 
-14 
-15    pr_info("example_mutex init\n"); 
-16 
-17    ret = mutex_trylock(&mymutex); 
-18    if (ret != 0) { 
-19        pr_info("mutex is locked\n"); 
-20 
-21        if (mutex_is_locked(&mymutex) == 0) 
-22            pr_info("The mutex failed to lock!\n"); 
-23 
-24        mutex_unlock(&mymutex); 
-25        pr_info("mutex is unlocked\n"); 
-26    } else 
-27        pr_info("Failed to lock\n"); 
-28 
-29    return 0; 
-30} 
-31 
-32static void example_mutex_exit(void) 
-33{ 
-34    pr_info("example_mutex exit\n"); 
-35} 
-36 
-37module_init(example_mutex_init); 
-38module_exit(example_mutex_exit); 
-39 
-40MODULE_DESCRIPTION("Mutex example"); 
-41MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  example_mutex.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/mutex.h> 
+8 
+9DEFINE_MUTEX(mymutex); 
+10 
+11static int example_mutex_init(void) 
+12{ 
+13    int ret; 
+14 
+15    pr_info("example_mutex init\n"); 
+16 
+17    ret = mutex_trylock(&mymutex); 
+18    if (ret != 0) { 
+19        pr_info("mutex is locked\n"); 
+20 
+21        if (mutex_is_locked(&mymutex) == 0) 
+22            pr_info("The mutex failed to lock!\n"); 
+23 
+24        mutex_unlock(&mymutex); 
+25        pr_info("mutex is unlocked\n"); 
+26    } else 
+27        pr_info("Failed to lock\n"); 
+28 
+29    return 0; 
+30} 
+31 
+32static void example_mutex_exit(void) 
+33{ 
+34    pr_info("example_mutex exit\n"); 
+35} 
+36 
+37module_init(example_mutex_init); 
+38module_exit(example_mutex_exit); 
+39 
+40MODULE_DESCRIPTION("Mutex example"); 
+41MODULE_LICENSE("GPL");
    +

    -

    0.12.2 Spinlocks

    -

    As the name suggests, spinlocks lock up the CPU that the code is running on, +

    0.12.2 Spinlocks

    +

    As the name suggests, spinlocks lock up the CPU that the code is running on, taking 100% of its resources. Because of this you should only use the spinlock mechanism around code which is likely to take no more than a few milliseconds to run and so won’t noticably slow anything down from the user’s point of @@ -3704,80 +3729,80 @@ view. -

    The example here is "irq safe" in that if interrupts happen during the lock then +

    The example here is "irq safe" in that if interrupts happen during the lock then they won’t be forgotten and will activate when the unlock happens, using the flags variable to retain their state.

    - 
    1/* 
-2 *  example_spinlock.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/interrupt.h> 
-6#include <linux/kernel.h> 
-7#include <linux/module.h> 
-8#include <linux/spinlock.h> 
-9 
-10DEFINE_SPINLOCK(sl_static); 
-11spinlock_t sl_dynamic; 
-12 
-13static void example_spinlock_static(void) 
-14{ 
-15    unsigned long flags; 
-16 
-17    spin_lock_irqsave(&sl_static, flags); 
-18    pr_info("Locked static spinlock\n"); 
-19 
-20    /* Do something or other safely. 
-21       Because this uses 100% CPU time this 
-22       code should take no more than a few 
-23       milliseconds to run */ 
-24 
-25    spin_unlock_irqrestore(&sl_static, flags); 
-26    pr_info("Unlocked static spinlock\n"); 
-27} 
-28 
-29static void example_spinlock_dynamic(void) 
-30{ 
-31    unsigned long flags; 
-32 
-33    spin_lock_init(&sl_dynamic); 
-34    spin_lock_irqsave(&sl_dynamic, flags); 
-35    pr_info("Locked dynamic spinlock\n"); 
-36 
-37    /* Do something or other safely. 
-38       Because this uses 100% CPU time this 
-39       code should take no more than a few 
-40       milliseconds to run */ 
-41 
-42    spin_unlock_irqrestore(&sl_dynamic, flags); 
-43    pr_info("Unlocked dynamic spinlock\n"); 
-44} 
-45 
-46static int example_spinlock_init(void) 
-47{ 
-48    pr_info("example spinlock started\n"); 
-49 
-50    example_spinlock_static(); 
-51    example_spinlock_dynamic(); 
-52 
-53    return 0; 
-54} 
-55 
-56static void example_spinlock_exit(void) 
-57{ 
-58    pr_info("example spinlock exit\n"); 
-59} 
-60 
-61module_init(example_spinlock_init); 
-62module_exit(example_spinlock_exit); 
-63 
-64MODULE_DESCRIPTION("Spinlock example"); 
-65MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  example_spinlock.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/interrupt.h> 
+6#include <linux/kernel.h> 
+7#include <linux/module.h> 
+8#include <linux/spinlock.h> 
+9 
+10DEFINE_SPINLOCK(sl_static); 
+11spinlock_t sl_dynamic; 
+12 
+13static void example_spinlock_static(void) 
+14{ 
+15    unsigned long flags; 
+16 
+17    spin_lock_irqsave(&sl_static, flags); 
+18    pr_info("Locked static spinlock\n"); 
+19 
+20    /* Do something or other safely. 
+21       Because this uses 100% CPU time this 
+22       code should take no more than a few 
+23       milliseconds to run */ 
+24 
+25    spin_unlock_irqrestore(&sl_static, flags); 
+26    pr_info("Unlocked static spinlock\n"); 
+27} 
+28 
+29static void example_spinlock_dynamic(void) 
+30{ 
+31    unsigned long flags; 
+32 
+33    spin_lock_init(&sl_dynamic); 
+34    spin_lock_irqsave(&sl_dynamic, flags); 
+35    pr_info("Locked dynamic spinlock\n"); 
+36 
+37    /* Do something or other safely. 
+38       Because this uses 100% CPU time this 
+39       code should take no more than a few 
+40       milliseconds to run */ 
+41 
+42    spin_unlock_irqrestore(&sl_dynamic, flags); 
+43    pr_info("Unlocked dynamic spinlock\n"); 
+44} 
+45 
+46static int example_spinlock_init(void) 
+47{ 
+48    pr_info("example spinlock started\n"); 
+49 
+50    example_spinlock_static(); 
+51    example_spinlock_dynamic(); 
+52 
+53    return 0; 
+54} 
+55 
+56static void example_spinlock_exit(void) 
+57{ 
+58    pr_info("example spinlock exit\n"); 
+59} 
+60 
+61module_init(example_spinlock_init); 
+62module_exit(example_spinlock_exit); 
+63 
+64MODULE_DESCRIPTION("Spinlock example"); 
+65MODULE_LICENSE("GPL");
    +

    -

    0.12.3 Read and write locks

    -

    Read and write locks are specialised kinds of spinlocks so that you can exclusively +

    0.12.3 Read and write locks

    +

    Read and write locks are specialised kinds of spinlocks so that you can exclusively read from something or write to something. Like the earlier spinlocks example the one below shows an "irq safe" situation in which if other functions were triggered from irqs which might also read and write to whatever you are concerned @@ -3787,68 +3812,68 @@ the system and cause users to start revolting against the tyranny of your module.

    - 
    1/* 
-2 *  example_rwlock.c 
-3 */ 
-4#include <linux/interrupt.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7 
-8DEFINE_RWLOCK(myrwlock); 
-9 
-10static void example_read_lock(void) 
-11{ 
-12    unsigned long flags; 
-13 
-14    read_lock_irqsave(&myrwlock, flags); 
-15    pr_info("Read Locked\n"); 
-16 
-17    /* Read from something */ 
-18 
-19    read_unlock_irqrestore(&myrwlock, flags); 
-20    pr_info("Read Unlocked\n"); 
-21} 
-22 
-23static void example_write_lock(void) 
-24{ 
-25    unsigned long flags; 
-26 
-27    write_lock_irqsave(&myrwlock, flags); 
-28    pr_info("Write Locked\n"); 
-29 
-30    /* Write to something */ 
-31 
-32    write_unlock_irqrestore(&myrwlock, flags); 
-33    pr_info("Write Unlocked\n"); 
-34} 
-35 
-36static int example_rwlock_init(void) 
-37{ 
-38    pr_info("example_rwlock started\n"); 
-39 
-40    example_read_lock(); 
-41    example_write_lock(); 
-42 
-43    return 0; 
-44} 
-45 
-46static void example_rwlock_exit(void) 
-47{ 
-48    pr_info("example_rwlock exit\n"); 
-49} 
-50 
-51module_init(example_rwlock_init); 
-52module_exit(example_rwlock_exit); 
-53 
-54MODULE_DESCRIPTION("Read/Write locks example"); 
-55MODULE_LICENSE("GPL");
    -

    Of course if you know for sure that there are no functions triggered by irqs + 

    1/* 
+2 *  example_rwlock.c 
+3 */ 
+4#include <linux/interrupt.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7 
+8DEFINE_RWLOCK(myrwlock); 
+9 
+10static void example_read_lock(void) 
+11{ 
+12    unsigned long flags; 
+13 
+14    read_lock_irqsave(&myrwlock, flags); 
+15    pr_info("Read Locked\n"); 
+16 
+17    /* Read from something */ 
+18 
+19    read_unlock_irqrestore(&myrwlock, flags); 
+20    pr_info("Read Unlocked\n"); 
+21} 
+22 
+23static void example_write_lock(void) 
+24{ 
+25    unsigned long flags; 
+26 
+27    write_lock_irqsave(&myrwlock, flags); 
+28    pr_info("Write Locked\n"); 
+29 
+30    /* Write to something */ 
+31 
+32    write_unlock_irqrestore(&myrwlock, flags); 
+33    pr_info("Write Unlocked\n"); 
+34} 
+35 
+36static int example_rwlock_init(void) 
+37{ 
+38    pr_info("example_rwlock started\n"); 
+39 
+40    example_read_lock(); 
+41    example_write_lock(); 
+42 
+43    return 0; 
+44} 
+45 
+46static void example_rwlock_exit(void) 
+47{ 
+48    pr_info("example_rwlock exit\n"); 
+49} 
+50 
+51module_init(example_rwlock_init); 
+52module_exit(example_rwlock_exit); 
+53 
+54MODULE_DESCRIPTION("Read/Write locks example"); 
+55MODULE_LICENSE("GPL");
    +

    Of course if you know for sure that there are no functions triggered by irqs which could possibly interfere with your logic then you can use the simpler read_lock(&myrwlock) and read_unlock(&myrwlock) or the corresponding write functions.

    -

    0.12.4 Atomic operations

    -

    If you’re doing simple arithmetic: adding, subtracting or bitwise operations then +

    0.12.4 Atomic operations

    +

    If you’re doing simple arithmetic: adding, subtracting or bitwise operations then there’s another way in the multi-CPU and multi-hyperthreaded world to stop other parts of the system from messing with your mojo. By using atomic operations you can be confident that your addition, subtraction or bit flip did actually happen @@ -3856,293 +3881,293 @@ and wasn’t overwritten by some other shenanigans. An example is shown below.

    - 
    1/* 
-2 *  example_atomic.c 
-3 */ 
-4#include <linux/interrupt.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7 
-8#define BYTE_TO_BINARY_PATTERN "%c%c%c%c%c%c%c%c" 
-9#define BYTE_TO_BINARY(byte)                                  \ 
-10    (byte & 0x80 ? '1' : '0'), (byte & 0x40 ? '1' : '0'),     \ 
-11        (byte & 0x20 ? '1' : '0'), (byte & 0x10 ? '1' : '0'), \ 
-12        (byte & 0x08 ? '1' : '0'), (byte & 0x04 ? '1' : '0'), \ 
-13        (byte & 0x02 ? '1' : '0'), (byte & 0x01 ? '1' : '0') 
-14 
-15static void atomic_add_subtract(void) 
-16{ 
-17    atomic_t debbie; 
-18    atomic_t chris = ATOMIC_INIT(50); 
-19 
-20    atomic_set(&debbie, 45); 
-21 
-22    /* subtract one */ 
-23    atomic_dec(&debbie); 
-24 
-25    atomic_add(7, &debbie); 
-26 
-27    /* add one */ 
-28    atomic_inc(&debbie); 
-29 
-30    pr_info("chris: %d, debbie: %d\n", atomic_read(&chris), 
-31            atomic_read(&debbie)); 
-32} 
-33 
-34static void atomic_bitwise(void) 
-35{ 
-36    unsigned long word = 0; 
-37 
-38    pr_info("Bits 0: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-39    set_bit(3, &word); 
-40    set_bit(5, &word); 
-41    pr_info("Bits 1: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-42    clear_bit(5, &word); 
-43    pr_info("Bits 2: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-44    change_bit(3, &word); 
-45 
-46    pr_info("Bits 3: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-47    if (test_and_set_bit(3, &word)) 
-48        pr_info("wrong\n"); 
-49    pr_info("Bits 4: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-50 
-51    word = 255; 
-52    pr_info("Bits 5: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-53} 
-54 
-55static int example_atomic_init(void) 
-56{ 
-57    pr_info("example_atomic started\n"); 
-58 
-59    atomic_add_subtract(); 
-60    atomic_bitwise(); 
-61 
-62    return 0; 
-63} 
-64 
-65static void example_atomic_exit(void) 
-66{ 
-67    pr_info("example_atomic exit\n"); 
-68} 
-69 
-70module_init(example_atomic_init); 
-71module_exit(example_atomic_exit); 
-72 
-73MODULE_DESCRIPTION("Atomic operations example"); 
-74MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  example_atomic.c 
+3 */ 
+4#include <linux/interrupt.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7 
+8#define BYTE_TO_BINARY_PATTERN "%c%c%c%c%c%c%c%c" 
+9#define BYTE_TO_BINARY(byte)                                  \ 
+10    (byte & 0x80 ? '1' : '0'), (byte & 0x40 ? '1' : '0'),     \ 
+11        (byte & 0x20 ? '1' : '0'), (byte & 0x10 ? '1' : '0'), \ 
+12        (byte & 0x08 ? '1' : '0'), (byte & 0x04 ? '1' : '0'), \ 
+13        (byte & 0x02 ? '1' : '0'), (byte & 0x01 ? '1' : '0') 
+14 
+15static void atomic_add_subtract(void) 
+16{ 
+17    atomic_t debbie; 
+18    atomic_t chris = ATOMIC_INIT(50); 
+19 
+20    atomic_set(&debbie, 45); 
+21 
+22    /* subtract one */ 
+23    atomic_dec(&debbie); 
+24 
+25    atomic_add(7, &debbie); 
+26 
+27    /* add one */ 
+28    atomic_inc(&debbie); 
+29 
+30    pr_info("chris: %d, debbie: %d\n", atomic_read(&chris), 
+31            atomic_read(&debbie)); 
+32} 
+33 
+34static void atomic_bitwise(void) 
+35{ 
+36    unsigned long word = 0; 
+37 
+38    pr_info("Bits 0: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+39    set_bit(3, &word); 
+40    set_bit(5, &word); 
+41    pr_info("Bits 1: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+42    clear_bit(5, &word); 
+43    pr_info("Bits 2: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+44    change_bit(3, &word); 
+45 
+46    pr_info("Bits 3: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+47    if (test_and_set_bit(3, &word)) 
+48        pr_info("wrong\n"); 
+49    pr_info("Bits 4: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+50 
+51    word = 255; 
+52    pr_info("Bits 5: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+53} 
+54 
+55static int example_atomic_init(void) 
+56{ 
+57    pr_info("example_atomic started\n"); 
+58 
+59    atomic_add_subtract(); 
+60    atomic_bitwise(); 
+61 
+62    return 0; 
+63} 
+64 
+65static void example_atomic_exit(void) 
+66{ 
+67    pr_info("example_atomic exit\n"); 
+68} 
+69 
+70module_init(example_atomic_init); 
+71module_exit(example_atomic_exit); 
+72 
+73MODULE_DESCRIPTION("Atomic operations example"); 
+74MODULE_LICENSE("GPL");
    -

    +

    -

    0.13 Replacing Print Macros

    -

    +

    0.13 Replacing Print Macros

    +

    -

    0.13.1 Replacement

    -

    In Section 1.2.1.2, I said that X and kernel module programming don’t mix. That’s +

    0.13.1 Replacement

    +

    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 the command to load the module came from. -

    "tty" is an abbreviation of teletype: originally a combination keyboard-printer +

    "tty" is an abbreviation of teletype: originally a combination keyboard-printer used to communicate with a Unix system, and today an abstraction for the text stream used for a Unix program, whether it’s a physical terminal, an xterm on an X display, a network connection used with ssh, etc. -

    The way this is done is by using current, a pointer to the currently running task, +

    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.

    - 
    1/* 
-2 *  print_string.c - Send output to the tty we're running on, regardless if it's 
-3 *  through X11, telnet, etc.  We do this by printing the string to the tty 
-4 *  associated with the current task. 
-5 */ 
-6#include <linux/init.h> 
-7#include <linux/kernel.h> 
-8#include <linux/module.h> 
-9#include <linux/sched.h> /* For current */ 
-10#include <linux/tty.h>   /* For the tty declarations */ 
-11 
-12MODULE_LICENSE("GPL"); 
-13 
-14static void print_string(char *str) 
-15{ 
-16    struct tty_struct *my_tty; 
-17    const struct tty_operations *ttyops; 
-18 
-19    /* 
-20     * The tty for the current task, for 2.6.6+ kernels 
-21     */ 
-22    my_tty = get_current_tty(); 
-23    ttyops = my_tty->driver->ops; 
-24 
-25    /* 
-26     * If my_tty is NULL, the current task has no tty you can print to 
-27     * (ie, if it's a daemon).  If so, there's nothing we can do. 
-28     */ 
-29    if (my_tty != NULL) { 
-30        /* 
-31         * my_tty->driver is a struct which holds the tty's functions, 
-32         * one of which (write) is used to write strings to the tty. 
-33         * It can be used to take a string either from the user's or 
-34         * kernel's memory segment. 
-35         * 
-36         * The function's 1st parameter is the tty to write to, 
-37         * because the same function would normally be used for all 
-38         * tty's of a certain type. 
-39         * The 2nd parameter is a pointer to a string. 
-40         * The 3rd parameter is the length of the string. 
-41         * 
-42         * As you will see below, sometimes it's necessary to use 
-43         * preprocessor stuff to create code that works for different 
-44         * kernel versions. The (naive) approach we've taken here 
-45         * does not scale well. The right way to deal with this 
-46         * is described in section 2 of 
-47         * linux/Documentation/SubmittingPatches 
-48         */ 
-49        (ttyops->write)(my_tty,       /* The tty itself */ 
-50                        str,          /* String                 */ 
-51                        strlen(str)); /* Length */ 
-52 
-53        /* 
-54         * ttys were originally hardware devices, which (usually) 
-55         * strictly followed the ASCII standard.  In ASCII, to move to 
-56         * a new line you need two characters, a carriage return and a 
-57         * line feed.  On Unix, the ASCII line feed is used for both 
-58         * purposes - so we can't just use \n, because it wouldn't have 
-59         * a carriage return and the next line will start at the 
-60         * column right after the line feed. 
-61         * 
-62         * This is why text files are different between Unix and 
-63         * MS Windows.  In CP/M and derivatives, like MS-DOS and 
-64         * MS Windows, the ASCII standard was strictly adhered to, 
-65         * and therefore a newline requirs both a LF and a CR. 
-66         */ 
-67        (ttyops->write)(my_tty, "\015\012", 2); 
-68    } 
-69} 
-70 
-71static int __init print_string_init(void) 
-72{ 
-73    print_string("The module has been inserted.  Hello world!"); 
-74    return 0; 
-75} 
-76 
-77static void __exit print_string_exit(void) 
-78{ 
-79    print_string("The module has been removed.  Farewell world!"); 
-80} 
-81 
-82module_init(print_string_init); 
-83module_exit(print_string_exit);
    -

    + 

    1/* 
+2 *  print_string.c - Send output to the tty we're running on, regardless if it's 
+3 *  through X11, telnet, etc.  We do this by printing the string to the tty 
+4 *  associated with the current task. 
+5 */ 
+6#include <linux/init.h> 
+7#include <linux/kernel.h> 
+8#include <linux/module.h> 
+9#include <linux/sched.h> /* For current */ 
+10#include <linux/tty.h>   /* For the tty declarations */ 
+11 
+12MODULE_LICENSE("GPL"); 
+13 
+14static void print_string(char *str) 
+15{ 
+16    struct tty_struct *my_tty; 
+17    const struct tty_operations *ttyops; 
+18 
+19    /* 
+20     * The tty for the current task, for 2.6.6+ kernels 
+21     */ 
+22    my_tty = get_current_tty(); 
+23    ttyops = my_tty->driver->ops; 
+24 
+25    /* 
+26     * If my_tty is NULL, the current task has no tty you can print to 
+27     * (ie, if it's a daemon).  If so, there's nothing we can do. 
+28     */ 
+29    if (my_tty != NULL) { 
+30        /* 
+31         * my_tty->driver is a struct which holds the tty's functions, 
+32         * one of which (write) is used to write strings to the tty. 
+33         * It can be used to take a string either from the user's or 
+34         * kernel's memory segment. 
+35         * 
+36         * The function's 1st parameter is the tty to write to, 
+37         * because the same function would normally be used for all 
+38         * tty's of a certain type. 
+39         * The 2nd parameter is a pointer to a string. 
+40         * The 3rd parameter is the length of the string. 
+41         * 
+42         * As you will see below, sometimes it's necessary to use 
+43         * preprocessor stuff to create code that works for different 
+44         * kernel versions. The (naive) approach we've taken here 
+45         * does not scale well. The right way to deal with this 
+46         * is described in section 2 of 
+47         * linux/Documentation/SubmittingPatches 
+48         */ 
+49        (ttyops->write)(my_tty,       /* The tty itself */ 
+50                        str,          /* String                 */ 
+51                        strlen(str)); /* Length */ 
+52 
+53        /* 
+54         * ttys were originally hardware devices, which (usually) 
+55         * strictly followed the ASCII standard.  In ASCII, to move to 
+56         * a new line you need two characters, a carriage return and a 
+57         * line feed.  On Unix, the ASCII line feed is used for both 
+58         * purposes - so we can't just use \n, because it wouldn't have 
+59         * a carriage return and the next line will start at the 
+60         * column right after the line feed. 
+61         * 
+62         * This is why text files are different between Unix and 
+63         * MS Windows.  In CP/M and derivatives, like MS-DOS and 
+64         * MS Windows, the ASCII standard was strictly adhered to, 
+65         * and therefore a newline requirs both a LF and a CR. 
+66         */ 
+67        (ttyops->write)(my_tty, "\015\012", 2); 
+68    } 
+69} 
+70 
+71static int __init print_string_init(void) 
+72{ 
+73    print_string("The module has been inserted.  Hello world!"); 
+74    return 0; 
+75} 
+76 
+77static void __exit print_string_exit(void) 
+78{ 
+79    print_string("The module has been removed.  Farewell world!"); 
+80} 
+81 
+82module_init(print_string_init); 
+83module_exit(print_string_exit);
    +

    -

    0.13.2 Flashing keyboard LEDs

    -

    In certain conditions, you may desire a simpler and more direct way to communicate +

    0.13.2 Flashing keyboard LEDs

    +

    In certain conditions, you may desire a simpler and more direct way to communicate to the external world. 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, compared to writing to a tty or a file. -

    The following source code illustrates a minimal kernel module which, when +

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

    - 
    1/* 
-2 *  kbleds.c - Blink keyboard leds until the module is unloaded. 
-3 */ 
-4 
-5#include <linux/init.h> 
-6#include <linux/kd.h> /* For KDSETLED */ 
-7#include <linux/module.h> 
-8#include <linux/tty.h> /* For fg_console, MAX_NR_CONSOLES */ 
-9#include <linux/vt.h> 
-10#include <linux/vt_kern.h> /* for fg_console */ 
-11 
-12#include <linux/console_struct.h> /* For vc_cons */ 
-13 
-14MODULE_DESCRIPTION("Example module illustrating the use of Keyboard LEDs."); 
-15MODULE_LICENSE("GPL"); 
-16 
-17struct timer_list my_timer; 
-18struct tty_driver *my_driver; 
-19char kbledstatus = 0; 
-20 
-21#define BLINK_DELAY HZ / 5 
-22#define ALL_LEDS_ON 0x07 
-23#define RESTORE_LEDS 0xFF 
-24 
-25/* 
-26 * Function my_timer_func blinks the keyboard LEDs periodically by invoking 
-27 * command KDSETLED of ioctl() on the keyboard driver. To learn more on virtual 
-28 * terminal ioctl operations, please see file: 
-29 *     /usr/src/linux/drivers/char/vt_ioctl.c, function vt_ioctl(). 
-30 * 
-31 * The argument to KDSETLED is alternatively set to 7 (thus causing the led 
-32 * mode to be set to LED_SHOW_IOCTL, and all the leds are lit) and to 0xFF 
-33 * (any value above 7 switches back the led mode to LED_SHOW_FLAGS, thus 
-34 * the LEDs reflect the actual keyboard status).  To learn more on this, 
-35 * please see file: 
-36 *     /usr/src/linux/drivers/char/keyboard.c, function setledstate(). 
-37 * 
-38 */ 
-39 
-40static void my_timer_func(unsigned long ptr) 
-41{ 
-42    unsigned long *pstatus = (unsigned long *) ptr; 
-43    struct tty_struct *t = vc_cons[fg_console].d->port.tty; 
-44 
-45    if (*pstatus == ALL_LEDS_ON) 
-46        *pstatus = RESTORE_LEDS; 
-47    else 
-48        *pstatus = ALL_LEDS_ON; 
-49 
-50    (my_driver->ops->ioctl)(t, KDSETLED, *pstatus); 
-51 
-52    my_timer.expires = jiffies + BLINK_DELAY; 
-53    add_timer(&my_timer); 
-54} 
-55 
-56static int __init kbleds_init(void) 
-57{ 
-58    int i; 
-59 
-60    pr_info("kbleds: loading\n"); 
-61    pr_info("kbleds: fgconsole is %x\n", fg_console); 
-62    for (i = 0; i < MAX_NR_CONSOLES; i++) { 
-63        if (!vc_cons[i].d) 
-64            break; 
-65        pr_info("poet_atkm: console[%i/%i] #%i, tty %lx\n", i, MAX_NR_CONSOLES, 
-66                vc_cons[i].d->vc_num, (unsigned long) vc_cons[i].d->port.tty); 
-67    } 
-68    pr_info("kbleds: finished scanning consoles\n"); 
-69 
-70    my_driver = vc_cons[fg_console].d->port.tty->driver; 
-71    pr_info("kbleds: tty driver magic %x\n", my_driver->magic); 
-72 
-73    /* 
-74     * Set up the LED blink timer the first time 
-75     */ 
-76    timer_setup(&my_timer, (void *) &my_timer_func, 
-77                (unsigned long) &kbledstatus); 
-78    my_timer.expires = jiffies + BLINK_DELAY; 
-79    add_timer(&my_timer); 
-80 
-81    return 0; 
-82} 
-83 
-84static void __exit kbleds_cleanup(void) 
-85{ 
-86    pr_info("kbleds: unloading...\n"); 
-87    del_timer(&my_timer); 
-88    (my_driver->ops->ioctl)(vc_cons[fg_console].d->port.tty, KDSETLED, 
-89                            RESTORE_LEDS); 
-90} 
-91 
-92module_init(kbleds_init); 
-93module_exit(kbleds_cleanup);
    -

    If none of the examples in this chapter fit your debugging needs there might yet + 

    1/* 
+2 *  kbleds.c - Blink keyboard leds until the module is unloaded. 
+3 */ 
+4 
+5#include <linux/init.h> 
+6#include <linux/kd.h> /* For KDSETLED */ 
+7#include <linux/module.h> 
+8#include <linux/tty.h> /* For fg_console, MAX_NR_CONSOLES */ 
+9#include <linux/vt.h> 
+10#include <linux/vt_kern.h> /* for fg_console */ 
+11 
+12#include <linux/console_struct.h> /* For vc_cons */ 
+13 
+14MODULE_DESCRIPTION("Example module illustrating the use of Keyboard LEDs."); 
+15MODULE_LICENSE("GPL"); 
+16 
+17struct timer_list my_timer; 
+18struct tty_driver *my_driver; 
+19char kbledstatus = 0; 
+20 
+21#define BLINK_DELAY HZ / 5 
+22#define ALL_LEDS_ON 0x07 
+23#define RESTORE_LEDS 0xFF 
+24 
+25/* 
+26 * Function my_timer_func blinks the keyboard LEDs periodically by invoking 
+27 * command KDSETLED of ioctl() on the keyboard driver. To learn more on virtual 
+28 * terminal ioctl operations, please see file: 
+29 *     /usr/src/linux/drivers/char/vt_ioctl.c, function vt_ioctl(). 
+30 * 
+31 * The argument to KDSETLED is alternatively set to 7 (thus causing the led 
+32 * mode to be set to LED_SHOW_IOCTL, and all the leds are lit) and to 0xFF 
+33 * (any value above 7 switches back the led mode to LED_SHOW_FLAGS, thus 
+34 * the LEDs reflect the actual keyboard status).  To learn more on this, 
+35 * please see file: 
+36 *     /usr/src/linux/drivers/char/keyboard.c, function setledstate(). 
+37 * 
+38 */ 
+39 
+40static void my_timer_func(unsigned long ptr) 
+41{ 
+42    unsigned long *pstatus = (unsigned long *) ptr; 
+43    struct tty_struct *t = vc_cons[fg_console].d->port.tty; 
+44 
+45    if (*pstatus == ALL_LEDS_ON) 
+46        *pstatus = RESTORE_LEDS; 
+47    else 
+48        *pstatus = ALL_LEDS_ON; 
+49 
+50    (my_driver->ops->ioctl)(t, KDSETLED, *pstatus); 
+51 
+52    my_timer.expires = jiffies + BLINK_DELAY; 
+53    add_timer(&my_timer); 
+54} 
+55 
+56static int __init kbleds_init(void) 
+57{ 
+58    int i; 
+59 
+60    pr_info("kbleds: loading\n"); 
+61    pr_info("kbleds: fgconsole is %x\n", fg_console); 
+62    for (i = 0; i < MAX_NR_CONSOLES; i++) { 
+63        if (!vc_cons[i].d) 
+64            break; 
+65        pr_info("poet_atkm: console[%i/%i] #%i, tty %lx\n", i, MAX_NR_CONSOLES, 
+66                vc_cons[i].d->vc_num, (unsigned long) vc_cons[i].d->port.tty); 
+67    } 
+68    pr_info("kbleds: finished scanning consoles\n"); 
+69 
+70    my_driver = vc_cons[fg_console].d->port.tty->driver; 
+71    pr_info("kbleds: tty driver magic %x\n", my_driver->magic); 
+72 
+73    /* 
+74     * Set up the LED blink timer the first time 
+75     */ 
+76    timer_setup(&my_timer, (void *) &my_timer_func, 
+77                (unsigned long) &kbledstatus); 
+78    my_timer.expires = jiffies + BLINK_DELAY; 
+79    add_timer(&my_timer); 
+80 
+81    return 0; 
+82} 
+83 
+84static void __exit kbleds_cleanup(void) 
+85{ 
+86    pr_info("kbleds: unloading...\n"); 
+87    del_timer(&my_timer); 
+88    (my_driver->ops->ioctl)(vc_cons[fg_console].d->port.tty, KDSETLED, 
+89                            RESTORE_LEDS); 
+90} 
+91 
+92module_init(kbleds_init); 
+93module_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 @@ -4155,64 +4180,64 @@ over a serial line. If you find yourself porting the kernel to some new and former unsupported architecture this is usually amongst 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 +

    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 sure it does not show up in production code. -

    +

    -

    0.14 Scheduling Tasks

    -

    There are two main ways of running tasks: tasklets and work queues. Tasklets are a +

    0.14 Scheduling Tasks

    +

    There are two main ways of running tasks: tasklets and work queues. Tasklets are a quick and easy way of scheduling a single function to be run, for example when triggered from an interrupt, whereas work queues are more complicated but also better suited to running multiple things in a sequence. -

    +

    -

    0.14.1 Tasklets

    -

    Here’s an example tasklet module. The tasklet_fn function runs for a few seconds +

    0.14.1 Tasklets

    +

    Here’s an example tasklet module. The tasklet_fn function runs for a few seconds and in the mean time execution of the example_tasklet_init function continues to the exit point.

    - 
    1/* 
-2 *  example_tasklet.c 
-3 */ 
-4#include <linux/delay.h> 
-5#include <linux/interrupt.h> 
-6#include <linux/kernel.h> 
-7#include <linux/module.h> 
-8 
-9static void tasklet_fn(unsigned long data) 
-10{ 
-11    pr_info("Example tasklet starts\n"); 
-12    mdelay(5000); 
-13    pr_info("Example tasklet ends\n"); 
-14} 
-15 
-16DECLARE_TASKLET(mytask, tasklet_fn, 0L); 
-17 
-18static int example_tasklet_init(void) 
-19{ 
-20    pr_info("tasklet example init\n"); 
-21    tasklet_schedule(&mytask); 
-22    mdelay(200); 
-23    pr_info("Example tasklet init continues...\n"); 
-24    return 0; 
-25} 
-26 
-27static void example_tasklet_exit(void) 
-28{ 
-29    pr_info("tasklet example exit\n"); 
-30    tasklet_kill(&mytask); 
-31} 
-32 
-33module_init(example_tasklet_init); 
-34module_exit(example_tasklet_exit); 
-35 
-36MODULE_DESCRIPTION("Tasklet example"); 
-37MODULE_LICENSE("GPL");
    -

    So with this example loaded dmesg should show: + 

    1/* 
+2 *  example_tasklet.c 
+3 */ 
+4#include <linux/delay.h> 
+5#include <linux/interrupt.h> 
+6#include <linux/kernel.h> 
+7#include <linux/module.h> 
+8 
+9static void tasklet_fn(unsigned long data) 
+10{ 
+11    pr_info("Example tasklet starts\n"); 
+12    mdelay(5000); 
+13    pr_info("Example tasklet ends\n"); 
+14} 
+15 
+16DECLARE_TASKLET(mytask, tasklet_fn, 0L); 
+17 
+18static int example_tasklet_init(void) 
+19{ 
+20    pr_info("tasklet example init\n"); 
+21    tasklet_schedule(&mytask); 
+22    mdelay(200); 
+23    pr_info("Example tasklet init continues...\n"); 
+24    return 0; 
+25} 
+26 
+27static void example_tasklet_exit(void) 
+28{ 
+29    pr_info("tasklet example exit\n"); 
+30    tasklet_kill(&mytask); 
+31} 
+32 
+33module_init(example_tasklet_init); 
+34module_exit(example_tasklet_exit); 
+35 
+36MODULE_DESCRIPTION("Tasklet example"); 
+37MODULE_LICENSE("GPL");
    +

    So with this example loaded dmesg should show: @@ -4223,57 +4248,57 @@ Example tasklet starts Example tasklet init continues... Example tasklet ends -

    -

    +

    +

    -

    0.14.2 Work queues

    -

    To add a task to the scheduler we can use a workqueue. The kernel then uses the +

    0.14.2 Work queues

    +

    To add a task to the scheduler we can use a workqueue. The kernel then uses the Completely Fair Scheduler (CFS) to execute work within the queue.

    - 
    1/* 
-2 *  sched.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/module.h> 
-6#include <linux/workqueue.h> 
-7 
-8static struct workqueue_struct *queue = NULL; 
-9static struct work_struct work; 
-10 
-11static void work_handler(struct work_struct *data) 
-12{ 
-13    pr_info("work handler function.\n"); 
-14} 
-15 
-16int init_module() 
-17{ 
-18    queue = alloc_workqueue("HELLOWORLD", WQ_UNBOUND, 1); 
-19    INIT_WORK(&work, work_handler); 
-20    schedule_work(&work); 
-21 
-22    return 0; 
-23} 
-24 
-25void cleanup_module() 
-26{ 
-27    destroy_workqueue(queue); 
-28} 
-29 
-30MODULE_LICENSE("GPL"); 
-31MODULE_DESCRIPTION("Workqueue example");
    -

    + 

    1/* 
+2 *  sched.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/module.h> 
+6#include <linux/workqueue.h> 
+7 
+8static struct workqueue_struct *queue = NULL; 
+9static struct work_struct work; 
+10 
+11static void work_handler(struct work_struct *data) 
+12{ 
+13    pr_info("work handler function.\n"); 
+14} 
+15 
+16int init_module() 
+17{ 
+18    queue = alloc_workqueue("HELLOWORLD", WQ_UNBOUND, 1); 
+19    INIT_WORK(&work, work_handler); 
+20    schedule_work(&work); 
+21 
+22    return 0; 
+23} 
+24 
+25void cleanup_module() 
+26{ 
+27    destroy_workqueue(queue); 
+28} 
+29 
+30MODULE_LICENSE("GPL"); 
+31MODULE_DESCRIPTION("Workqueue example");
    +

    -

    0.15 Interrupt Handlers

    -

    +

    0.15 Interrupt Handlers

    +

    -

    0.15.1 Interrupt Handlers

    -

    Except for the last chapter, everything we did in the kernel so far we’ve done as a +

    0.15.1 Interrupt Handlers

    +

    Except for the last chapter, everything we did in the kernel so far we’ve done as a response to a process asking for it, either by dealing with a special file, sending an ioctl(), or issuing a system call. But the job of the kernel isn’t just to respond to process requests. Another job, which is every bit as important, is to speak to the hardware connected to the machine. -

    There are two types of interaction between the CPU and the rest of the +

    There are two types of interaction between the CPU and the rest of the computer’s hardware. The first type is when the CPU gives orders to the hardware, the other is when the hardware needs to tell the CPU something. The second, called interrupts, is much harder to implement because it has to be dealt with when @@ -4283,14 +4308,14 @@ lost. -

    Under Linux, hardware interrupts are called IRQ’s (Interrupt ReQuests). There +

    Under Linux, hardware interrupts are called IRQ’s (Interrupt ReQuests). There are two types of IRQ’s, short and long. A short IRQ is one which is expected to take a very short period of time, during which the rest of the machine will be blocked and no other interrupts will be handled. A long IRQ is one which can take longer, and during which other interrupts may occur (but not interrupts from the same device). If at all possible, it’s better to declare an interrupt handler to be long. -

    When the CPU receives an interrupt, it stops whatever it’s doing (unless it’s +

    When the CPU receives an interrupt, it stops whatever it’s doing (unless it’s processing a more important interrupt, in which case it will deal with this one only when the more important one is done), saves certain parameters on the stack and calls the interrupt handler. This means that certain things @@ -4302,9 +4327,9 @@ the new information at a later time (this is called the "bottom half") and return. The kernel is then guaranteed to call the bottom half as soon as possible – and when it does, everything allowed in kernel modules will be allowed. -

    The way to implement this is to call request_irq() to get your interrupt handler +

    The way to implement this is to call request_irq() to get your interrupt handler called when the relevant IRQ is received. -

    In practice IRQ handling can be a bit more complex. Hardware is often +

    In practice IRQ handling can be a bit more complex. Hardware is often designed in a way that chains two interrupt controllers, so that all the IRQs from interrupt controller B are cascaded to a certain IRQ from interrupt controller A. Of course that requires that the kernel finds out which IRQ it @@ -4318,7 +4343,7 @@ another truckload of problems. It’s not enough to know if a certain IRQs has happend, it’s also important for what CPU(s) it was for. People still interested in more details, might want to do a web search for "APIC" now ;) -

    This function receives the IRQ number, the name of the function, flags, a name +

    This function receives the IRQ number, the name of the function, flags, a name for /proc/interrupts and a parameter to pass to the interrupt handler. Usually there is a certain number of IRQs available. How many IRQs there are is hardware-dependent. The flags can include SA_SHIRQ to indicate you’re willing to @@ -4329,333 +4354,333 @@ or if you’re both willing to share. -

    +

    -

    0.15.2 Detecting button presses

    -

    Many popular single board computers, such as Raspberry Pis or Beagleboards, have +

    0.15.2 Detecting button presses

    +

    Many popular single board computers, such as Raspberry Pis or Beagleboards, have a bunch of GPIO pins. Attaching buttons to those and then having a button press do something is a classic case in which you might need to use interrupts so that instead of having the CPU waste time and battery power polling for a change in input state it’s better for the input to trigger the CPU to then run a particular handling function. -

    Here’s an example where buttons are connected to GPIO numbers 17 and 18 and +

    Here’s an example where buttons are connected to GPIO numbers 17 and 18 and an LED is connected to GPIO 4. You can change those numbers to whatever is appropriate for your board.

    - 
    1/* 
-2 *  intrpt.c - Handling GPIO with interrupts 
-3 * 
-4 *  Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
-5 *  from: 
-6 *    https://github.com/wendlers/rpi-kmod-samples 
-7 * 
-8 *  Press one button to turn on a LED and another to turn it off 
-9 */ 
-10 
-11#include <linux/gpio.h> 
-12#include <linux/interrupt.h> 
-13#include <linux/kernel.h> 
-14#include <linux/module.h> 
-15 
-16static int button_irqs[] = {-1, -1}; 
-17 
-18/* Define GPIOs for LEDs. 
-19   Change the numbers for the GPIO on your board. */ 
-20static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
-21 
-22/* Define GPIOs for BUTTONS 
-23   Change the numbers for the GPIO on your board. */ 
-24static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
-25                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
-26 
-27/* 
-28 * interrupt function triggered when a button is pressed 
-29 */ 
-30static irqreturn_t button_isr(int irq, void *data) 
-31{ 
-32    /* first button */ 
-33    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
-34        gpio_set_value(leds[0].gpio, 1); 
-35    /* second button */ 
-36    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
-37        gpio_set_value(leds[0].gpio, 0); 
-38 
-39    return IRQ_HANDLED; 
-40} 
-41 
-42int init_module() 
-43{ 
-44    int ret = 0; 
-45 
-46    pr_info("%s\n", __func__); 
-47 
-48    /* register LED gpios */ 
-49    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
-50 
-51    if (ret) { 
-52        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
-53        return ret; 
-54    } 
-55 
-56    /* register BUTTON gpios */ 
-57    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
-58 
-59    if (ret) { 
-60        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
-61        goto fail1; 
-62    } 
-63 
-64    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
-65 
-66    ret = gpio_to_irq(buttons[0].gpio); 
-67 
-68    if (ret < 0) { 
-69        pr_err("Unable to request IRQ: %d\n", ret); 
-70        goto fail2; 
-71    } 
-72 
-73    button_irqs[0] = ret; 
-74 
-75    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
-76 
-77    ret = request_irq(button_irqs[0], button_isr, 
-78                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-79                      "gpiomod#button1", NULL); 
-80 
-81    if (ret) { 
-82        pr_err("Unable to request IRQ: %d\n", ret); 
-83        goto fail2; 
-84    } 
-85 
-86 
-87    ret = gpio_to_irq(buttons[1].gpio); 
-88 
-89    if (ret < 0) { 
-90        pr_err("Unable to request IRQ: %d\n", ret); 
-91        goto fail2; 
-92    } 
-93 
-94    button_irqs[1] = ret; 
-95 
-96    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
-97 
-98    ret = request_irq(button_irqs[1], button_isr, 
-99                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-100                      "gpiomod#button2", NULL); 
-101 
-102    if (ret) { 
-103        pr_err("Unable to request IRQ: %d\n", ret); 
-104        goto fail3; 
-105    } 
-106 
-107    return 0; 
-108 
-109/* cleanup what has been setup so far */ 
-110fail3: 
-111    free_irq(button_irqs[0], NULL); 
-112 
-113fail2: 
-114    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
-115 
-116fail1: 
-117    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-118 
-119    return ret; 
-120} 
-121 
-122void cleanup_module() 
-123{ 
-124    int i; 
-125 
-126    pr_info("%s\n", __func__); 
-127 
-128    /* free irqs */ 
-129    free_irq(button_irqs[0], NULL); 
-130    free_irq(button_irqs[1], NULL); 
-131 
-132    /* turn all LEDs off */ 
-133    for (i = 0; i < ARRAY_SIZE(leds); i++) 
-134        gpio_set_value(leds[i].gpio, 0); 
-135 
-136    /* unregister */ 
-137    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-138    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
-139} 
-140 
-141MODULE_LICENSE("GPL"); 
-142MODULE_DESCRIPTION("Handle some GPIO interrupts");
    -

    + 

    1/* 
+2 *  intrpt.c - Handling GPIO with interrupts 
+3 * 
+4 *  Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
+5 *  from: 
+6 *    https://github.com/wendlers/rpi-kmod-samples 
+7 * 
+8 *  Press one button to turn on a LED and another to turn it off 
+9 */ 
+10 
+11#include <linux/gpio.h> 
+12#include <linux/interrupt.h> 
+13#include <linux/kernel.h> 
+14#include <linux/module.h> 
+15 
+16static int button_irqs[] = {-1, -1}; 
+17 
+18/* Define GPIOs for LEDs. 
+19   Change the numbers for the GPIO on your board. */ 
+20static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
+21 
+22/* Define GPIOs for BUTTONS 
+23   Change the numbers for the GPIO on your board. */ 
+24static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
+25                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
+26 
+27/* 
+28 * interrupt function triggered when a button is pressed 
+29 */ 
+30static irqreturn_t button_isr(int irq, void *data) 
+31{ 
+32    /* first button */ 
+33    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
+34        gpio_set_value(leds[0].gpio, 1); 
+35    /* second button */ 
+36    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
+37        gpio_set_value(leds[0].gpio, 0); 
+38 
+39    return IRQ_HANDLED; 
+40} 
+41 
+42int init_module() 
+43{ 
+44    int ret = 0; 
+45 
+46    pr_info("%s\n", __func__); 
+47 
+48    /* register LED gpios */ 
+49    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
+50 
+51    if (ret) { 
+52        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
+53        return ret; 
+54    } 
+55 
+56    /* register BUTTON gpios */ 
+57    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
+58 
+59    if (ret) { 
+60        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
+61        goto fail1; 
+62    } 
+63 
+64    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
+65 
+66    ret = gpio_to_irq(buttons[0].gpio); 
+67 
+68    if (ret < 0) { 
+69        pr_err("Unable to request IRQ: %d\n", ret); 
+70        goto fail2; 
+71    } 
+72 
+73    button_irqs[0] = ret; 
+74 
+75    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
+76 
+77    ret = request_irq(button_irqs[0], button_isr, 
+78                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+79                      "gpiomod#button1", NULL); 
+80 
+81    if (ret) { 
+82        pr_err("Unable to request IRQ: %d\n", ret); 
+83        goto fail2; 
+84    } 
+85 
+86 
+87    ret = gpio_to_irq(buttons[1].gpio); 
+88 
+89    if (ret < 0) { 
+90        pr_err("Unable to request IRQ: %d\n", ret); 
+91        goto fail2; 
+92    } 
+93 
+94    button_irqs[1] = ret; 
+95 
+96    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
+97 
+98    ret = request_irq(button_irqs[1], button_isr, 
+99                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+100                      "gpiomod#button2", NULL); 
+101 
+102    if (ret) { 
+103        pr_err("Unable to request IRQ: %d\n", ret); 
+104        goto fail3; 
+105    } 
+106 
+107    return 0; 
+108 
+109/* cleanup what has been setup so far */ 
+110fail3: 
+111    free_irq(button_irqs[0], NULL); 
+112 
+113fail2: 
+114    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
+115 
+116fail1: 
+117    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+118 
+119    return ret; 
+120} 
+121 
+122void cleanup_module() 
+123{ 
+124    int i; 
+125 
+126    pr_info("%s\n", __func__); 
+127 
+128    /* free irqs */ 
+129    free_irq(button_irqs[0], NULL); 
+130    free_irq(button_irqs[1], NULL); 
+131 
+132    /* turn all LEDs off */ 
+133    for (i = 0; i < ARRAY_SIZE(leds); i++) 
+134        gpio_set_value(leds[i].gpio, 0); 
+135 
+136    /* unregister */ 
+137    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+138    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
+139} 
+140 
+141MODULE_LICENSE("GPL"); 
+142MODULE_DESCRIPTION("Handle some GPIO interrupts");
    +

    -

    0.15.3 Bottom Half

    -

    Suppose you want to do a bunch of stuff inside of an interrupt routine. A common +

    0.15.3 Bottom Half

    +

    Suppose you want to do a bunch of stuff inside of an interrupt routine. A common way to do that without rendering the interrupt unavailable for a significant duration is to combine it with a tasklet. This pushes the bulk of the work off into the scheduler. -

    The example below modifies the previous example to also run an additional task +

    The example below modifies the previous example to also run an additional task when an interrupt is triggered.

    - 
    1/* 
-2 * bottomhalf.c - Top and bottom half interrupt handling 
-3 * 
-4 * Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
-5 * from: 
-6 *    https://github.com/wendlers/rpi-kmod-samples 
-7 * 
-8 *  Press one button to turn on a LED and another to turn it off 
-9 */ 
-10 
-11#include <linux/delay.h> 
-12#include <linux/gpio.h> 
-13#include <linux/interrupt.h> 
-14#include <linux/kernel.h> 
-15#include <linux/module.h> 
-16 
-17static int button_irqs[] = {-1, -1}; 
-18 
-19/* Define GPIOs for LEDs. 
-20   Change the numbers for the GPIO on your board. */ 
-21static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
-22 
-23/* Define GPIOs for BUTTONS 
-24   Change the numbers for the GPIO on your board. */ 
-25static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
-26                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
-27 
-28/* Tasklet containing some non-trivial amount of processing */ 
-29static void bottomhalf_tasklet_fn(unsigned long data) 
-30{ 
-31    pr_info("Bottom half tasklet starts\n"); 
-32    /* do something which takes a while */ 
-33    mdelay(500); 
-34    pr_info("Bottom half tasklet ends\n"); 
-35} 
-36 
-37DECLARE_TASKLET(buttontask, bottomhalf_tasklet_fn, 0L); 
-38 
-39/* 
-40 * interrupt function triggered when a button is pressed 
-41 */ 
-42static irqreturn_t button_isr(int irq, void *data) 
-43{ 
-44    /* Do something quickly right now */ 
-45    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
-46        gpio_set_value(leds[0].gpio, 1); 
-47    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
-48        gpio_set_value(leds[0].gpio, 0); 
-49 
-50    /* Do the rest at leisure via the scheduler */ 
-51    tasklet_schedule(&buttontask); 
-52 
-53    return IRQ_HANDLED; 
-54} 
-55 
-56int init_module() 
-57{ 
-58    int ret = 0; 
-59 
-60    pr_info("%s\n", __func__); 
-61 
-62    /* register LED gpios */ 
-63    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
-64 
-65    if (ret) { 
-66        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
-67        return ret; 
-68    } 
-69 
-70    /* register BUTTON gpios */ 
-71    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
-72 
-73    if (ret) { 
-74        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
-75        goto fail1; 
-76    } 
-77 
-78    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
-79 
-80    ret = gpio_to_irq(buttons[0].gpio); 
-81 
-82    if (ret < 0) { 
-83        pr_err("Unable to request IRQ: %d\n", ret); 
-84        goto fail2; 
-85    } 
-86 
-87    button_irqs[0] = ret; 
-88 
-89    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
-90 
-91    ret = request_irq(button_irqs[0], button_isr, 
-92                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-93                      "gpiomod#button1", NULL); 
-94 
-95    if (ret) { 
-96        pr_err("Unable to request IRQ: %d\n", ret); 
-97        goto fail2; 
-98    } 
-99 
-100 
-101    ret = gpio_to_irq(buttons[1].gpio); 
-102 
-103    if (ret < 0) { 
-104        pr_err("Unable to request IRQ: %d\n", ret); 
-105        goto fail2; 
-106    } 
-107 
-108    button_irqs[1] = ret; 
-109 
-110    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
-111 
-112    ret = request_irq(button_irqs[1], button_isr, 
-113                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-114                      "gpiomod#button2", NULL); 
-115 
-116    if (ret) { 
-117        pr_err("Unable to request IRQ: %d\n", ret); 
-118        goto fail3; 
-119    } 
-120 
-121    return 0; 
-122 
-123/* cleanup what has been setup so far */ 
-124fail3: 
-125    free_irq(button_irqs[0], NULL); 
-126 
-127fail2: 
-128    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
-129 
-130fail1: 
-131    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-132 
-133    return ret; 
-134} 
-135 
-136void cleanup_module() 
-137{ 
-138    int i; 
-139 
-140    pr_info("%s\n", __func__); 
-141 
-142    /* free irqs */ 
-143    free_irq(button_irqs[0], NULL); 
-144    free_irq(button_irqs[1], NULL); 
-145 
-146    /* turn all LEDs off */ 
-147    for (i = 0; i < ARRAY_SIZE(leds); i++) 
-148        gpio_set_value(leds[i].gpio, 0); 
-149 
-150    /* unregister */ 
-151    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-152    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
-153} 
-154 
-155MODULE_LICENSE("GPL"); 
-156MODULE_DESCRIPTION("Interrupt with top and bottom half");
    -

    + 

    1/* 
+2 * bottomhalf.c - Top and bottom half interrupt handling 
+3 * 
+4 * Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
+5 * from: 
+6 *    https://github.com/wendlers/rpi-kmod-samples 
+7 * 
+8 *  Press one button to turn on a LED and another to turn it off 
+9 */ 
+10 
+11#include <linux/delay.h> 
+12#include <linux/gpio.h> 
+13#include <linux/interrupt.h> 
+14#include <linux/kernel.h> 
+15#include <linux/module.h> 
+16 
+17static int button_irqs[] = {-1, -1}; 
+18 
+19/* Define GPIOs for LEDs. 
+20   Change the numbers for the GPIO on your board. */ 
+21static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
+22 
+23/* Define GPIOs for BUTTONS 
+24   Change the numbers for the GPIO on your board. */ 
+25static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
+26                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
+27 
+28/* Tasklet containing some non-trivial amount of processing */ 
+29static void bottomhalf_tasklet_fn(unsigned long data) 
+30{ 
+31    pr_info("Bottom half tasklet starts\n"); 
+32    /* do something which takes a while */ 
+33    mdelay(500); 
+34    pr_info("Bottom half tasklet ends\n"); 
+35} 
+36 
+37DECLARE_TASKLET(buttontask, bottomhalf_tasklet_fn, 0L); 
+38 
+39/* 
+40 * interrupt function triggered when a button is pressed 
+41 */ 
+42static irqreturn_t button_isr(int irq, void *data) 
+43{ 
+44    /* Do something quickly right now */ 
+45    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
+46        gpio_set_value(leds[0].gpio, 1); 
+47    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
+48        gpio_set_value(leds[0].gpio, 0); 
+49 
+50    /* Do the rest at leisure via the scheduler */ 
+51    tasklet_schedule(&buttontask); 
+52 
+53    return IRQ_HANDLED; 
+54} 
+55 
+56int init_module() 
+57{ 
+58    int ret = 0; 
+59 
+60    pr_info("%s\n", __func__); 
+61 
+62    /* register LED gpios */ 
+63    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
+64 
+65    if (ret) { 
+66        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
+67        return ret; 
+68    } 
+69 
+70    /* register BUTTON gpios */ 
+71    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
+72 
+73    if (ret) { 
+74        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
+75        goto fail1; 
+76    } 
+77 
+78    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
+79 
+80    ret = gpio_to_irq(buttons[0].gpio); 
+81 
+82    if (ret < 0) { 
+83        pr_err("Unable to request IRQ: %d\n", ret); 
+84        goto fail2; 
+85    } 
+86 
+87    button_irqs[0] = ret; 
+88 
+89    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
+90 
+91    ret = request_irq(button_irqs[0], button_isr, 
+92                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+93                      "gpiomod#button1", NULL); 
+94 
+95    if (ret) { 
+96        pr_err("Unable to request IRQ: %d\n", ret); 
+97        goto fail2; 
+98    } 
+99 
+100 
+101    ret = gpio_to_irq(buttons[1].gpio); 
+102 
+103    if (ret < 0) { 
+104        pr_err("Unable to request IRQ: %d\n", ret); 
+105        goto fail2; 
+106    } 
+107 
+108    button_irqs[1] = ret; 
+109 
+110    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
+111 
+112    ret = request_irq(button_irqs[1], button_isr, 
+113                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+114                      "gpiomod#button2", NULL); 
+115 
+116    if (ret) { 
+117        pr_err("Unable to request IRQ: %d\n", ret); 
+118        goto fail3; 
+119    } 
+120 
+121    return 0; 
+122 
+123/* cleanup what has been setup so far */ 
+124fail3: 
+125    free_irq(button_irqs[0], NULL); 
+126 
+127fail2: 
+128    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
+129 
+130fail1: 
+131    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+132 
+133    return ret; 
+134} 
+135 
+136void cleanup_module() 
+137{ 
+138    int i; 
+139 
+140    pr_info("%s\n", __func__); 
+141 
+142    /* free irqs */ 
+143    free_irq(button_irqs[0], NULL); 
+144    free_irq(button_irqs[1], NULL); 
+145 
+146    /* turn all LEDs off */ 
+147    for (i = 0; i < ARRAY_SIZE(leds); i++) 
+148        gpio_set_value(leds[i].gpio, 0); 
+149 
+150    /* unregister */ 
+151    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+152    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
+153} 
+154 
+155MODULE_LICENSE("GPL"); 
+156MODULE_DESCRIPTION("Interrupt with top and bottom half");
    +

    -

    0.16 Crypto

    -

    At the dawn of the internet everybody trusted everybody completely…but that didn’t +

    0.16 Crypto

    +

    At the dawn of the internet everybody trusted everybody completely…but that didn’t work out so well. When this guide was originally written it was a more innocent era in which almost nobody actually gave a damn about crypto - least of all kernel developers. That’s certainly no longer the case now. To handle crypto stuff the kernel @@ -4664,295 +4689,295 @@ favourite hash functions. -

    +

    -

    0.16.1 Hash functions

    -

    Calculating and checking the hashes of things is a common operation. Here is a +

    0.16.1 Hash functions

    +

    Calculating and checking the hashes of things is a common operation. Here is a demonstration of how to calculate a sha256 hash within a kernel module.

    - 
    1/* 
-2 *  cryptosha256.c 
-3 */ 
-4#include <crypto/internal/hash.h> 
-5#include <linux/module.h> 
-6 
-7#define SHA256_LENGTH 32 
-8 
-9static void show_hash_result(char *plaintext, char *hash_sha256) 
-10{ 
-11    int i; 
-12    char str[SHA256_LENGTH * 2 + 1]; 
-13 
-14    pr_info("sha256 test for string: \"%s\"\n", plaintext); 
-15    for (i = 0; i < SHA256_LENGTH; i++) 
-16        sprintf(&str[i * 2], "%02x", (unsigned char) hash_sha256[i]); 
-17    str[i * 2] = 0; 
-18    pr_info("%s\n", str); 
-19} 
-20 
-21int cryptosha256_init(void) 
-22{ 
-23    char *plaintext = "This is a test"; 
-24    char hash_sha256[SHA256_LENGTH]; 
-25    struct crypto_shash *sha256; 
-26    struct shash_desc *shash; 
-27 
-28    sha256 = crypto_alloc_shash("sha256", 0, 0); 
-29    if (IS_ERR(sha256)) 
-30        return -1; 
-31 
-32    shash = kmalloc(sizeof(struct shash_desc) + crypto_shash_descsize(sha256), 
-33                    GFP_KERNEL); 
-34    if (!shash) 
-35        return -ENOMEM; 
-36 
-37    shash->tfm = sha256; 
-38 
-39    if (crypto_shash_init(shash)) 
-40        return -1; 
-41 
-42    if (crypto_shash_update(shash, plaintext, strlen(plaintext))) 
-43        return -1; 
-44 
-45    if (crypto_shash_final(shash, hash_sha256)) 
-46        return -1; 
-47 
-48    kfree(shash); 
-49    crypto_free_shash(sha256); 
-50 
-51    show_hash_result(plaintext, hash_sha256); 
-52 
-53    return 0; 
-54} 
-55 
-56void cryptosha256_exit(void) {} 
-57 
-58module_init(cryptosha256_init); 
-59module_exit(cryptosha256_exit); 
-60 
-61MODULE_DESCRIPTION("sha256 hash test"); 
-62MODULE_LICENSE("GPL");
    -

    Make and install the module: + 

    1/* 
+2 *  cryptosha256.c 
+3 */ 
+4#include <crypto/internal/hash.h> 
+5#include <linux/module.h> 
+6 
+7#define SHA256_LENGTH 32 
+8 
+9static void show_hash_result(char *plaintext, char *hash_sha256) 
+10{ 
+11    int i; 
+12    char str[SHA256_LENGTH * 2 + 1]; 
+13 
+14    pr_info("sha256 test for string: \"%s\"\n", plaintext); 
+15    for (i = 0; i < SHA256_LENGTH; i++) 
+16        sprintf(&str[i * 2], "%02x", (unsigned char) hash_sha256[i]); 
+17    str[i * 2] = 0; 
+18    pr_info("%s\n", str); 
+19} 
+20 
+21int cryptosha256_init(void) 
+22{ 
+23    char *plaintext = "This is a test"; 
+24    char hash_sha256[SHA256_LENGTH]; 
+25    struct crypto_shash *sha256; 
+26    struct shash_desc *shash; 
+27 
+28    sha256 = crypto_alloc_shash("sha256", 0, 0); 
+29    if (IS_ERR(sha256)) 
+30        return -1; 
+31 
+32    shash = kmalloc(sizeof(struct shash_desc) + crypto_shash_descsize(sha256), 
+33                    GFP_KERNEL); 
+34    if (!shash) 
+35        return -ENOMEM; 
+36 
+37    shash->tfm = sha256; 
+38 
+39    if (crypto_shash_init(shash)) 
+40        return -1; 
+41 
+42    if (crypto_shash_update(shash, plaintext, strlen(plaintext))) 
+43        return -1; 
+44 
+45    if (crypto_shash_final(shash, hash_sha256)) 
+46        return -1; 
+47 
+48    kfree(shash); 
+49    crypto_free_shash(sha256); 
+50 
+51    show_hash_result(plaintext, hash_sha256); 
+52 
+53    return 0; 
+54} 
+55 
+56void cryptosha256_exit(void) {} 
+57 
+58module_init(cryptosha256_init); 
+59module_exit(cryptosha256_exit); 
+60 
+61MODULE_DESCRIPTION("sha256 hash test"); 
+62MODULE_LICENSE("GPL");
    +

    Make and install the module:

    - 
    1make 
-2sudo insmod cryptosha256.ko 
-3dmesg
    -

    And you should see that the hash was calculated for the test string. -

    Finally, remove the test module: + 

    1make 
+2sudo insmod cryptosha256.ko 
+3dmesg
    +

    And you should see that the hash was calculated for the test string. +

    Finally, remove the test module:

    - 
    1sudo rmmod cryptosha256
    -

    + 

    1sudo rmmod cryptosha256
    +

    -

    0.16.2 Symmetric key encryption

    -

    Here is an example of symmetrically encrypting a string using the AES algorithm +

    0.16.2 Symmetric key encryption

    +

    Here is an example of symmetrically encrypting a string using the AES algorithm and a password.

    - 
    1/* 
-2 *  cryptosk.c 
-3 */ 
-4#include <crypto/internal/skcipher.h> 
-5#include <linux/crypto.h> 
-6#include <linux/module.h> 
-7 
-8#define SYMMETRIC_KEY_LENGTH 32 
-9#define CIPHER_BLOCK_SIZE 16 
-10 
-11struct tcrypt_result { 
-12    struct completion completion; 
-13    int err; 
-14}; 
-15 
-16struct skcipher_def { 
-17    struct scatterlist sg; 
-18    struct crypto_skcipher *tfm; 
-19    struct skcipher_request *req; 
-20    struct tcrypt_result result; 
-21    char *scratchpad; 
-22    char *ciphertext; 
-23    char *ivdata; 
-24}; 
-25 
-26static struct skcipher_def sk; 
-27 
-28static void test_skcipher_finish(struct skcipher_def *sk) 
-29{ 
-30    if (sk->tfm) 
-31        crypto_free_skcipher(sk->tfm); 
-32    if (sk->req) 
-33        skcipher_request_free(sk->req); 
-34    if (sk->ivdata) 
-35        kfree(sk->ivdata); 
-36    if (sk->scratchpad) 
-37        kfree(sk->scratchpad); 
-38    if (sk->ciphertext) 
-39        kfree(sk->ciphertext); 
-40} 
-41 
-42static int test_skcipher_result(struct skcipher_def *sk, int rc) 
-43{ 
-44    switch (rc) { 
-45    case 0: 
-46        break; 
-47    case -EINPROGRESS || -EBUSY: 
-48        rc = wait_for_completion_interruptible(&sk->result.completion); 
-49        if (!rc && !sk->result.err) { 
-50            reinit_completion(&sk->result.completion); 
-51            break; 
-52        } 
-53        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
-54                sk->result.err); 
-55        break; 
-56    default: 
-57        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
-58                sk->result.err); 
-59        break; 
-60    } 
-61 
-62    init_completion(&sk->result.completion); 
-63 
-64    return rc; 
-65} 
-66 
-67static void test_skcipher_callback(struct crypto_async_request *req, int error) 
-68{ 
-69    struct tcrypt_result *result = req->data; 
-70    /* int ret; */ 
-71 
-72    if (error == -EINPROGRESS) 
-73        return; 
-74 
-75    result->err = error; 
-76    complete(&result->completion); 
-77    pr_info("Encryption finished successfully\n"); 
-78 
-79    /* decrypt data */ 
-80    /* 
-81    memset((void*)sk.scratchpad, '-', CIPHER_BLOCK_SIZE); 
-82    ret = crypto_skcipher_decrypt(sk.req); 
-83    ret = test_skcipher_result(&sk, ret); 
-84    if (ret) 
-85        return; 
-86 
-87    sg_copy_from_buffer(&sk.sg, 1, sk.scratchpad, CIPHER_BLOCK_SIZE); 
-88    sk.scratchpad[CIPHER_BLOCK_SIZE-1] = 0; 
-89 
-90    pr_info("Decryption request successful\n"); 
-91    pr_info("Decrypted: %s\n", sk.scratchpad); 
-92    */ 
-93} 
-94 
-95static int test_skcipher_encrypt(char *plaintext, 
-96                                 char *password, 
-97                                 struct skcipher_def *sk) 
-98{ 
-99    int ret = -EFAULT; 
-100    unsigned char key[SYMMETRIC_KEY_LENGTH]; 
-101 
-102    if (!sk->tfm) { 
-103        sk->tfm = crypto_alloc_skcipher("cbc-aes-aesni", 0, 0); 
-104        if (IS_ERR(sk->tfm)) { 
-105            pr_info("could not allocate skcipher handle\n"); 
-106            return PTR_ERR(sk->tfm); 
-107        } 
-108    } 
-109 
-110    if (!sk->req) { 
-111        sk->req = skcipher_request_alloc(sk->tfm, GFP_KERNEL); 
-112        if (!sk->req) { 
-113            pr_info("could not allocate skcipher request\n"); 
-114            ret = -ENOMEM; 
-115            goto out; 
-116        } 
-117    } 
-118 
-119    skcipher_request_set_callback(sk->req, CRYPTO_TFM_REQ_MAY_BACKLOG, 
-120                                  test_skcipher_callback, &sk->result); 
-121 
-122    /* clear the key */ 
-123    memset((void *) key, '\0', SYMMETRIC_KEY_LENGTH); 
-124 
-125    /* Use the world's favourite password */ 
-126    sprintf((char *) key, "%s", password); 
-127 
-128    /* AES 256 with given symmetric key */ 
-129    if (crypto_skcipher_setkey(sk->tfm, key, SYMMETRIC_KEY_LENGTH)) { 
-130        pr_info("key could not be set\n"); 
-131        ret = -EAGAIN; 
-132        goto out; 
-133    } 
-134    pr_info("Symmetric key: %s\n", key); 
-135    pr_info("Plaintext: %s\n", plaintext); 
-136 
-137    if (!sk->ivdata) { 
-138        /* see https://en.wikipedia.org/wiki/Initialization_vector */ 
-139        sk->ivdata = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
-140        if (!sk->ivdata) { 
-141            pr_info("could not allocate ivdata\n"); 
-142            goto out; 
-143        } 
-144        get_random_bytes(sk->ivdata, CIPHER_BLOCK_SIZE); 
-145    } 
-146 
-147    if (!sk->scratchpad) { 
-148        /* The text to be encrypted */ 
-149        sk->scratchpad = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
-150        if (!sk->scratchpad) { 
-151            pr_info("could not allocate scratchpad\n"); 
-152            goto out; 
-153        } 
-154    } 
-155    sprintf((char *) sk->scratchpad, "%s", plaintext); 
-156 
-157    sg_init_one(&sk->sg, sk->scratchpad, CIPHER_BLOCK_SIZE); 
-158    skcipher_request_set_crypt(sk->req, &sk->sg, &sk->sg, CIPHER_BLOCK_SIZE, 
-159                               sk->ivdata); 
-160    init_completion(&sk->result.completion); 
-161 
-162    /* encrypt data */ 
-163    ret = crypto_skcipher_encrypt(sk->req); 
-164    ret = test_skcipher_result(sk, ret); 
-165    if (ret) 
-166        goto out; 
-167 
-168    pr_info("Encryption request successful\n"); 
-169 
-170out: 
-171    return ret; 
-172} 
-173 
-174int cryptoapi_init(void) 
-175{ 
-176    /* The world's favorite password */ 
-177    char *password = "password123"; 
-178 
-179    sk.tfm = NULL; 
-180    sk.req = NULL; 
-181    sk.scratchpad = NULL; 
-182    sk.ciphertext = NULL; 
-183    sk.ivdata = NULL; 
-184 
-185    test_skcipher_encrypt("Testing", password, &sk); 
-186    return 0; 
-187} 
-188 
-189void cryptoapi_exit(void) 
-190{ 
-191    test_skcipher_finish(&sk); 
-192} 
-193 
-194module_init(cryptoapi_init); 
-195module_exit(cryptoapi_exit); 
-196 
-197MODULE_DESCRIPTION("Symmetric key encryption example"); 
-198MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  cryptosk.c 
+3 */ 
+4#include <crypto/internal/skcipher.h> 
+5#include <linux/crypto.h> 
+6#include <linux/module.h> 
+7 
+8#define SYMMETRIC_KEY_LENGTH 32 
+9#define CIPHER_BLOCK_SIZE 16 
+10 
+11struct tcrypt_result { 
+12    struct completion completion; 
+13    int err; 
+14}; 
+15 
+16struct skcipher_def { 
+17    struct scatterlist sg; 
+18    struct crypto_skcipher *tfm; 
+19    struct skcipher_request *req; 
+20    struct tcrypt_result result; 
+21    char *scratchpad; 
+22    char *ciphertext; 
+23    char *ivdata; 
+24}; 
+25 
+26static struct skcipher_def sk; 
+27 
+28static void test_skcipher_finish(struct skcipher_def *sk) 
+29{ 
+30    if (sk->tfm) 
+31        crypto_free_skcipher(sk->tfm); 
+32    if (sk->req) 
+33        skcipher_request_free(sk->req); 
+34    if (sk->ivdata) 
+35        kfree(sk->ivdata); 
+36    if (sk->scratchpad) 
+37        kfree(sk->scratchpad); 
+38    if (sk->ciphertext) 
+39        kfree(sk->ciphertext); 
+40} 
+41 
+42static int test_skcipher_result(struct skcipher_def *sk, int rc) 
+43{ 
+44    switch (rc) { 
+45    case 0: 
+46        break; 
+47    case -EINPROGRESS || -EBUSY: 
+48        rc = wait_for_completion_interruptible(&sk->result.completion); 
+49        if (!rc && !sk->result.err) { 
+50            reinit_completion(&sk->result.completion); 
+51            break; 
+52        } 
+53        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
+54                sk->result.err); 
+55        break; 
+56    default: 
+57        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
+58                sk->result.err); 
+59        break; 
+60    } 
+61 
+62    init_completion(&sk->result.completion); 
+63 
+64    return rc; 
+65} 
+66 
+67static void test_skcipher_callback(struct crypto_async_request *req, int error) 
+68{ 
+69    struct tcrypt_result *result = req->data; 
+70    /* int ret; */ 
+71 
+72    if (error == -EINPROGRESS) 
+73        return; 
+74 
+75    result->err = error; 
+76    complete(&result->completion); 
+77    pr_info("Encryption finished successfully\n"); 
+78 
+79    /* decrypt data */ 
+80    /* 
+81    memset((void*)sk.scratchpad, '-', CIPHER_BLOCK_SIZE); 
+82    ret = crypto_skcipher_decrypt(sk.req); 
+83    ret = test_skcipher_result(&sk, ret); 
+84    if (ret) 
+85        return; 
+86 
+87    sg_copy_from_buffer(&sk.sg, 1, sk.scratchpad, CIPHER_BLOCK_SIZE); 
+88    sk.scratchpad[CIPHER_BLOCK_SIZE-1] = 0; 
+89 
+90    pr_info("Decryption request successful\n"); 
+91    pr_info("Decrypted: %s\n", sk.scratchpad); 
+92    */ 
+93} 
+94 
+95static int test_skcipher_encrypt(char *plaintext, 
+96                                 char *password, 
+97                                 struct skcipher_def *sk) 
+98{ 
+99    int ret = -EFAULT; 
+100    unsigned char key[SYMMETRIC_KEY_LENGTH]; 
+101 
+102    if (!sk->tfm) { 
+103        sk->tfm = crypto_alloc_skcipher("cbc-aes-aesni", 0, 0); 
+104        if (IS_ERR(sk->tfm)) { 
+105            pr_info("could not allocate skcipher handle\n"); 
+106            return PTR_ERR(sk->tfm); 
+107        } 
+108    } 
+109 
+110    if (!sk->req) { 
+111        sk->req = skcipher_request_alloc(sk->tfm, GFP_KERNEL); 
+112        if (!sk->req) { 
+113            pr_info("could not allocate skcipher request\n"); 
+114            ret = -ENOMEM; 
+115            goto out; 
+116        } 
+117    } 
+118 
+119    skcipher_request_set_callback(sk->req, CRYPTO_TFM_REQ_MAY_BACKLOG, 
+120                                  test_skcipher_callback, &sk->result); 
+121 
+122    /* clear the key */ 
+123    memset((void *) key, '\0', SYMMETRIC_KEY_LENGTH); 
+124 
+125    /* Use the world's favourite password */ 
+126    sprintf((char *) key, "%s", password); 
+127 
+128    /* AES 256 with given symmetric key */ 
+129    if (crypto_skcipher_setkey(sk->tfm, key, SYMMETRIC_KEY_LENGTH)) { 
+130        pr_info("key could not be set\n"); 
+131        ret = -EAGAIN; 
+132        goto out; 
+133    } 
+134    pr_info("Symmetric key: %s\n", key); 
+135    pr_info("Plaintext: %s\n", plaintext); 
+136 
+137    if (!sk->ivdata) { 
+138        /* see https://en.wikipedia.org/wiki/Initialization_vector */ 
+139        sk->ivdata = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
+140        if (!sk->ivdata) { 
+141            pr_info("could not allocate ivdata\n"); 
+142            goto out; 
+143        } 
+144        get_random_bytes(sk->ivdata, CIPHER_BLOCK_SIZE); 
+145    } 
+146 
+147    if (!sk->scratchpad) { 
+148        /* The text to be encrypted */ 
+149        sk->scratchpad = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
+150        if (!sk->scratchpad) { 
+151            pr_info("could not allocate scratchpad\n"); 
+152            goto out; 
+153        } 
+154    } 
+155    sprintf((char *) sk->scratchpad, "%s", plaintext); 
+156 
+157    sg_init_one(&sk->sg, sk->scratchpad, CIPHER_BLOCK_SIZE); 
+158    skcipher_request_set_crypt(sk->req, &sk->sg, &sk->sg, CIPHER_BLOCK_SIZE, 
+159                               sk->ivdata); 
+160    init_completion(&sk->result.completion); 
+161 
+162    /* encrypt data */ 
+163    ret = crypto_skcipher_encrypt(sk->req); 
+164    ret = test_skcipher_result(sk, ret); 
+165    if (ret) 
+166        goto out; 
+167 
+168    pr_info("Encryption request successful\n"); 
+169 
+170out: 
+171    return ret; 
+172} 
+173 
+174int cryptoapi_init(void) 
+175{ 
+176    /* The world's favorite password */ 
+177    char *password = "password123"; 
+178 
+179    sk.tfm = NULL; 
+180    sk.req = NULL; 
+181    sk.scratchpad = NULL; 
+182    sk.ciphertext = NULL; 
+183    sk.ivdata = NULL; 
+184 
+185    test_skcipher_encrypt("Testing", password, &sk); 
+186    return 0; 
+187} 
+188 
+189void cryptoapi_exit(void) 
+190{ 
+191    test_skcipher_finish(&sk); 
+192} 
+193 
+194module_init(cryptoapi_init); 
+195module_exit(cryptoapi_exit); 
+196 
+197MODULE_DESCRIPTION("Symmetric key encryption example"); 
+198MODULE_LICENSE("GPL");
    +

    -

    0.17 Standardising the interfaces: The Device Model

    -

    Up to this point we’ve seen all kinds of modules doing all kinds of things, but there +

    0.17 Standardising the interfaces: The Device Model

    +

    Up to this point we’ve seen all kinds of modules doing all kinds of things, but there was no consistency in their interfaces with the rest of the kernel. To impose some consistency such that there is at minimum a standardised way to start, suspend and resume a device a device model was added. An example is show below, and you can @@ -4960,188 +4985,188 @@ use this as a template to add your own suspend, resume or other interface functions.

    - 
    1/* 
-2 *  devicemodel.c 
-3 */ 
-4#include <linux/kernel.h> 
-5#include <linux/module.h> 
-6#include <linux/platform_device.h> 
-7 
-8struct devicemodel_data { 
-9    char *greeting; 
-10    int number; 
-11}; 
-12 
-13static int devicemodel_probe(struct platform_device *dev) 
-14{ 
-15    struct devicemodel_data *pd = 
-16        (struct devicemodel_data *) (dev->dev.platform_data); 
-17 
-18    pr_info("devicemodel probe\n"); 
-19    pr_info("devicemodel greeting: %s; %d\n", pd->greeting, pd->number); 
-20 
-21    /* Your device initialisation code */ 
-22 
-23    return 0; 
-24} 
-25 
-26static int devicemodel_remove(struct platform_device *dev) 
-27{ 
-28    pr_info("devicemodel example removed\n"); 
-29 
-30    /* Your device removal code */ 
-31 
-32    return 0; 
-33} 
-34 
-35static int devicemodel_suspend(struct device *dev) 
-36{ 
-37    pr_info("devicemodel example suspend\n"); 
-38 
-39    /* Your device suspend code */ 
-40 
-41    return 0; 
-42} 
-43 
-44static int devicemodel_resume(struct device *dev) 
-45{ 
-46    pr_info("devicemodel example resume\n"); 
-47 
-48    /* Your device resume code */ 
-49 
-50    return 0; 
-51} 
-52 
-53static const struct dev_pm_ops devicemodel_pm_ops = { 
-54    .suspend = devicemodel_suspend, 
-55    .resume = devicemodel_resume, 
-56    .poweroff = devicemodel_suspend, 
-57    .freeze = devicemodel_suspend, 
-58    .thaw = devicemodel_resume, 
-59    .restore = devicemodel_resume, 
-60}; 
-61 
-62static struct platform_driver devicemodel_driver = { 
-63    .driver = 
-64        { 
-65            .name = "devicemodel_example", 
-66            .owner = THIS_MODULE, 
-67            .pm = &devicemodel_pm_ops, 
-68        }, 
-69    .probe = devicemodel_probe, 
-70    .remove = devicemodel_remove, 
-71}; 
-72 
-73static int devicemodel_init(void) 
-74{ 
-75    int ret; 
-76 
-77    pr_info("devicemodel init\n"); 
-78 
-79    ret = platform_driver_register(&devicemodel_driver); 
-80 
-81    if (ret) { 
-82        pr_err("Unable to register driver\n"); 
-83        return ret; 
-84    } 
-85 
-86    return 0; 
-87} 
-88 
-89static void devicemodel_exit(void) 
-90{ 
-91    pr_info("devicemodel exit\n"); 
-92    platform_driver_unregister(&devicemodel_driver); 
-93} 
-94 
-95MODULE_LICENSE("GPL"); 
-96MODULE_DESCRIPTION("Linux Device Model example"); 
-97 
-98module_init(devicemodel_init); 
-99module_exit(devicemodel_exit);
    + 
    1/* 
+2 *  devicemodel.c 
+3 */ 
+4#include <linux/kernel.h> 
+5#include <linux/module.h> 
+6#include <linux/platform_device.h> 
+7 
+8struct devicemodel_data { 
+9    char *greeting; 
+10    int number; 
+11}; 
+12 
+13static int devicemodel_probe(struct platform_device *dev) 
+14{ 
+15    struct devicemodel_data *pd = 
+16        (struct devicemodel_data *) (dev->dev.platform_data); 
+17 
+18    pr_info("devicemodel probe\n"); 
+19    pr_info("devicemodel greeting: %s; %d\n", pd->greeting, pd->number); 
+20 
+21    /* Your device initialisation code */ 
+22 
+23    return 0; 
+24} 
+25 
+26static int devicemodel_remove(struct platform_device *dev) 
+27{ 
+28    pr_info("devicemodel example removed\n"); 
+29 
+30    /* Your device removal code */ 
+31 
+32    return 0; 
+33} 
+34 
+35static int devicemodel_suspend(struct device *dev) 
+36{ 
+37    pr_info("devicemodel example suspend\n"); 
+38 
+39    /* Your device suspend code */ 
+40 
+41    return 0; 
+42} 
+43 
+44static int devicemodel_resume(struct device *dev) 
+45{ 
+46    pr_info("devicemodel example resume\n"); 
+47 
+48    /* Your device resume code */ 
+49 
+50    return 0; 
+51} 
+52 
+53static const struct dev_pm_ops devicemodel_pm_ops = { 
+54    .suspend = devicemodel_suspend, 
+55    .resume = devicemodel_resume, 
+56    .poweroff = devicemodel_suspend, 
+57    .freeze = devicemodel_suspend, 
+58    .thaw = devicemodel_resume, 
+59    .restore = devicemodel_resume, 
+60}; 
+61 
+62static struct platform_driver devicemodel_driver = { 
+63    .driver = 
+64        { 
+65            .name = "devicemodel_example", 
+66            .owner = THIS_MODULE, 
+67            .pm = &devicemodel_pm_ops, 
+68        }, 
+69    .probe = devicemodel_probe, 
+70    .remove = devicemodel_remove, 
+71}; 
+72 
+73static int devicemodel_init(void) 
+74{ 
+75    int ret; 
+76 
+77    pr_info("devicemodel init\n"); 
+78 
+79    ret = platform_driver_register(&devicemodel_driver); 
+80 
+81    if (ret) { 
+82        pr_err("Unable to register driver\n"); 
+83        return ret; 
+84    } 
+85 
+86    return 0; 
+87} 
+88 
+89static void devicemodel_exit(void) 
+90{ 
+91    pr_info("devicemodel exit\n"); 
+92    platform_driver_unregister(&devicemodel_driver); 
+93} 
+94 
+95MODULE_LICENSE("GPL"); 
+96MODULE_DESCRIPTION("Linux Device Model example"); 
+97 
+98module_init(devicemodel_init); 
+99module_exit(devicemodel_exit);
    -

    +

    -

    0.18 Optimizations

    -

    +

    0.18 Optimizations

    +

    -

    0.18.1 Likely and Unlikely conditions

    -

    Sometimes you might want your code to run as quickly as possible, especially if +

    0.18.1 Likely and Unlikely conditions

    +

    Sometimes you might want your code to run as quickly as possible, especially if it’s handling an interrupt or doing something which might cause noticible latency. If your code contains boolean conditions and if you know that the conditions are almost always likely to evaluate as either true or false, then you can allow the compiler to optimise for this using the likely and unlikely macros. -

    For example, when allocating memory you’re almost always expecting this to +

    For example, when allocating memory you’re almost always expecting this to succeed.

    - 
    1bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx); 
-2if (unlikely(!bvl)) { 
-3  mempool_free(bio, bio_pool); 
-4  bio = NULL; 
-5  goto out; 
-6}
    -

    When the unlikely macro is used the compiler alters its machine instruction + 

    1bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx); 
+2if (unlikely(!bvl)) { 
+3  mempool_free(bio, bio_pool); 
+4  bio = NULL; 
+5  goto out; 
+6}
    +

    When the unlikely macro is used the compiler alters its machine instruction output so that it continues along the false branch and only jumps if the condition is true. That avoids flushing the processor pipeline. The opposite happens if you use the likely macro.

    -

    0.19 Common Pitfalls

    -

    Before I send you on your way to go out into the world and write kernel modules, +

    0.19 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. -

    +

    -

    0.19.1 Using standard libraries

    -

    You can’t do that. In a kernel module you can only use kernel functions, which are +

    0.19.1 Using standard libraries

    +

    You can’t do that. In a kernel module you can only use kernel functions, which are the functions you can see in /proc/kallsyms. -

    +

    -

    0.19.2 Disabling interrupts

    -

    You might need to do this for a short time and that is OK, but if you don’t +

    0.19.2 Disabling interrupts

    +

    You might need to do this for a short time and that is OK, but if you don’t enable them afterwards, your system will be stuck and you’ll have to power it off. -

    +

    -

    0.19.3 Sticking your head inside a large carnivore

    -

    I probably don’t have to warn you about this, but I figured I will anyway, just in +

    0.19.3 Sticking your head inside a large carnivore

    +

    I probably don’t have to warn you about this, but I figured I will anyway, just in case. -

    +

    -

    0.20 Where To Go From Here?

    -

    I could easily have squeezed a few more chapters into this book. I could have added a +

    0.20 Where To Go From Here?

    +

    I could easily have squeezed a few more chapters into this book. I could have added a chapter about creating new file systems, or about adding new protocol stacks (as if there’s a need for that – you’d have to dig underground to find a protocol stack not supported by Linux). I could have added explanations of the kernel mechanisms we haven’t touched upon, such as bootstrapping or the disk interface. -

    However, I chose not to. My purpose in writing this book was to provide initiation +

    However, I chose not to. My purpose in writing this book was to provide initiation into the mysteries of kernel module programming and to teach the common techniques for that purpose. For people seriously interested in kernel programming, I recommend kernelnewbies.org and the Documentation subdirectory within the kernel source code which isn’t always easy to understand but can be a starting point for further investigation. Also, as Linus said, the best way to learn the kernel is to read the source code yourself. -

    If you’re interested in more examples of short kernel modules then searching on +

    If you’re interested in more examples of short kernel modules then searching on sites such as Github and Gitlab is a good way to start, although there is a lot of duplication of older LKMPG examples which may not compile with newer kernel versions. You will also be able to find examples of the use of kernel modules to attack or compromise systems or exfiltrate data and those can be useful for thinking about how to defend systems and learning about existing security mechanisms within the kernel. -

    I hope I have helped you in your quest to become a better programmer, or at +

    I hope I have helped you in your quest to become a better programmer, or at least to have fun through technology. And, if you do write useful kernel modules, I hope you publish them under the GPL, so I can use them too. -

    If you’d like to contribute to this guide or notice anything glaringly wrong, please +

    If you’d like to contribute to this guide or notice anything glaringly wrong, please create an issue at https://github.com/sysprog21/lkmpg. -

    Happy hacking. +

    Happy hacking.

    diff --git a/lkmpg.css b/lkmpg.css index b6f0cfb..1c9159c 100644 --- a/lkmpg.css +++ b/lkmpg.css @@ -623,66 +623,66 @@ span#textcolor272{color:rgb(43,145,175)} span#textcolor273{color:rgb(43,145,175)} span#textcolor274{color:rgb(0,0,255)} span#textcolor275{color:rgb(0,0,255)} -span#textcolor276{color:rgb(0,0,255)} -span#textcolor277{color:rgb(43,145,175)} -span#textcolor278{color:rgb(43,145,175)} +span#textcolor276{color:rgb(43,145,175)} +span#textcolor277{color:rgb(0,0,255)} +span#textcolor278{color:rgb(0,0,255)} span#textcolor279{color:rgb(43,145,175)} span#textcolor280{color:rgb(0,0,255)} -span#textcolor281{color:rgb(0,0,255)} -span#textcolor282{color:rgb(0,0,255)} -span#textcolor283{color:rgb(43,145,175)} -span#textcolor284{color:rgb(43,145,175)} +span#textcolor281{color:rgb(43,145,175)} +span#textcolor282{color:rgb(43,145,175)} +span#textcolor283{color:rgb(0,0,255)} +span#textcolor284{color:rgb(0,0,255)} span#textcolor285{color:rgb(43,145,175)} span#textcolor286{color:rgb(0,0,255)} span#textcolor287{color:rgb(0,0,255)} -span#textcolor288{color:rgb(43,145,175)} +span#textcolor288{color:rgb(0,0,255)} span#textcolor289{color:rgb(0,0,255)} -span#textcolor290{color:rgb(0,0,255)} -span#textcolor291{color:rgb(43,145,175)} -span#textcolor292{color:rgb(0,0,255)} +span#textcolor290{color:rgb(43,145,175)} +span#textcolor291{color:rgb(0,0,255)} +span#textcolor292{color:rgb(43,145,175)} span#textcolor293{color:rgb(43,145,175)} span#textcolor294{color:rgb(43,145,175)} span#textcolor295{color:rgb(43,145,175)} span#textcolor296{color:rgb(43,145,175)} -span#textcolor297{color:rgb(43,145,175)} -span#textcolor298{color:rgb(0,0,255)} +span#textcolor297{color:rgb(0,0,255)} +span#textcolor298{color:rgb(43,145,175)} span#textcolor299{color:rgb(43,145,175)} span#textcolor300{color:rgb(43,145,175)} span#textcolor301{color:rgb(43,145,175)} span#textcolor302{color:rgb(43,145,175)} -span#textcolor303{color:rgb(43,145,175)} +span#textcolor303{color:rgb(0,0,255)} span#textcolor304{color:rgb(0,0,255)} -span#textcolor305{color:rgb(0,0,255)} +span#textcolor305{color:rgb(43,145,175)} span#textcolor306{color:rgb(43,145,175)} -span#textcolor307{color:rgb(0,0,255)} +span#textcolor307{color:rgb(43,145,175)} span#textcolor308{color:rgb(0,0,255)} -span#textcolor309{color:rgb(43,145,175)} -span#textcolor310{color:rgb(0,0,255)} -span#textcolor311{color:rgb(43,145,175)} -span#textcolor312{color:rgb(0,0,255)} +span#textcolor309{color:rgb(0,0,255)} +span#textcolor310{color:rgb(43,145,175)} +span#textcolor311{color:rgb(0,0,255)} +span#textcolor312{color:rgb(43,145,175)} span#textcolor313{color:rgb(0,0,255)} -span#textcolor314{color:rgb(43,145,175)} -span#textcolor315{color:rgb(0,0,255)} -span#textcolor316{color:rgb(43,145,175)} +span#textcolor314{color:rgb(0,0,255)} +span#textcolor315{color:rgb(43,145,175)} +span#textcolor316{color:rgb(0,0,255)} span#textcolor317{color:rgb(43,145,175)} -span#textcolor318{color:rgb(0,0,255)} +span#textcolor318{color:rgb(43,145,175)} span#textcolor319{color:rgb(43,145,175)} -span#textcolor320{color:rgb(43,145,175)} +span#textcolor320{color:rgb(0,0,255)} span#textcolor321{color:rgb(43,145,175)} -span#textcolor322{color:rgb(0,0,255)} -span#textcolor323{color:rgb(43,145,175)} +span#textcolor322{color:rgb(43,145,175)} +span#textcolor323{color:rgb(0,0,255)} span#textcolor324{color:rgb(43,145,175)} span#textcolor325{color:rgb(0,0,255)} span#textcolor326{color:rgb(43,145,175)} span#textcolor327{color:rgb(0,0,255)} -span#textcolor328{color:rgb(43,145,175)} -span#textcolor329{color:rgb(0,0,255)} -span#textcolor330{color:rgb(0,0,255)} +span#textcolor328{color:rgb(0,0,255)} +span#textcolor329{color:rgb(43,145,175)} +span#textcolor330{color:rgb(43,145,175)} span#textcolor331{color:rgb(43,145,175)} span#textcolor332{color:rgb(43,145,175)} -span#textcolor333{color:rgb(43,145,175)} +span#textcolor333{color:rgb(0,0,255)} span#textcolor334{color:rgb(43,145,175)} -span#textcolor335{color:rgb(0,0,255)} +span#textcolor335{color:rgb(43,145,175)} span#textcolor336{color:rgb(43,145,175)} span#textcolor337{color:rgb(43,145,175)} span#textcolor338{color:rgb(43,145,175)} @@ -692,56 +692,68 @@ span#textcolor341{color:rgb(43,145,175)} span#textcolor342{color:rgb(43,145,175)} span#textcolor343{color:rgb(43,145,175)} span#textcolor344{color:rgb(43,145,175)} -span#textcolor345{color:rgb(43,145,175)} +span#textcolor345{color:rgb(0,0,255)} span#textcolor346{color:rgb(43,145,175)} span#textcolor347{color:rgb(0,0,255)} span#textcolor348{color:rgb(43,145,175)} span#textcolor349{color:rgb(0,0,255)} -span#textcolor350{color:rgb(43,145,175)} -span#textcolor351{color:rgb(0,0,255)} -span#textcolor352{color:rgb(0,0,255)} +span#textcolor350{color:rgb(0,0,255)} +span#textcolor351{color:rgb(43,145,175)} +span#textcolor352{color:rgb(43,145,175)} span#textcolor353{color:rgb(43,145,175)} span#textcolor354{color:rgb(43,145,175)} -span#textcolor355{color:rgb(43,145,175)} -span#textcolor356{color:rgb(43,145,175)} -span#textcolor357{color:rgb(0,0,255)} -span#textcolor358{color:rgb(0,0,255)} +span#textcolor355{color:rgb(0,0,255)} +span#textcolor356{color:rgb(0,0,255)} +span#textcolor357{color:rgb(43,145,175)} +span#textcolor358{color:rgb(43,145,175)} span#textcolor359{color:rgb(43,145,175)} span#textcolor360{color:rgb(43,145,175)} -span#textcolor361{color:rgb(43,145,175)} +span#textcolor361{color:rgb(0,0,255)} span#textcolor362{color:rgb(43,145,175)} span#textcolor363{color:rgb(0,0,255)} span#textcolor364{color:rgb(43,145,175)} -span#textcolor365{color:rgb(0,0,255)} -span#textcolor366{color:rgb(43,145,175)} -span#textcolor367{color:rgb(0,0,255)} +span#textcolor365{color:rgb(43,145,175)} +span#textcolor366{color:rgb(0,0,255)} +span#textcolor367{color:rgb(43,145,175)} span#textcolor368{color:rgb(43,145,175)} -span#textcolor369{color:rgb(43,145,175)} +span#textcolor369{color:rgb(0,0,255)} span#textcolor370{color:rgb(0,0,255)} -span#textcolor371{color:rgb(0,0,255)} +span#textcolor371{color:rgb(43,145,175)} +span#textcolor372{color:rgb(0,0,255)} +span#textcolor373{color:rgb(0,0,255)} +span#textcolor374{color:rgb(43,145,175)} +span#textcolor375{color:rgb(43,145,175)} +span#textcolor376{color:rgb(43,145,175)} +span#textcolor377{color:rgb(0,0,255)} +span#textcolor378{color:rgb(0,0,255)} +span#textcolor379{color:rgb(43,145,175)} +span#textcolor380{color:rgb(43,145,175)} +span#textcolor381{color:rgb(43,145,175)} +span#textcolor382{color:rgb(0,0,255)} +span#textcolor383{color:rgb(43,145,175)} pre#fancyvrb32{padding:5.69054pt;} pre#fancyvrb32{ border-top: solid 0.4pt; } pre#fancyvrb32{ border-left: solid 0.4pt; } pre#fancyvrb32{ border-bottom: solid 0.4pt; } pre#fancyvrb32{ border-right: solid 0.4pt; } -span#textcolor372{color:rgb(0,0,255)} +span#textcolor384{color:rgb(0,0,255)} pre#fancyvrb33{padding:5.69054pt;} pre#fancyvrb33{ border-top: solid 0.4pt; } pre#fancyvrb33{ border-left: solid 0.4pt; } pre#fancyvrb33{ border-bottom: solid 0.4pt; } pre#fancyvrb33{ border-right: solid 0.4pt; } -span#textcolor373{color:rgb(0,0,255)} +span#textcolor385{color:rgb(0,0,255)} pre#fancyvrb34{padding:5.69054pt;} pre#fancyvrb34{ border-top: solid 0.4pt; } pre#fancyvrb34{ border-left: solid 0.4pt; } pre#fancyvrb34{ border-bottom: solid 0.4pt; } pre#fancyvrb34{ border-right: solid 0.4pt; } -span#textcolor374{color:rgb(43,145,175)} -span#textcolor375{color:rgb(43,145,175)} -span#textcolor376{color:rgb(43,145,175)} -span#textcolor377{color:rgb(0,0,255)} -span#textcolor378{color:rgb(43,145,175)} -span#textcolor379{color:rgb(0,0,255)} +span#textcolor386{color:rgb(43,145,175)} +span#textcolor387{color:rgb(43,145,175)} +span#textcolor388{color:rgb(43,145,175)} +span#textcolor389{color:rgb(0,0,255)} +span#textcolor390{color:rgb(43,145,175)} +span#textcolor391{color:rgb(0,0,255)} pre#fancyvrb35{padding:5.69054pt;} pre#fancyvrb35{ border-top: solid 0.4pt; } pre#fancyvrb35{ border-left: solid 0.4pt; } @@ -752,21 +764,9 @@ pre#fancyvrb36{ border-top: solid 0.4pt; } pre#fancyvrb36{ border-left: solid 0.4pt; } pre#fancyvrb36{ border-bottom: solid 0.4pt; } pre#fancyvrb36{ border-right: solid 0.4pt; } -span#textcolor380{color:rgb(0,127,0)} -span#textcolor381{color:rgb(0,127,0)} -span#textcolor382{color:rgb(0,127,0)} -span#textcolor383{color:rgb(0,127,0)} -span#textcolor384{color:rgb(0,0,255)} -span#textcolor385{color:rgb(0,127,0)} -span#textcolor386{color:rgb(0,0,255)} -span#textcolor387{color:rgb(0,127,0)} -span#textcolor388{color:rgb(0,0,255)} -span#textcolor389{color:rgb(0,127,0)} -span#textcolor390{color:rgb(0,0,255)} -span#textcolor391{color:rgb(0,127,0)} -span#textcolor392{color:rgb(0,0,255)} +span#textcolor392{color:rgb(0,127,0)} span#textcolor393{color:rgb(0,127,0)} -span#textcolor394{color:rgb(0,0,255)} +span#textcolor394{color:rgb(0,127,0)} span#textcolor395{color:rgb(0,127,0)} span#textcolor396{color:rgb(0,0,255)} span#textcolor397{color:rgb(0,127,0)} @@ -774,567 +774,579 @@ span#textcolor398{color:rgb(0,0,255)} span#textcolor399{color:rgb(0,127,0)} span#textcolor400{color:rgb(0,0,255)} span#textcolor401{color:rgb(0,127,0)} -span#textcolor402{color:rgb(0,127,0)} +span#textcolor402{color:rgb(0,0,255)} span#textcolor403{color:rgb(0,127,0)} -span#textcolor404{color:rgb(0,127,0)} -span#textcolor405{color:rgb(43,145,175)} -span#textcolor406{color:rgb(43,145,175)} -span#textcolor407{color:rgb(43,145,175)} -span#textcolor408{color:rgb(43,145,175)} -span#textcolor409{color:rgb(0,0,255)} -span#textcolor410{color:rgb(43,145,175)} -span#textcolor411{color:rgb(0,0,255)} +span#textcolor404{color:rgb(0,0,255)} +span#textcolor405{color:rgb(0,127,0)} +span#textcolor406{color:rgb(0,0,255)} +span#textcolor407{color:rgb(0,127,0)} +span#textcolor408{color:rgb(0,0,255)} +span#textcolor409{color:rgb(0,127,0)} +span#textcolor410{color:rgb(0,0,255)} +span#textcolor411{color:rgb(0,127,0)} span#textcolor412{color:rgb(0,0,255)} -span#textcolor413{color:rgb(0,0,255)} -span#textcolor414{color:rgb(43,145,175)} -span#textcolor415{color:rgb(0,0,255)} -span#textcolor416{color:rgb(0,0,255)} -span#textcolor417{color:rgb(0,0,255)} +span#textcolor413{color:rgb(0,127,0)} +span#textcolor414{color:rgb(0,127,0)} +span#textcolor415{color:rgb(0,127,0)} +span#textcolor416{color:rgb(0,127,0)} +span#textcolor417{color:rgb(43,145,175)} span#textcolor418{color:rgb(43,145,175)} -span#textcolor419{color:rgb(0,0,255)} +span#textcolor419{color:rgb(43,145,175)} span#textcolor420{color:rgb(43,145,175)} -span#textcolor421{color:rgb(43,145,175)} -span#textcolor422{color:rgb(0,0,255)} -span#textcolor423{color:rgb(43,145,175)} +span#textcolor421{color:rgb(0,0,255)} +span#textcolor422{color:rgb(43,145,175)} +span#textcolor423{color:rgb(0,0,255)} span#textcolor424{color:rgb(0,0,255)} span#textcolor425{color:rgb(0,0,255)} span#textcolor426{color:rgb(43,145,175)} -span#textcolor427{color:rgb(43,145,175)} +span#textcolor427{color:rgb(0,0,255)} span#textcolor428{color:rgb(0,0,255)} span#textcolor429{color:rgb(0,0,255)} -span#textcolor430{color:rgb(0,127,0)} +span#textcolor430{color:rgb(43,145,175)} span#textcolor431{color:rgb(0,0,255)} -span#textcolor432{color:rgb(0,127,0)} -span#textcolor433{color:rgb(0,127,0)} -span#textcolor434{color:rgb(0,127,0)} -span#textcolor435{color:rgb(0,127,0)} +span#textcolor432{color:rgb(43,145,175)} +span#textcolor433{color:rgb(43,145,175)} +span#textcolor434{color:rgb(0,0,255)} +span#textcolor435{color:rgb(43,145,175)} span#textcolor436{color:rgb(0,0,255)} -span#textcolor437{color:rgb(43,145,175)} -span#textcolor438{color:rgb(0,127,0)} -span#textcolor439{color:rgb(0,0,255)} -span#textcolor440{color:rgb(43,145,175)} -span#textcolor441{color:rgb(0,127,0)} +span#textcolor437{color:rgb(0,0,255)} +span#textcolor438{color:rgb(43,145,175)} +span#textcolor439{color:rgb(43,145,175)} +span#textcolor440{color:rgb(0,0,255)} +span#textcolor441{color:rgb(0,0,255)} span#textcolor442{color:rgb(0,127,0)} span#textcolor443{color:rgb(0,0,255)} -span#textcolor444{color:rgb(43,145,175)} +span#textcolor444{color:rgb(0,127,0)} span#textcolor445{color:rgb(0,127,0)} -span#textcolor446{color:rgb(0,0,255)} -span#textcolor447{color:rgb(43,145,175)} +span#textcolor446{color:rgb(0,127,0)} +span#textcolor447{color:rgb(0,127,0)} span#textcolor448{color:rgb(0,0,255)} -span#textcolor449{color:rgb(0,0,255)} -span#textcolor450{color:rgb(0,0,255)} +span#textcolor449{color:rgb(43,145,175)} +span#textcolor450{color:rgb(0,127,0)} span#textcolor451{color:rgb(0,0,255)} -span#textcolor452{color:rgb(0,127,0)} +span#textcolor452{color:rgb(43,145,175)} span#textcolor453{color:rgb(0,127,0)} span#textcolor454{color:rgb(0,127,0)} -span#textcolor455{color:rgb(43,145,175)} +span#textcolor455{color:rgb(0,0,255)} span#textcolor456{color:rgb(43,145,175)} -span#textcolor457{color:rgb(0,0,255)} -span#textcolor458{color:rgb(163,20,20)} -span#textcolor459{color:rgb(163,20,20)} -span#textcolor460{color:rgb(163,20,20)} +span#textcolor457{color:rgb(0,127,0)} +span#textcolor458{color:rgb(0,0,255)} +span#textcolor459{color:rgb(43,145,175)} +span#textcolor460{color:rgb(0,0,255)} span#textcolor461{color:rgb(0,0,255)} -span#textcolor462{color:rgb(163,20,20)} -span#textcolor463{color:rgb(163,20,20)} -span#textcolor464{color:rgb(163,20,20)} -span#textcolor465{color:rgb(163,20,20)} -span#textcolor466{color:rgb(163,20,20)} -span#textcolor467{color:rgb(163,20,20)} -span#textcolor468{color:rgb(0,0,255)} -span#textcolor469{color:rgb(0,127,0)} -span#textcolor470{color:rgb(0,127,0)} -span#textcolor471{color:rgb(0,127,0)} -span#textcolor472{color:rgb(43,145,175)} -span#textcolor473{color:rgb(43,145,175)} -span#textcolor474{color:rgb(0,127,0)} -span#textcolor475{color:rgb(0,127,0)} -span#textcolor476{color:rgb(0,127,0)} -span#textcolor477{color:rgb(0,127,0)} -span#textcolor478{color:rgb(0,127,0)} -span#textcolor479{color:rgb(0,127,0)} -span#textcolor480{color:rgb(0,127,0)} +span#textcolor462{color:rgb(0,0,255)} +span#textcolor463{color:rgb(0,0,255)} +span#textcolor464{color:rgb(0,127,0)} +span#textcolor465{color:rgb(0,127,0)} +span#textcolor466{color:rgb(0,127,0)} +span#textcolor467{color:rgb(43,145,175)} +span#textcolor468{color:rgb(43,145,175)} +span#textcolor469{color:rgb(0,0,255)} +span#textcolor470{color:rgb(163,20,20)} +span#textcolor471{color:rgb(163,20,20)} +span#textcolor472{color:rgb(163,20,20)} +span#textcolor473{color:rgb(0,0,255)} +span#textcolor474{color:rgb(163,20,20)} +span#textcolor475{color:rgb(163,20,20)} +span#textcolor476{color:rgb(163,20,20)} +span#textcolor477{color:rgb(163,20,20)} +span#textcolor478{color:rgb(163,20,20)} +span#textcolor479{color:rgb(163,20,20)} +span#textcolor480{color:rgb(0,0,255)} span#textcolor481{color:rgb(0,127,0)} span#textcolor482{color:rgb(0,127,0)} span#textcolor483{color:rgb(0,127,0)} -span#textcolor484{color:rgb(0,0,255)} +span#textcolor484{color:rgb(43,145,175)} span#textcolor485{color:rgb(43,145,175)} -span#textcolor486{color:rgb(0,0,255)} -span#textcolor487{color:rgb(0,0,255)} -span#textcolor488{color:rgb(0,0,255)} -span#textcolor489{color:rgb(43,145,175)} -span#textcolor490{color:rgb(0,0,255)} -span#textcolor491{color:rgb(0,0,255)} -span#textcolor492{color:rgb(163,20,20)} -span#textcolor493{color:rgb(163,20,20)} -span#textcolor494{color:rgb(163,20,20)} -span#textcolor495{color:rgb(0,0,255)} -span#textcolor496{color:rgb(0,127,0)} -span#textcolor497{color:rgb(0,127,0)} -span#textcolor498{color:rgb(0,127,0)} +span#textcolor486{color:rgb(0,127,0)} +span#textcolor487{color:rgb(0,127,0)} +span#textcolor488{color:rgb(0,127,0)} +span#textcolor489{color:rgb(0,127,0)} +span#textcolor490{color:rgb(0,127,0)} +span#textcolor491{color:rgb(0,127,0)} +span#textcolor492{color:rgb(0,127,0)} +span#textcolor493{color:rgb(0,127,0)} +span#textcolor494{color:rgb(0,127,0)} +span#textcolor495{color:rgb(0,127,0)} +span#textcolor496{color:rgb(0,0,255)} +span#textcolor497{color:rgb(43,145,175)} +span#textcolor498{color:rgb(0,0,255)} span#textcolor499{color:rgb(0,0,255)} -span#textcolor500{color:rgb(43,145,175)} -span#textcolor501{color:rgb(0,0,255)} +span#textcolor500{color:rgb(0,0,255)} +span#textcolor501{color:rgb(43,145,175)} span#textcolor502{color:rgb(0,0,255)} -span#textcolor503{color:rgb(0,127,0)} -span#textcolor504{color:rgb(0,127,0)} -span#textcolor505{color:rgb(0,127,0)} -span#textcolor506{color:rgb(0,127,0)} -span#textcolor507{color:rgb(0,127,0)} -span#textcolor508{color:rgb(0,0,255)} +span#textcolor503{color:rgb(0,0,255)} +span#textcolor504{color:rgb(163,20,20)} +span#textcolor505{color:rgb(163,20,20)} +span#textcolor506{color:rgb(163,20,20)} +span#textcolor507{color:rgb(0,0,255)} +span#textcolor508{color:rgb(0,127,0)} span#textcolor509{color:rgb(0,127,0)} span#textcolor510{color:rgb(0,127,0)} -span#textcolor511{color:rgb(0,127,0)} -span#textcolor512{color:rgb(0,127,0)} +span#textcolor511{color:rgb(0,0,255)} +span#textcolor512{color:rgb(43,145,175)} span#textcolor513{color:rgb(0,0,255)} -span#textcolor514{color:rgb(43,145,175)} -span#textcolor515{color:rgb(0,0,255)} +span#textcolor514{color:rgb(0,0,255)} +span#textcolor515{color:rgb(0,127,0)} span#textcolor516{color:rgb(0,127,0)} -span#textcolor517{color:rgb(43,145,175)} +span#textcolor517{color:rgb(0,127,0)} span#textcolor518{color:rgb(0,127,0)} -span#textcolor519{color:rgb(43,145,175)} -span#textcolor520{color:rgb(0,127,0)} +span#textcolor519{color:rgb(0,127,0)} +span#textcolor520{color:rgb(0,0,255)} span#textcolor521{color:rgb(0,127,0)} span#textcolor522{color:rgb(0,127,0)} span#textcolor523{color:rgb(0,127,0)} -span#textcolor524{color:rgb(43,145,175)} -span#textcolor525{color:rgb(0,127,0)} -span#textcolor526{color:rgb(0,127,0)} -span#textcolor527{color:rgb(0,127,0)} +span#textcolor524{color:rgb(0,127,0)} +span#textcolor525{color:rgb(0,0,255)} +span#textcolor526{color:rgb(43,145,175)} +span#textcolor527{color:rgb(0,0,255)} span#textcolor528{color:rgb(0,127,0)} -span#textcolor529{color:rgb(0,0,255)} -span#textcolor530{color:rgb(0,0,255)} -span#textcolor531{color:rgb(0,127,0)} +span#textcolor529{color:rgb(43,145,175)} +span#textcolor530{color:rgb(0,127,0)} +span#textcolor531{color:rgb(43,145,175)} span#textcolor532{color:rgb(0,127,0)} span#textcolor533{color:rgb(0,127,0)} -span#textcolor534{color:rgb(0,0,255)} +span#textcolor534{color:rgb(0,127,0)} span#textcolor535{color:rgb(0,127,0)} -span#textcolor536{color:rgb(0,127,0)} +span#textcolor536{color:rgb(43,145,175)} span#textcolor537{color:rgb(0,127,0)} span#textcolor538{color:rgb(0,127,0)} span#textcolor539{color:rgb(0,127,0)} span#textcolor540{color:rgb(0,127,0)} -span#textcolor541{color:rgb(0,127,0)} -span#textcolor542{color:rgb(0,127,0)} +span#textcolor541{color:rgb(0,0,255)} +span#textcolor542{color:rgb(0,0,255)} span#textcolor543{color:rgb(0,127,0)} -span#textcolor544{color:rgb(0,0,255)} +span#textcolor544{color:rgb(0,127,0)} span#textcolor545{color:rgb(0,127,0)} -span#textcolor546{color:rgb(0,127,0)} +span#textcolor546{color:rgb(0,0,255)} span#textcolor547{color:rgb(0,127,0)} -span#textcolor548{color:rgb(0,0,255)} -span#textcolor549{color:rgb(43,145,175)} -span#textcolor550{color:rgb(0,0,255)} -span#textcolor551{color:rgb(0,0,255)} -span#textcolor552{color:rgb(43,145,175)} -span#textcolor553{color:rgb(43,145,175)} -span#textcolor554{color:rgb(163,20,20)} -span#textcolor555{color:rgb(163,20,20)} -span#textcolor556{color:rgb(163,20,20)} -span#textcolor557{color:rgb(0,0,255)} -span#textcolor558{color:rgb(163,20,20)} +span#textcolor548{color:rgb(0,127,0)} +span#textcolor549{color:rgb(0,127,0)} +span#textcolor550{color:rgb(0,127,0)} +span#textcolor551{color:rgb(0,127,0)} +span#textcolor552{color:rgb(0,127,0)} +span#textcolor553{color:rgb(0,127,0)} +span#textcolor554{color:rgb(0,127,0)} +span#textcolor555{color:rgb(0,127,0)} +span#textcolor556{color:rgb(0,0,255)} +span#textcolor557{color:rgb(0,127,0)} +span#textcolor558{color:rgb(0,127,0)} +span#textcolor559{color:rgb(0,127,0)} +span#textcolor560{color:rgb(0,0,255)} +span#textcolor561{color:rgb(43,145,175)} +span#textcolor562{color:rgb(0,0,255)} +span#textcolor563{color:rgb(0,0,255)} +span#textcolor564{color:rgb(43,145,175)} +span#textcolor565{color:rgb(43,145,175)} +span#textcolor566{color:rgb(163,20,20)} +span#textcolor567{color:rgb(163,20,20)} +span#textcolor568{color:rgb(163,20,20)} +span#textcolor569{color:rgb(0,0,255)} +span#textcolor570{color:rgb(163,20,20)} pre#fancyvrb37{padding:5.69054pt;} pre#fancyvrb37{ border-top: solid 0.4pt; } pre#fancyvrb37{ border-left: solid 0.4pt; } pre#fancyvrb37{ border-bottom: solid 0.4pt; } pre#fancyvrb37{ border-right: solid 0.4pt; } -span#textcolor559{color:rgb(0,127,0)} -span#textcolor560{color:rgb(0,127,0)} -span#textcolor561{color:rgb(0,127,0)} -span#textcolor562{color:rgb(0,0,255)} -span#textcolor563{color:rgb(0,127,0)} -span#textcolor564{color:rgb(0,0,255)} -span#textcolor565{color:rgb(0,127,0)} -span#textcolor566{color:rgb(0,0,255)} -span#textcolor567{color:rgb(0,127,0)} -span#textcolor568{color:rgb(0,0,255)} -span#textcolor569{color:rgb(0,127,0)} -span#textcolor570{color:rgb(0,0,255)} span#textcolor571{color:rgb(0,127,0)} -span#textcolor572{color:rgb(0,0,255)} -span#textcolor573{color:rgb(0,0,255)} +span#textcolor572{color:rgb(0,127,0)} +span#textcolor573{color:rgb(0,127,0)} span#textcolor574{color:rgb(0,0,255)} -span#textcolor575{color:rgb(0,0,255)} +span#textcolor575{color:rgb(0,127,0)} span#textcolor576{color:rgb(0,0,255)} -span#textcolor577{color:rgb(43,145,175)} +span#textcolor577{color:rgb(0,127,0)} span#textcolor578{color:rgb(0,0,255)} -span#textcolor579{color:rgb(43,145,175)} -span#textcolor580{color:rgb(43,145,175)} -span#textcolor581{color:rgb(43,145,175)} -span#textcolor582{color:rgb(163,20,20)} -span#textcolor583{color:rgb(163,20,20)} -span#textcolor584{color:rgb(163,20,20)} -span#textcolor585{color:rgb(43,145,175)} +span#textcolor579{color:rgb(0,127,0)} +span#textcolor580{color:rgb(0,0,255)} +span#textcolor581{color:rgb(0,127,0)} +span#textcolor582{color:rgb(0,0,255)} +span#textcolor583{color:rgb(0,127,0)} +span#textcolor584{color:rgb(0,0,255)} +span#textcolor585{color:rgb(0,0,255)} span#textcolor586{color:rgb(0,0,255)} -span#textcolor587{color:rgb(43,145,175)} +span#textcolor587{color:rgb(0,0,255)} span#textcolor588{color:rgb(0,0,255)} -span#textcolor589{color:rgb(163,20,20)} -span#textcolor590{color:rgb(163,20,20)} -span#textcolor591{color:rgb(163,20,20)} -span#textcolor592{color:rgb(0,0,255)} -span#textcolor593{color:rgb(163,20,20)} +span#textcolor589{color:rgb(43,145,175)} +span#textcolor590{color:rgb(0,0,255)} +span#textcolor591{color:rgb(43,145,175)} +span#textcolor592{color:rgb(43,145,175)} +span#textcolor593{color:rgb(43,145,175)} span#textcolor594{color:rgb(163,20,20)} span#textcolor595{color:rgb(163,20,20)} -span#textcolor596{color:rgb(0,0,255)} -span#textcolor597{color:rgb(0,0,255)} +span#textcolor596{color:rgb(163,20,20)} +span#textcolor597{color:rgb(43,145,175)} span#textcolor598{color:rgb(0,0,255)} -span#textcolor599{color:rgb(0,0,255)} +span#textcolor599{color:rgb(43,145,175)} span#textcolor600{color:rgb(0,0,255)} -span#textcolor601{color:rgb(0,0,255)} -span#textcolor602{color:rgb(0,0,255)} -span#textcolor603{color:rgb(0,0,255)} +span#textcolor601{color:rgb(163,20,20)} +span#textcolor602{color:rgb(163,20,20)} +span#textcolor603{color:rgb(163,20,20)} span#textcolor604{color:rgb(0,0,255)} -span#textcolor605{color:rgb(0,0,255)} -span#textcolor606{color:rgb(43,145,175)} -span#textcolor607{color:rgb(0,0,255)} -span#textcolor608{color:rgb(163,20,20)} -span#textcolor609{color:rgb(163,20,20)} -span#textcolor610{color:rgb(163,20,20)} +span#textcolor605{color:rgb(163,20,20)} +span#textcolor606{color:rgb(163,20,20)} +span#textcolor607{color:rgb(163,20,20)} +span#textcolor608{color:rgb(0,0,255)} +span#textcolor609{color:rgb(0,0,255)} +span#textcolor610{color:rgb(0,0,255)} span#textcolor611{color:rgb(0,0,255)} -span#textcolor612{color:rgb(163,20,20)} -span#textcolor613{color:rgb(163,20,20)} -span#textcolor614{color:rgb(163,20,20)} +span#textcolor612{color:rgb(0,0,255)} +span#textcolor613{color:rgb(0,0,255)} +span#textcolor614{color:rgb(0,0,255)} span#textcolor615{color:rgb(0,0,255)} -span#textcolor616{color:rgb(43,145,175)} -span#textcolor617{color:rgb(163,20,20)} -span#textcolor618{color:rgb(163,20,20)} -span#textcolor619{color:rgb(163,20,20)} +span#textcolor616{color:rgb(0,0,255)} +span#textcolor617{color:rgb(0,0,255)} +span#textcolor618{color:rgb(43,145,175)} +span#textcolor619{color:rgb(0,0,255)} span#textcolor620{color:rgb(163,20,20)} +span#textcolor621{color:rgb(163,20,20)} +span#textcolor622{color:rgb(163,20,20)} +span#textcolor623{color:rgb(0,0,255)} +span#textcolor624{color:rgb(163,20,20)} +span#textcolor625{color:rgb(163,20,20)} +span#textcolor626{color:rgb(163,20,20)} +span#textcolor627{color:rgb(0,0,255)} +span#textcolor628{color:rgb(43,145,175)} +span#textcolor629{color:rgb(163,20,20)} +span#textcolor630{color:rgb(163,20,20)} +span#textcolor631{color:rgb(163,20,20)} +span#textcolor632{color:rgb(163,20,20)} pre#fancyvrb38{padding:5.69054pt;} pre#fancyvrb38{ border-top: solid 0.4pt; } pre#fancyvrb38{ border-left: solid 0.4pt; } pre#fancyvrb38{ border-bottom: solid 0.4pt; } pre#fancyvrb38{ border-right: solid 0.4pt; } -span#textcolor621{color:rgb(0,127,0)} -span#textcolor622{color:rgb(0,127,0)} -span#textcolor623{color:rgb(0,127,0)} -span#textcolor624{color:rgb(0,0,255)} -span#textcolor625{color:rgb(0,127,0)} -span#textcolor626{color:rgb(0,0,255)} -span#textcolor627{color:rgb(0,127,0)} -span#textcolor628{color:rgb(0,0,255)} -span#textcolor629{color:rgb(0,127,0)} -span#textcolor630{color:rgb(0,0,255)} -span#textcolor631{color:rgb(0,127,0)} -span#textcolor632{color:rgb(0,0,255)} span#textcolor633{color:rgb(0,127,0)} -span#textcolor634{color:rgb(0,0,255)} -span#textcolor635{color:rgb(0,0,255)} +span#textcolor634{color:rgb(0,127,0)} +span#textcolor635{color:rgb(0,127,0)} span#textcolor636{color:rgb(0,0,255)} -span#textcolor637{color:rgb(0,0,255)} +span#textcolor637{color:rgb(0,127,0)} span#textcolor638{color:rgb(0,0,255)} span#textcolor639{color:rgb(0,127,0)} -span#textcolor640{color:rgb(0,127,0)} +span#textcolor640{color:rgb(0,0,255)} span#textcolor641{color:rgb(0,127,0)} -span#textcolor642{color:rgb(0,127,0)} -span#textcolor643{color:rgb(0,0,255)} +span#textcolor642{color:rgb(0,0,255)} +span#textcolor643{color:rgb(0,127,0)} span#textcolor644{color:rgb(0,0,255)} span#textcolor645{color:rgb(0,127,0)} -span#textcolor646{color:rgb(0,127,0)} -span#textcolor647{color:rgb(0,127,0)} -span#textcolor648{color:rgb(0,127,0)} +span#textcolor646{color:rgb(0,0,255)} +span#textcolor647{color:rgb(0,0,255)} +span#textcolor648{color:rgb(0,0,255)} span#textcolor649{color:rgb(0,0,255)} -span#textcolor650{color:rgb(43,145,175)} +span#textcolor650{color:rgb(0,0,255)} span#textcolor651{color:rgb(0,127,0)} span#textcolor652{color:rgb(0,127,0)} span#textcolor653{color:rgb(0,127,0)} span#textcolor654{color:rgb(0,127,0)} span#textcolor655{color:rgb(0,0,255)} -span#textcolor656{color:rgb(43,145,175)} -span#textcolor657{color:rgb(43,145,175)} +span#textcolor656{color:rgb(0,0,255)} +span#textcolor657{color:rgb(0,127,0)} span#textcolor658{color:rgb(0,127,0)} span#textcolor659{color:rgb(0,127,0)} span#textcolor660{color:rgb(0,127,0)} -span#textcolor661{color:rgb(0,127,0)} +span#textcolor661{color:rgb(0,0,255)} span#textcolor662{color:rgb(43,145,175)} -span#textcolor663{color:rgb(0,0,255)} -span#textcolor664{color:rgb(43,145,175)} -span#textcolor665{color:rgb(43,145,175)} -span#textcolor666{color:rgb(43,145,175)} -span#textcolor667{color:rgb(163,20,20)} -span#textcolor668{color:rgb(163,20,20)} -span#textcolor669{color:rgb(163,20,20)} -span#textcolor670{color:rgb(43,145,175)} -span#textcolor671{color:rgb(0,0,255)} -span#textcolor672{color:rgb(43,145,175)} -span#textcolor673{color:rgb(0,0,255)} -span#textcolor674{color:rgb(163,20,20)} -span#textcolor675{color:rgb(163,20,20)} -span#textcolor676{color:rgb(163,20,20)} -span#textcolor677{color:rgb(0,0,255)} -span#textcolor678{color:rgb(163,20,20)} +span#textcolor663{color:rgb(0,127,0)} +span#textcolor664{color:rgb(0,127,0)} +span#textcolor665{color:rgb(0,127,0)} +span#textcolor666{color:rgb(0,127,0)} +span#textcolor667{color:rgb(0,0,255)} +span#textcolor668{color:rgb(43,145,175)} +span#textcolor669{color:rgb(43,145,175)} +span#textcolor670{color:rgb(0,127,0)} +span#textcolor671{color:rgb(0,127,0)} +span#textcolor672{color:rgb(0,127,0)} +span#textcolor673{color:rgb(0,127,0)} +span#textcolor674{color:rgb(43,145,175)} +span#textcolor675{color:rgb(0,0,255)} +span#textcolor676{color:rgb(43,145,175)} +span#textcolor677{color:rgb(43,145,175)} +span#textcolor678{color:rgb(43,145,175)} span#textcolor679{color:rgb(163,20,20)} span#textcolor680{color:rgb(163,20,20)} -span#textcolor681{color:rgb(0,0,255)} -span#textcolor682{color:rgb(0,127,0)} -span#textcolor683{color:rgb(0,127,0)} -span#textcolor684{color:rgb(0,127,0)} -span#textcolor685{color:rgb(0,127,0)} -span#textcolor686{color:rgb(0,0,255)} -span#textcolor687{color:rgb(43,145,175)} -span#textcolor688{color:rgb(0,0,255)} +span#textcolor681{color:rgb(163,20,20)} +span#textcolor682{color:rgb(43,145,175)} +span#textcolor683{color:rgb(0,0,255)} +span#textcolor684{color:rgb(43,145,175)} +span#textcolor685{color:rgb(0,0,255)} +span#textcolor686{color:rgb(163,20,20)} +span#textcolor687{color:rgb(163,20,20)} +span#textcolor688{color:rgb(163,20,20)} span#textcolor689{color:rgb(0,0,255)} -span#textcolor690{color:rgb(43,145,175)} -span#textcolor691{color:rgb(43,145,175)} -span#textcolor692{color:rgb(0,0,255)} +span#textcolor690{color:rgb(163,20,20)} +span#textcolor691{color:rgb(163,20,20)} +span#textcolor692{color:rgb(163,20,20)} span#textcolor693{color:rgb(0,0,255)} -span#textcolor694{color:rgb(0,0,255)} -span#textcolor695{color:rgb(163,20,20)} -span#textcolor696{color:rgb(0,0,255)} -span#textcolor697{color:rgb(0,0,255)} +span#textcolor694{color:rgb(0,127,0)} +span#textcolor695{color:rgb(0,127,0)} +span#textcolor696{color:rgb(0,127,0)} +span#textcolor697{color:rgb(0,127,0)} span#textcolor698{color:rgb(0,0,255)} -span#textcolor699{color:rgb(0,0,255)} +span#textcolor699{color:rgb(43,145,175)} span#textcolor700{color:rgb(0,0,255)} span#textcolor701{color:rgb(0,0,255)} -span#textcolor702{color:rgb(0,0,255)} -span#textcolor703{color:rgb(0,0,255)} +span#textcolor702{color:rgb(43,145,175)} +span#textcolor703{color:rgb(43,145,175)} span#textcolor704{color:rgb(0,0,255)} span#textcolor705{color:rgb(0,0,255)} -span#textcolor706{color:rgb(0,127,0)} -span#textcolor707{color:rgb(0,127,0)} -span#textcolor708{color:rgb(0,127,0)} -span#textcolor709{color:rgb(0,127,0)} -span#textcolor710{color:rgb(43,145,175)} +span#textcolor706{color:rgb(0,0,255)} +span#textcolor707{color:rgb(163,20,20)} +span#textcolor708{color:rgb(0,0,255)} +span#textcolor709{color:rgb(0,0,255)} +span#textcolor710{color:rgb(0,0,255)} span#textcolor711{color:rgb(0,0,255)} -span#textcolor712{color:rgb(163,20,20)} -span#textcolor713{color:rgb(163,20,20)} -span#textcolor714{color:rgb(163,20,20)} +span#textcolor712{color:rgb(0,0,255)} +span#textcolor713{color:rgb(0,0,255)} +span#textcolor714{color:rgb(0,0,255)} span#textcolor715{color:rgb(0,0,255)} -span#textcolor716{color:rgb(163,20,20)} -span#textcolor717{color:rgb(163,20,20)} -span#textcolor718{color:rgb(163,20,20)} -span#textcolor719{color:rgb(0,0,255)} +span#textcolor716{color:rgb(0,0,255)} +span#textcolor717{color:rgb(0,0,255)} +span#textcolor718{color:rgb(0,127,0)} +span#textcolor719{color:rgb(0,127,0)} span#textcolor720{color:rgb(0,127,0)} span#textcolor721{color:rgb(0,127,0)} -span#textcolor722{color:rgb(0,127,0)} -span#textcolor723{color:rgb(0,127,0)} -span#textcolor724{color:rgb(43,145,175)} +span#textcolor722{color:rgb(43,145,175)} +span#textcolor723{color:rgb(0,0,255)} +span#textcolor724{color:rgb(163,20,20)} span#textcolor725{color:rgb(163,20,20)} span#textcolor726{color:rgb(163,20,20)} -span#textcolor727{color:rgb(163,20,20)} +span#textcolor727{color:rgb(0,0,255)} span#textcolor728{color:rgb(163,20,20)} +span#textcolor729{color:rgb(163,20,20)} +span#textcolor730{color:rgb(163,20,20)} +span#textcolor731{color:rgb(0,0,255)} +span#textcolor732{color:rgb(0,127,0)} +span#textcolor733{color:rgb(0,127,0)} +span#textcolor734{color:rgb(0,127,0)} +span#textcolor735{color:rgb(0,127,0)} +span#textcolor736{color:rgb(43,145,175)} +span#textcolor737{color:rgb(163,20,20)} +span#textcolor738{color:rgb(163,20,20)} +span#textcolor739{color:rgb(163,20,20)} +span#textcolor740{color:rgb(163,20,20)} pre#fancyvrb39{padding:5.69054pt;} pre#fancyvrb39{ border-top: solid 0.4pt; } pre#fancyvrb39{ border-left: solid 0.4pt; } pre#fancyvrb39{ border-bottom: solid 0.4pt; } pre#fancyvrb39{ border-right: solid 0.4pt; } -span#textcolor729{color:rgb(0,127,0)} -span#textcolor730{color:rgb(0,127,0)} -span#textcolor731{color:rgb(0,127,0)} -span#textcolor732{color:rgb(0,0,255)} -span#textcolor733{color:rgb(0,127,0)} -span#textcolor734{color:rgb(0,0,255)} -span#textcolor735{color:rgb(0,127,0)} -span#textcolor736{color:rgb(0,0,255)} -span#textcolor737{color:rgb(0,127,0)} -span#textcolor738{color:rgb(0,0,255)} -span#textcolor739{color:rgb(0,127,0)} -span#textcolor740{color:rgb(0,0,255)} span#textcolor741{color:rgb(0,127,0)} -span#textcolor742{color:rgb(0,0,255)} +span#textcolor742{color:rgb(0,127,0)} span#textcolor743{color:rgb(0,127,0)} span#textcolor744{color:rgb(0,0,255)} -span#textcolor745{color:rgb(0,0,255)} +span#textcolor745{color:rgb(0,127,0)} span#textcolor746{color:rgb(0,0,255)} -span#textcolor747{color:rgb(0,0,255)} +span#textcolor747{color:rgb(0,127,0)} span#textcolor748{color:rgb(0,0,255)} -span#textcolor749{color:rgb(0,0,255)} +span#textcolor749{color:rgb(0,127,0)} span#textcolor750{color:rgb(0,0,255)} -span#textcolor751{color:rgb(43,145,175)} +span#textcolor751{color:rgb(0,127,0)} span#textcolor752{color:rgb(0,0,255)} -span#textcolor753{color:rgb(43,145,175)} -span#textcolor754{color:rgb(43,145,175)} -span#textcolor755{color:rgb(0,0,255)} -span#textcolor756{color:rgb(43,145,175)} +span#textcolor753{color:rgb(0,127,0)} +span#textcolor754{color:rgb(0,0,255)} +span#textcolor755{color:rgb(0,127,0)} +span#textcolor756{color:rgb(0,0,255)} span#textcolor757{color:rgb(0,0,255)} -span#textcolor758{color:rgb(43,145,175)} -span#textcolor759{color:rgb(43,145,175)} +span#textcolor758{color:rgb(0,0,255)} +span#textcolor759{color:rgb(0,0,255)} span#textcolor760{color:rgb(0,0,255)} -span#textcolor761{color:rgb(43,145,175)} +span#textcolor761{color:rgb(0,0,255)} span#textcolor762{color:rgb(0,0,255)} -span#textcolor763{color:rgb(163,20,20)} -span#textcolor764{color:rgb(163,20,20)} -span#textcolor765{color:rgb(163,20,20)} -span#textcolor766{color:rgb(0,0,255)} +span#textcolor763{color:rgb(43,145,175)} +span#textcolor764{color:rgb(0,0,255)} +span#textcolor765{color:rgb(43,145,175)} +span#textcolor766{color:rgb(43,145,175)} span#textcolor767{color:rgb(0,0,255)} -span#textcolor768{color:rgb(0,0,255)} -span#textcolor769{color:rgb(163,20,20)} -span#textcolor770{color:rgb(163,20,20)} -span#textcolor771{color:rgb(163,20,20)} +span#textcolor768{color:rgb(43,145,175)} +span#textcolor769{color:rgb(0,0,255)} +span#textcolor770{color:rgb(43,145,175)} +span#textcolor771{color:rgb(43,145,175)} span#textcolor772{color:rgb(0,0,255)} -span#textcolor773{color:rgb(0,0,255)} -span#textcolor774{color:rgb(43,145,175)} -span#textcolor775{color:rgb(0,0,255)} -span#textcolor776{color:rgb(0,0,255)} -span#textcolor777{color:rgb(43,145,175)} -span#textcolor778{color:rgb(43,145,175)} +span#textcolor773{color:rgb(43,145,175)} +span#textcolor774{color:rgb(0,0,255)} +span#textcolor775{color:rgb(163,20,20)} +span#textcolor776{color:rgb(163,20,20)} +span#textcolor777{color:rgb(163,20,20)} +span#textcolor778{color:rgb(0,0,255)} span#textcolor779{color:rgb(0,0,255)} span#textcolor780{color:rgb(0,0,255)} -span#textcolor781{color:rgb(0,0,255)} -span#textcolor782{color:rgb(0,0,255)} +span#textcolor781{color:rgb(163,20,20)} +span#textcolor782{color:rgb(163,20,20)} span#textcolor783{color:rgb(163,20,20)} -span#textcolor784{color:rgb(163,20,20)} -span#textcolor785{color:rgb(163,20,20)} -span#textcolor786{color:rgb(0,0,255)} -span#textcolor787{color:rgb(43,145,175)} +span#textcolor784{color:rgb(0,0,255)} +span#textcolor785{color:rgb(0,0,255)} +span#textcolor786{color:rgb(43,145,175)} +span#textcolor787{color:rgb(0,0,255)} span#textcolor788{color:rgb(0,0,255)} -span#textcolor789{color:rgb(0,0,255)} -span#textcolor790{color:rgb(0,0,255)} -span#textcolor791{color:rgb(43,145,175)} +span#textcolor789{color:rgb(43,145,175)} +span#textcolor790{color:rgb(43,145,175)} +span#textcolor791{color:rgb(0,0,255)} span#textcolor792{color:rgb(0,0,255)} span#textcolor793{color:rgb(0,0,255)} span#textcolor794{color:rgb(0,0,255)} -span#textcolor795{color:rgb(0,0,255)} -span#textcolor796{color:rgb(0,0,255)} -span#textcolor797{color:rgb(0,0,255)} +span#textcolor795{color:rgb(163,20,20)} +span#textcolor796{color:rgb(163,20,20)} +span#textcolor797{color:rgb(163,20,20)} span#textcolor798{color:rgb(0,0,255)} -span#textcolor799{color:rgb(0,0,255)} +span#textcolor799{color:rgb(43,145,175)} span#textcolor800{color:rgb(0,0,255)} span#textcolor801{color:rgb(0,0,255)} span#textcolor802{color:rgb(0,0,255)} span#textcolor803{color:rgb(43,145,175)} span#textcolor804{color:rgb(0,0,255)} -span#textcolor805{color:rgb(163,20,20)} -span#textcolor806{color:rgb(163,20,20)} -span#textcolor807{color:rgb(163,20,20)} +span#textcolor805{color:rgb(0,0,255)} +span#textcolor806{color:rgb(0,0,255)} +span#textcolor807{color:rgb(0,0,255)} span#textcolor808{color:rgb(0,0,255)} -span#textcolor809{color:rgb(163,20,20)} -span#textcolor810{color:rgb(163,20,20)} -span#textcolor811{color:rgb(163,20,20)} +span#textcolor809{color:rgb(0,0,255)} +span#textcolor810{color:rgb(0,0,255)} +span#textcolor811{color:rgb(0,0,255)} span#textcolor812{color:rgb(0,0,255)} -span#textcolor813{color:rgb(43,145,175)} -span#textcolor814{color:rgb(163,20,20)} -span#textcolor815{color:rgb(163,20,20)} -span#textcolor816{color:rgb(163,20,20)} +span#textcolor813{color:rgb(0,0,255)} +span#textcolor814{color:rgb(0,0,255)} +span#textcolor815{color:rgb(43,145,175)} +span#textcolor816{color:rgb(0,0,255)} span#textcolor817{color:rgb(163,20,20)} +span#textcolor818{color:rgb(163,20,20)} +span#textcolor819{color:rgb(163,20,20)} +span#textcolor820{color:rgb(0,0,255)} +span#textcolor821{color:rgb(163,20,20)} +span#textcolor822{color:rgb(163,20,20)} +span#textcolor823{color:rgb(163,20,20)} +span#textcolor824{color:rgb(0,0,255)} +span#textcolor825{color:rgb(43,145,175)} +span#textcolor826{color:rgb(163,20,20)} +span#textcolor827{color:rgb(163,20,20)} +span#textcolor828{color:rgb(163,20,20)} +span#textcolor829{color:rgb(163,20,20)} pre#fancyvrb40{padding:5.69054pt;} pre#fancyvrb40{ border-top: solid 0.4pt; } pre#fancyvrb40{ border-left: solid 0.4pt; } pre#fancyvrb40{ border-bottom: solid 0.4pt; } pre#fancyvrb40{ border-right: solid 0.4pt; } -span#textcolor818{color:rgb(0,127,0)} -span#textcolor819{color:rgb(0,127,0)} -span#textcolor820{color:rgb(0,127,0)} -span#textcolor821{color:rgb(0,127,0)} -span#textcolor822{color:rgb(0,127,0)} -span#textcolor823{color:rgb(0,0,255)} -span#textcolor824{color:rgb(0,127,0)} -span#textcolor825{color:rgb(0,0,255)} -span#textcolor826{color:rgb(0,127,0)} -span#textcolor827{color:rgb(0,0,255)} -span#textcolor828{color:rgb(0,127,0)} -span#textcolor829{color:rgb(0,0,255)} span#textcolor830{color:rgb(0,127,0)} -span#textcolor831{color:rgb(0,0,255)} +span#textcolor831{color:rgb(0,127,0)} span#textcolor832{color:rgb(0,127,0)} -span#textcolor833{color:rgb(0,0,255)} -span#textcolor834{color:rgb(0,0,255)} +span#textcolor833{color:rgb(0,127,0)} +span#textcolor834{color:rgb(0,127,0)} span#textcolor835{color:rgb(0,0,255)} -span#textcolor836{color:rgb(0,0,255)} -span#textcolor837{color:rgb(163,20,20)} +span#textcolor836{color:rgb(0,127,0)} +span#textcolor837{color:rgb(0,0,255)} span#textcolor838{color:rgb(0,127,0)} -span#textcolor839{color:rgb(0,127,0)} +span#textcolor839{color:rgb(0,0,255)} span#textcolor840{color:rgb(0,127,0)} -span#textcolor841{color:rgb(0,127,0)} +span#textcolor841{color:rgb(0,0,255)} span#textcolor842{color:rgb(0,127,0)} -span#textcolor843{color:rgb(0,127,0)} +span#textcolor843{color:rgb(0,0,255)} span#textcolor844{color:rgb(0,127,0)} span#textcolor845{color:rgb(0,0,255)} -span#textcolor846{color:rgb(43,145,175)} +span#textcolor846{color:rgb(0,0,255)} span#textcolor847{color:rgb(0,0,255)} span#textcolor848{color:rgb(0,0,255)} -span#textcolor849{color:rgb(43,145,175)} -span#textcolor850{color:rgb(43,145,175)} +span#textcolor849{color:rgb(163,20,20)} +span#textcolor850{color:rgb(0,127,0)} span#textcolor851{color:rgb(0,127,0)} -span#textcolor852{color:rgb(0,0,255)} +span#textcolor852{color:rgb(0,127,0)} span#textcolor853{color:rgb(0,127,0)} -span#textcolor854{color:rgb(0,0,255)} -span#textcolor855{color:rgb(0,0,255)} +span#textcolor854{color:rgb(0,127,0)} +span#textcolor855{color:rgb(0,127,0)} span#textcolor856{color:rgb(0,127,0)} span#textcolor857{color:rgb(0,0,255)} -span#textcolor858{color:rgb(0,127,0)} -span#textcolor859{color:rgb(0,127,0)} -span#textcolor860{color:rgb(0,127,0)} -span#textcolor861{color:rgb(0,127,0)} -span#textcolor862{color:rgb(0,127,0)} -span#textcolor863{color:rgb(0,0,255)} -span#textcolor864{color:rgb(43,145,175)} -span#textcolor865{color:rgb(0,0,255)} -span#textcolor866{color:rgb(43,145,175)} -span#textcolor867{color:rgb(43,145,175)} -span#textcolor868{color:rgb(43,145,175)} -span#textcolor869{color:rgb(43,145,175)} -span#textcolor870{color:rgb(43,145,175)} -span#textcolor871{color:rgb(0,0,255)} +span#textcolor858{color:rgb(43,145,175)} +span#textcolor859{color:rgb(0,0,255)} +span#textcolor860{color:rgb(0,0,255)} +span#textcolor861{color:rgb(43,145,175)} +span#textcolor862{color:rgb(43,145,175)} +span#textcolor863{color:rgb(0,127,0)} +span#textcolor864{color:rgb(0,0,255)} +span#textcolor865{color:rgb(0,127,0)} +span#textcolor866{color:rgb(0,0,255)} +span#textcolor867{color:rgb(0,0,255)} +span#textcolor868{color:rgb(0,127,0)} +span#textcolor869{color:rgb(0,0,255)} +span#textcolor870{color:rgb(0,127,0)} +span#textcolor871{color:rgb(0,127,0)} span#textcolor872{color:rgb(0,127,0)} span#textcolor873{color:rgb(0,127,0)} span#textcolor874{color:rgb(0,127,0)} -span#textcolor875{color:rgb(0,127,0)} -span#textcolor876{color:rgb(0,0,255)} -span#textcolor877{color:rgb(43,145,175)} -span#textcolor878{color:rgb(0,0,255)} +span#textcolor875{color:rgb(0,0,255)} +span#textcolor876{color:rgb(43,145,175)} +span#textcolor877{color:rgb(0,0,255)} +span#textcolor878{color:rgb(43,145,175)} span#textcolor879{color:rgb(43,145,175)} -span#textcolor880{color:rgb(0,127,0)} -span#textcolor881{color:rgb(0,127,0)} -span#textcolor882{color:rgb(0,127,0)} -span#textcolor883{color:rgb(0,127,0)} +span#textcolor880{color:rgb(43,145,175)} +span#textcolor881{color:rgb(43,145,175)} +span#textcolor882{color:rgb(43,145,175)} +span#textcolor883{color:rgb(0,0,255)} span#textcolor884{color:rgb(0,127,0)} -span#textcolor885{color:rgb(0,0,255)} -span#textcolor886{color:rgb(43,145,175)} -span#textcolor887{color:rgb(0,0,255)} -span#textcolor888{color:rgb(43,145,175)} -span#textcolor889{color:rgb(163,20,20)} -span#textcolor890{color:rgb(163,20,20)} -span#textcolor891{color:rgb(163,20,20)} -span#textcolor892{color:rgb(0,0,255)} +span#textcolor885{color:rgb(0,127,0)} +span#textcolor886{color:rgb(0,127,0)} +span#textcolor887{color:rgb(0,127,0)} +span#textcolor888{color:rgb(0,0,255)} +span#textcolor889{color:rgb(43,145,175)} +span#textcolor890{color:rgb(0,0,255)} +span#textcolor891{color:rgb(43,145,175)} +span#textcolor892{color:rgb(0,127,0)} span#textcolor893{color:rgb(0,127,0)} span#textcolor894{color:rgb(0,127,0)} span#textcolor895{color:rgb(0,127,0)} span#textcolor896{color:rgb(0,127,0)} span#textcolor897{color:rgb(0,0,255)} -span#textcolor898{color:rgb(0,0,255)} -span#textcolor899{color:rgb(0,127,0)} -span#textcolor900{color:rgb(0,127,0)} -span#textcolor901{color:rgb(0,127,0)} -span#textcolor902{color:rgb(0,127,0)} -span#textcolor903{color:rgb(0,0,255)} -span#textcolor904{color:rgb(43,145,175)} -span#textcolor905{color:rgb(0,0,255)} -span#textcolor906{color:rgb(0,0,255)} -span#textcolor907{color:rgb(0,0,255)} +span#textcolor898{color:rgb(43,145,175)} +span#textcolor899{color:rgb(0,0,255)} +span#textcolor900{color:rgb(43,145,175)} +span#textcolor901{color:rgb(163,20,20)} +span#textcolor902{color:rgb(163,20,20)} +span#textcolor903{color:rgb(163,20,20)} +span#textcolor904{color:rgb(0,0,255)} +span#textcolor905{color:rgb(0,127,0)} +span#textcolor906{color:rgb(0,127,0)} +span#textcolor907{color:rgb(0,127,0)} span#textcolor908{color:rgb(0,127,0)} -span#textcolor909{color:rgb(0,127,0)} -span#textcolor910{color:rgb(0,127,0)} +span#textcolor909{color:rgb(0,0,255)} +span#textcolor910{color:rgb(0,0,255)} span#textcolor911{color:rgb(0,127,0)} -span#textcolor912{color:rgb(0,0,255)} -span#textcolor913{color:rgb(0,0,255)} -span#textcolor914{color:rgb(0,0,255)} +span#textcolor912{color:rgb(0,127,0)} +span#textcolor913{color:rgb(0,127,0)} +span#textcolor914{color:rgb(0,127,0)} span#textcolor915{color:rgb(0,0,255)} -span#textcolor916{color:rgb(0,0,255)} +span#textcolor916{color:rgb(43,145,175)} span#textcolor917{color:rgb(0,0,255)} span#textcolor918{color:rgb(0,0,255)} span#textcolor919{color:rgb(0,0,255)} -span#textcolor920{color:rgb(0,0,255)} +span#textcolor920{color:rgb(0,127,0)} span#textcolor921{color:rgb(0,127,0)} span#textcolor922{color:rgb(0,127,0)} span#textcolor923{color:rgb(0,127,0)} -span#textcolor924{color:rgb(0,127,0)} -span#textcolor925{color:rgb(43,145,175)} -span#textcolor926{color:rgb(43,145,175)} +span#textcolor924{color:rgb(0,0,255)} +span#textcolor925{color:rgb(0,0,255)} +span#textcolor926{color:rgb(0,0,255)} span#textcolor927{color:rgb(0,0,255)} span#textcolor928{color:rgb(0,0,255)} -span#textcolor929{color:rgb(163,20,20)} -span#textcolor930{color:rgb(163,20,20)} -span#textcolor931{color:rgb(163,20,20)} +span#textcolor929{color:rgb(0,0,255)} +span#textcolor930{color:rgb(0,0,255)} +span#textcolor931{color:rgb(0,0,255)} span#textcolor932{color:rgb(0,0,255)} -span#textcolor933{color:rgb(0,0,255)} +span#textcolor933{color:rgb(0,127,0)} span#textcolor934{color:rgb(0,127,0)} span#textcolor935{color:rgb(0,127,0)} span#textcolor936{color:rgb(0,127,0)} -span#textcolor937{color:rgb(0,127,0)} +span#textcolor937{color:rgb(43,145,175)} span#textcolor938{color:rgb(43,145,175)} -span#textcolor939{color:rgb(43,145,175)} -span#textcolor940{color:rgb(163,20,20)} +span#textcolor939{color:rgb(0,0,255)} +span#textcolor940{color:rgb(0,0,255)} span#textcolor941{color:rgb(163,20,20)} span#textcolor942{color:rgb(163,20,20)} +span#textcolor943{color:rgb(163,20,20)} +span#textcolor944{color:rgb(0,0,255)} +span#textcolor945{color:rgb(0,0,255)} +span#textcolor946{color:rgb(0,127,0)} +span#textcolor947{color:rgb(0,127,0)} +span#textcolor948{color:rgb(0,127,0)} +span#textcolor949{color:rgb(0,127,0)} +span#textcolor950{color:rgb(43,145,175)} +span#textcolor951{color:rgb(43,145,175)} +span#textcolor952{color:rgb(163,20,20)} +span#textcolor953{color:rgb(163,20,20)} +span#textcolor954{color:rgb(163,20,20)} pre#fancyvrb41{padding:5.69054pt;} pre#fancyvrb41{ border-top: solid 0.4pt; } pre#fancyvrb41{ border-left: solid 0.4pt; } @@ -1345,69 +1357,69 @@ pre#fancyvrb42{ border-top: solid 0.4pt; } pre#fancyvrb42{ border-left: solid 0.4pt; } pre#fancyvrb42{ border-bottom: solid 0.4pt; } pre#fancyvrb42{ border-right: solid 0.4pt; } -span#textcolor943{color:rgb(0,127,0)} -span#textcolor944{color:rgb(0,127,0)} -span#textcolor945{color:rgb(0,127,0)} -span#textcolor946{color:rgb(0,0,255)} -span#textcolor947{color:rgb(0,127,0)} -span#textcolor948{color:rgb(0,0,255)} -span#textcolor949{color:rgb(0,127,0)} -span#textcolor950{color:rgb(0,0,255)} -span#textcolor951{color:rgb(0,127,0)} -span#textcolor952{color:rgb(0,0,255)} -span#textcolor953{color:rgb(0,127,0)} -span#textcolor954{color:rgb(0,0,255)} span#textcolor955{color:rgb(0,127,0)} -span#textcolor956{color:rgb(0,0,255)} +span#textcolor956{color:rgb(0,127,0)} span#textcolor957{color:rgb(0,127,0)} -span#textcolor958{color:rgb(163,20,20)} -span#textcolor959{color:rgb(0,0,255)} +span#textcolor958{color:rgb(0,0,255)} +span#textcolor959{color:rgb(0,127,0)} span#textcolor960{color:rgb(0,0,255)} span#textcolor961{color:rgb(0,127,0)} span#textcolor962{color:rgb(0,0,255)} -span#textcolor963{color:rgb(43,145,175)} +span#textcolor963{color:rgb(0,127,0)} span#textcolor964{color:rgb(0,0,255)} -span#textcolor965{color:rgb(43,145,175)} +span#textcolor965{color:rgb(0,127,0)} span#textcolor966{color:rgb(0,0,255)} -span#textcolor967{color:rgb(0,0,255)} -span#textcolor968{color:rgb(43,145,175)} -span#textcolor969{color:rgb(0,0,255)} +span#textcolor967{color:rgb(0,127,0)} +span#textcolor968{color:rgb(0,0,255)} +span#textcolor969{color:rgb(0,127,0)} span#textcolor970{color:rgb(163,20,20)} -span#textcolor971{color:rgb(163,20,20)} -span#textcolor972{color:rgb(163,20,20)} -span#textcolor973{color:rgb(0,0,255)} -span#textcolor974{color:rgb(43,145,175)} -span#textcolor975{color:rgb(0,0,255)} +span#textcolor971{color:rgb(0,0,255)} +span#textcolor972{color:rgb(0,0,255)} +span#textcolor973{color:rgb(0,127,0)} +span#textcolor974{color:rgb(0,0,255)} +span#textcolor975{color:rgb(43,145,175)} span#textcolor976{color:rgb(0,0,255)} span#textcolor977{color:rgb(43,145,175)} -span#textcolor978{color:rgb(43,145,175)} -span#textcolor979{color:rgb(163,20,20)} -span#textcolor980{color:rgb(0,0,255)} +span#textcolor978{color:rgb(0,0,255)} +span#textcolor979{color:rgb(0,0,255)} +span#textcolor980{color:rgb(43,145,175)} span#textcolor981{color:rgb(0,0,255)} -span#textcolor982{color:rgb(0,0,255)} -span#textcolor983{color:rgb(43,145,175)} -span#textcolor984{color:rgb(0,0,255)} -span#textcolor985{color:rgb(43,145,175)} +span#textcolor982{color:rgb(163,20,20)} +span#textcolor983{color:rgb(163,20,20)} +span#textcolor984{color:rgb(163,20,20)} +span#textcolor985{color:rgb(0,0,255)} span#textcolor986{color:rgb(43,145,175)} -span#textcolor987{color:rgb(43,145,175)} -span#textcolor988{color:rgb(163,20,20)} -span#textcolor989{color:rgb(163,20,20)} -span#textcolor990{color:rgb(163,20,20)} +span#textcolor987{color:rgb(0,0,255)} +span#textcolor988{color:rgb(0,0,255)} +span#textcolor989{color:rgb(43,145,175)} +span#textcolor990{color:rgb(43,145,175)} span#textcolor991{color:rgb(163,20,20)} span#textcolor992{color:rgb(0,0,255)} span#textcolor993{color:rgb(0,0,255)} span#textcolor994{color:rgb(0,0,255)} -span#textcolor995{color:rgb(163,20,20)} -span#textcolor996{color:rgb(163,20,20)} -span#textcolor997{color:rgb(163,20,20)} -span#textcolor998{color:rgb(163,20,20)} -span#textcolor999{color:rgb(0,0,255)} -span#textcolor1000{color:rgb(0,0,255)} -span#textcolor1001{color:rgb(43,145,175)} -span#textcolor1002{color:rgb(43,145,175)} +span#textcolor995{color:rgb(43,145,175)} +span#textcolor996{color:rgb(0,0,255)} +span#textcolor997{color:rgb(43,145,175)} +span#textcolor998{color:rgb(43,145,175)} +span#textcolor999{color:rgb(43,145,175)} +span#textcolor1000{color:rgb(163,20,20)} +span#textcolor1001{color:rgb(163,20,20)} +span#textcolor1002{color:rgb(163,20,20)} span#textcolor1003{color:rgb(163,20,20)} -span#textcolor1004{color:rgb(163,20,20)} -span#textcolor1005{color:rgb(163,20,20)} +span#textcolor1004{color:rgb(0,0,255)} +span#textcolor1005{color:rgb(0,0,255)} +span#textcolor1006{color:rgb(0,0,255)} +span#textcolor1007{color:rgb(163,20,20)} +span#textcolor1008{color:rgb(163,20,20)} +span#textcolor1009{color:rgb(163,20,20)} +span#textcolor1010{color:rgb(163,20,20)} +span#textcolor1011{color:rgb(0,0,255)} +span#textcolor1012{color:rgb(0,0,255)} +span#textcolor1013{color:rgb(43,145,175)} +span#textcolor1014{color:rgb(43,145,175)} +span#textcolor1015{color:rgb(163,20,20)} +span#textcolor1016{color:rgb(163,20,20)} +span#textcolor1017{color:rgb(163,20,20)} pre#fancyvrb43{padding:5.69054pt;} pre#fancyvrb43{ border-top: solid 0.4pt; } pre#fancyvrb43{ border-left: solid 0.4pt; } @@ -1428,7 +1440,7 @@ pre#fancyvrb46{ border-top: solid 0.4pt; } pre#fancyvrb46{ border-left: solid 0.4pt; } pre#fancyvrb46{ border-bottom: solid 0.4pt; } pre#fancyvrb46{ border-right: solid 0.4pt; } -span#textcolor1006{color:rgb(163,20,20)} +span#textcolor1018{color:rgb(163,20,20)} pre#fancyvrb47{padding:5.69054pt;} pre#fancyvrb47{ border-top: solid 0.4pt; } pre#fancyvrb47{ border-left: solid 0.4pt; } @@ -1439,20 +1451,8 @@ pre#fancyvrb48{ border-top: solid 0.4pt; } pre#fancyvrb48{ border-left: solid 0.4pt; } pre#fancyvrb48{ border-bottom: solid 0.4pt; } pre#fancyvrb48{ border-right: solid 0.4pt; } -span#textcolor1007{color:rgb(0,127,0)} -span#textcolor1008{color:rgb(0,127,0)} -span#textcolor1009{color:rgb(0,127,0)} -span#textcolor1010{color:rgb(0,0,255)} -span#textcolor1011{color:rgb(0,127,0)} -span#textcolor1012{color:rgb(0,0,255)} -span#textcolor1013{color:rgb(0,127,0)} -span#textcolor1014{color:rgb(0,0,255)} -span#textcolor1015{color:rgb(0,127,0)} -span#textcolor1016{color:rgb(0,0,255)} -span#textcolor1017{color:rgb(0,127,0)} -span#textcolor1018{color:rgb(0,0,255)} span#textcolor1019{color:rgb(0,127,0)} -span#textcolor1020{color:rgb(0,0,255)} +span#textcolor1020{color:rgb(0,127,0)} span#textcolor1021{color:rgb(0,127,0)} span#textcolor1022{color:rgb(0,0,255)} span#textcolor1023{color:rgb(0,127,0)} @@ -1463,162 +1463,162 @@ span#textcolor1027{color:rgb(0,127,0)} span#textcolor1028{color:rgb(0,0,255)} span#textcolor1029{color:rgb(0,127,0)} span#textcolor1030{color:rgb(0,0,255)} -span#textcolor1031{color:rgb(0,0,255)} +span#textcolor1031{color:rgb(0,127,0)} span#textcolor1032{color:rgb(0,0,255)} span#textcolor1033{color:rgb(0,127,0)} -span#textcolor1034{color:rgb(0,127,0)} +span#textcolor1034{color:rgb(0,0,255)} span#textcolor1035{color:rgb(0,127,0)} -span#textcolor1036{color:rgb(0,127,0)} -span#textcolor1037{color:rgb(0,0,255)} -span#textcolor1038{color:rgb(43,145,175)} +span#textcolor1036{color:rgb(0,0,255)} +span#textcolor1037{color:rgb(0,127,0)} +span#textcolor1038{color:rgb(0,0,255)} span#textcolor1039{color:rgb(0,127,0)} -span#textcolor1040{color:rgb(0,127,0)} +span#textcolor1040{color:rgb(0,0,255)} span#textcolor1041{color:rgb(0,127,0)} span#textcolor1042{color:rgb(0,0,255)} -span#textcolor1043{color:rgb(43,145,175)} -span#textcolor1044{color:rgb(0,127,0)} +span#textcolor1043{color:rgb(0,0,255)} +span#textcolor1044{color:rgb(0,0,255)} span#textcolor1045{color:rgb(0,127,0)} span#textcolor1046{color:rgb(0,127,0)} span#textcolor1047{color:rgb(0,127,0)} span#textcolor1048{color:rgb(0,127,0)} span#textcolor1049{color:rgb(0,0,255)} span#textcolor1050{color:rgb(43,145,175)} -span#textcolor1051{color:rgb(0,0,255)} -span#textcolor1052{color:rgb(43,145,175)} +span#textcolor1051{color:rgb(0,127,0)} +span#textcolor1052{color:rgb(0,127,0)} span#textcolor1053{color:rgb(0,127,0)} span#textcolor1054{color:rgb(0,0,255)} -span#textcolor1055{color:rgb(0,0,255)} +span#textcolor1055{color:rgb(43,145,175)} span#textcolor1056{color:rgb(0,127,0)} span#textcolor1057{color:rgb(0,127,0)} span#textcolor1058{color:rgb(0,127,0)} -span#textcolor1059{color:rgb(0,0,255)} -span#textcolor1060{color:rgb(43,145,175)} +span#textcolor1059{color:rgb(0,127,0)} +span#textcolor1060{color:rgb(0,127,0)} span#textcolor1061{color:rgb(0,0,255)} -span#textcolor1062{color:rgb(0,0,255)} -span#textcolor1063{color:rgb(163,20,20)} -span#textcolor1064{color:rgb(163,20,20)} -span#textcolor1065{color:rgb(163,20,20)} -span#textcolor1066{color:rgb(0,127,0)} -span#textcolor1067{color:rgb(0,127,0)} +span#textcolor1062{color:rgb(43,145,175)} +span#textcolor1063{color:rgb(0,0,255)} +span#textcolor1064{color:rgb(43,145,175)} +span#textcolor1065{color:rgb(0,127,0)} +span#textcolor1066{color:rgb(0,0,255)} +span#textcolor1067{color:rgb(0,0,255)} span#textcolor1068{color:rgb(0,127,0)} -span#textcolor1069{color:rgb(0,0,255)} -span#textcolor1070{color:rgb(0,0,255)} -span#textcolor1071{color:rgb(0,127,0)} -span#textcolor1072{color:rgb(0,127,0)} -span#textcolor1073{color:rgb(0,127,0)} +span#textcolor1069{color:rgb(0,127,0)} +span#textcolor1070{color:rgb(0,127,0)} +span#textcolor1071{color:rgb(0,0,255)} +span#textcolor1072{color:rgb(43,145,175)} +span#textcolor1073{color:rgb(0,0,255)} span#textcolor1074{color:rgb(0,0,255)} -span#textcolor1075{color:rgb(0,0,255)} -span#textcolor1076{color:rgb(43,145,175)} -span#textcolor1077{color:rgb(0,0,255)} -span#textcolor1078{color:rgb(0,0,255)} -span#textcolor1079{color:rgb(163,20,20)} -span#textcolor1080{color:rgb(163,20,20)} -span#textcolor1081{color:rgb(163,20,20)} -span#textcolor1082{color:rgb(0,127,0)} +span#textcolor1075{color:rgb(163,20,20)} +span#textcolor1076{color:rgb(163,20,20)} +span#textcolor1077{color:rgb(163,20,20)} +span#textcolor1078{color:rgb(0,127,0)} +span#textcolor1079{color:rgb(0,127,0)} +span#textcolor1080{color:rgb(0,127,0)} +span#textcolor1081{color:rgb(0,0,255)} +span#textcolor1082{color:rgb(0,0,255)} span#textcolor1083{color:rgb(0,127,0)} span#textcolor1084{color:rgb(0,127,0)} -span#textcolor1085{color:rgb(0,0,255)} -span#textcolor1086{color:rgb(0,127,0)} -span#textcolor1087{color:rgb(0,127,0)} -span#textcolor1088{color:rgb(0,127,0)} -span#textcolor1089{color:rgb(0,127,0)} +span#textcolor1085{color:rgb(0,127,0)} +span#textcolor1086{color:rgb(0,0,255)} +span#textcolor1087{color:rgb(0,0,255)} +span#textcolor1088{color:rgb(43,145,175)} +span#textcolor1089{color:rgb(0,0,255)} span#textcolor1090{color:rgb(0,0,255)} -span#textcolor1091{color:rgb(43,145,175)} -span#textcolor1092{color:rgb(0,0,255)} -span#textcolor1093{color:rgb(0,127,0)} -span#textcolor1094{color:rgb(43,145,175)} +span#textcolor1091{color:rgb(163,20,20)} +span#textcolor1092{color:rgb(163,20,20)} +span#textcolor1093{color:rgb(163,20,20)} +span#textcolor1094{color:rgb(0,127,0)} span#textcolor1095{color:rgb(0,127,0)} span#textcolor1096{color:rgb(0,127,0)} -span#textcolor1097{color:rgb(43,145,175)} +span#textcolor1097{color:rgb(0,0,255)} span#textcolor1098{color:rgb(0,127,0)} span#textcolor1099{color:rgb(0,127,0)} span#textcolor1100{color:rgb(0,127,0)} span#textcolor1101{color:rgb(0,127,0)} -span#textcolor1102{color:rgb(43,145,175)} -span#textcolor1103{color:rgb(163,20,20)} -span#textcolor1104{color:rgb(163,20,20)} -span#textcolor1105{color:rgb(163,20,20)} -span#textcolor1106{color:rgb(0,127,0)} +span#textcolor1102{color:rgb(0,0,255)} +span#textcolor1103{color:rgb(43,145,175)} +span#textcolor1104{color:rgb(0,0,255)} +span#textcolor1105{color:rgb(0,127,0)} +span#textcolor1106{color:rgb(43,145,175)} span#textcolor1107{color:rgb(0,127,0)} span#textcolor1108{color:rgb(0,127,0)} -span#textcolor1109{color:rgb(0,127,0)} -span#textcolor1110{color:rgb(0,0,255)} -span#textcolor1111{color:rgb(0,0,255)} +span#textcolor1109{color:rgb(43,145,175)} +span#textcolor1110{color:rgb(0,127,0)} +span#textcolor1111{color:rgb(0,127,0)} span#textcolor1112{color:rgb(0,127,0)} span#textcolor1113{color:rgb(0,127,0)} -span#textcolor1114{color:rgb(0,127,0)} -span#textcolor1115{color:rgb(0,0,255)} -span#textcolor1116{color:rgb(0,127,0)} -span#textcolor1117{color:rgb(0,127,0)} +span#textcolor1114{color:rgb(43,145,175)} +span#textcolor1115{color:rgb(163,20,20)} +span#textcolor1116{color:rgb(163,20,20)} +span#textcolor1117{color:rgb(163,20,20)} span#textcolor1118{color:rgb(0,127,0)} span#textcolor1119{color:rgb(0,127,0)} span#textcolor1120{color:rgb(0,127,0)} span#textcolor1121{color:rgb(0,127,0)} -span#textcolor1122{color:rgb(0,127,0)} -span#textcolor1123{color:rgb(163,20,20)} -span#textcolor1124{color:rgb(163,20,20)} -span#textcolor1125{color:rgb(163,20,20)} +span#textcolor1122{color:rgb(0,0,255)} +span#textcolor1123{color:rgb(0,0,255)} +span#textcolor1124{color:rgb(0,127,0)} +span#textcolor1125{color:rgb(0,127,0)} span#textcolor1126{color:rgb(0,127,0)} -span#textcolor1127{color:rgb(0,127,0)} +span#textcolor1127{color:rgb(0,0,255)} span#textcolor1128{color:rgb(0,127,0)} span#textcolor1129{color:rgb(0,127,0)} -span#textcolor1130{color:rgb(0,0,255)} +span#textcolor1130{color:rgb(0,127,0)} span#textcolor1131{color:rgb(0,127,0)} span#textcolor1132{color:rgb(0,127,0)} span#textcolor1133{color:rgb(0,127,0)} span#textcolor1134{color:rgb(0,127,0)} -span#textcolor1135{color:rgb(0,0,255)} -span#textcolor1136{color:rgb(43,145,175)} -span#textcolor1137{color:rgb(0,0,255)} -span#textcolor1138{color:rgb(0,0,255)} -span#textcolor1139{color:rgb(43,145,175)} -span#textcolor1140{color:rgb(43,145,175)} -span#textcolor1141{color:rgb(43,145,175)} -span#textcolor1142{color:rgb(163,20,20)} -span#textcolor1143{color:rgb(0,0,255)} +span#textcolor1135{color:rgb(163,20,20)} +span#textcolor1136{color:rgb(163,20,20)} +span#textcolor1137{color:rgb(163,20,20)} +span#textcolor1138{color:rgb(0,127,0)} +span#textcolor1139{color:rgb(0,127,0)} +span#textcolor1140{color:rgb(0,127,0)} +span#textcolor1141{color:rgb(0,127,0)} +span#textcolor1142{color:rgb(0,0,255)} +span#textcolor1143{color:rgb(0,127,0)} span#textcolor1144{color:rgb(0,127,0)} span#textcolor1145{color:rgb(0,127,0)} span#textcolor1146{color:rgb(0,127,0)} span#textcolor1147{color:rgb(0,0,255)} -span#textcolor1148{color:rgb(0,127,0)} -span#textcolor1149{color:rgb(0,127,0)} -span#textcolor1150{color:rgb(0,127,0)} -span#textcolor1151{color:rgb(0,127,0)} -span#textcolor1152{color:rgb(0,127,0)} -span#textcolor1153{color:rgb(0,127,0)} -span#textcolor1154{color:rgb(0,127,0)} -span#textcolor1155{color:rgb(0,127,0)} +span#textcolor1148{color:rgb(43,145,175)} +span#textcolor1149{color:rgb(0,0,255)} +span#textcolor1150{color:rgb(0,0,255)} +span#textcolor1151{color:rgb(43,145,175)} +span#textcolor1152{color:rgb(43,145,175)} +span#textcolor1153{color:rgb(43,145,175)} +span#textcolor1154{color:rgb(163,20,20)} +span#textcolor1155{color:rgb(0,0,255)} span#textcolor1156{color:rgb(0,127,0)} span#textcolor1157{color:rgb(0,127,0)} -span#textcolor1158{color:rgb(43,145,175)} +span#textcolor1158{color:rgb(0,127,0)} span#textcolor1159{color:rgb(0,0,255)} span#textcolor1160{color:rgb(0,127,0)} -span#textcolor1161{color:rgb(43,145,175)} -span#textcolor1162{color:rgb(43,145,175)} +span#textcolor1161{color:rgb(0,127,0)} +span#textcolor1162{color:rgb(0,127,0)} span#textcolor1163{color:rgb(0,127,0)} -span#textcolor1164{color:rgb(43,145,175)} -span#textcolor1165{color:rgb(43,145,175)} -span#textcolor1166{color:rgb(43,145,175)} -span#textcolor1167{color:rgb(43,145,175)} -span#textcolor1168{color:rgb(43,145,175)} +span#textcolor1164{color:rgb(0,127,0)} +span#textcolor1165{color:rgb(0,127,0)} +span#textcolor1166{color:rgb(0,127,0)} +span#textcolor1167{color:rgb(0,127,0)} +span#textcolor1168{color:rgb(0,127,0)} span#textcolor1169{color:rgb(0,127,0)} -span#textcolor1170{color:rgb(0,127,0)} -span#textcolor1171{color:rgb(0,127,0)} -span#textcolor1172{color:rgb(0,0,255)} -span#textcolor1173{color:rgb(0,0,255)} -span#textcolor1174{color:rgb(0,127,0)} +span#textcolor1170{color:rgb(43,145,175)} +span#textcolor1171{color:rgb(0,0,255)} +span#textcolor1172{color:rgb(0,127,0)} +span#textcolor1173{color:rgb(43,145,175)} +span#textcolor1174{color:rgb(43,145,175)} span#textcolor1175{color:rgb(0,127,0)} -span#textcolor1176{color:rgb(0,127,0)} -span#textcolor1177{color:rgb(0,127,0)} -span#textcolor1178{color:rgb(0,127,0)} +span#textcolor1176{color:rgb(43,145,175)} +span#textcolor1177{color:rgb(43,145,175)} +span#textcolor1178{color:rgb(43,145,175)} span#textcolor1179{color:rgb(43,145,175)} -span#textcolor1180{color:rgb(0,127,0)} +span#textcolor1180{color:rgb(43,145,175)} span#textcolor1181{color:rgb(0,127,0)} span#textcolor1182{color:rgb(0,127,0)} -span#textcolor1183{color:rgb(0,0,255)} -span#textcolor1184{color:rgb(43,145,175)} +span#textcolor1183{color:rgb(0,127,0)} +span#textcolor1184{color:rgb(0,0,255)} span#textcolor1185{color:rgb(0,0,255)} -span#textcolor1186{color:rgb(0,0,255)} +span#textcolor1186{color:rgb(0,127,0)} span#textcolor1187{color:rgb(0,127,0)} span#textcolor1188{color:rgb(0,127,0)} span#textcolor1189{color:rgb(0,127,0)} @@ -1627,101 +1627,101 @@ span#textcolor1191{color:rgb(43,145,175)} span#textcolor1192{color:rgb(0,127,0)} span#textcolor1193{color:rgb(0,127,0)} span#textcolor1194{color:rgb(0,127,0)} -span#textcolor1195{color:rgb(0,127,0)} -span#textcolor1196{color:rgb(163,20,20)} -span#textcolor1197{color:rgb(43,145,175)} +span#textcolor1195{color:rgb(0,0,255)} +span#textcolor1196{color:rgb(43,145,175)} +span#textcolor1197{color:rgb(0,0,255)} span#textcolor1198{color:rgb(0,0,255)} -span#textcolor1199{color:rgb(0,0,255)} +span#textcolor1199{color:rgb(0,127,0)} span#textcolor1200{color:rgb(0,127,0)} span#textcolor1201{color:rgb(0,127,0)} span#textcolor1202{color:rgb(0,127,0)} -span#textcolor1203{color:rgb(0,127,0)} -span#textcolor1204{color:rgb(0,0,255)} -span#textcolor1205{color:rgb(0,0,255)} -span#textcolor1206{color:rgb(0,0,255)} +span#textcolor1203{color:rgb(43,145,175)} +span#textcolor1204{color:rgb(0,127,0)} +span#textcolor1205{color:rgb(0,127,0)} +span#textcolor1206{color:rgb(0,127,0)} span#textcolor1207{color:rgb(0,127,0)} -span#textcolor1208{color:rgb(0,127,0)} -span#textcolor1209{color:rgb(0,127,0)} -span#textcolor1210{color:rgb(0,127,0)} -span#textcolor1211{color:rgb(0,127,0)} +span#textcolor1208{color:rgb(163,20,20)} +span#textcolor1209{color:rgb(43,145,175)} +span#textcolor1210{color:rgb(0,0,255)} +span#textcolor1211{color:rgb(0,0,255)} span#textcolor1212{color:rgb(0,127,0)} span#textcolor1213{color:rgb(0,127,0)} span#textcolor1214{color:rgb(0,127,0)} -span#textcolor1215{color:rgb(0,0,255)} -span#textcolor1216{color:rgb(0,127,0)} -span#textcolor1217{color:rgb(0,127,0)} -span#textcolor1218{color:rgb(0,127,0)} +span#textcolor1215{color:rgb(0,127,0)} +span#textcolor1216{color:rgb(0,0,255)} +span#textcolor1217{color:rgb(0,0,255)} +span#textcolor1218{color:rgb(0,0,255)} span#textcolor1219{color:rgb(0,127,0)} -span#textcolor1220{color:rgb(43,145,175)} -span#textcolor1221{color:rgb(43,145,175)} +span#textcolor1220{color:rgb(0,127,0)} +span#textcolor1221{color:rgb(0,127,0)} span#textcolor1222{color:rgb(0,127,0)} span#textcolor1223{color:rgb(0,127,0)} span#textcolor1224{color:rgb(0,127,0)} span#textcolor1225{color:rgb(0,127,0)} span#textcolor1226{color:rgb(0,127,0)} -span#textcolor1227{color:rgb(0,127,0)} -span#textcolor1228{color:rgb(0,0,255)} -span#textcolor1229{color:rgb(163,20,20)} -span#textcolor1230{color:rgb(163,20,20)} -span#textcolor1231{color:rgb(163,20,20)} -span#textcolor1232{color:rgb(163,20,20)} -span#textcolor1233{color:rgb(0,0,255)} -span#textcolor1234{color:rgb(163,20,20)} -span#textcolor1235{color:rgb(163,20,20)} -span#textcolor1236{color:rgb(163,20,20)} -span#textcolor1237{color:rgb(0,0,255)} +span#textcolor1227{color:rgb(0,0,255)} +span#textcolor1228{color:rgb(0,127,0)} +span#textcolor1229{color:rgb(0,127,0)} +span#textcolor1230{color:rgb(0,127,0)} +span#textcolor1231{color:rgb(0,127,0)} +span#textcolor1232{color:rgb(43,145,175)} +span#textcolor1233{color:rgb(43,145,175)} +span#textcolor1234{color:rgb(0,127,0)} +span#textcolor1235{color:rgb(0,127,0)} +span#textcolor1236{color:rgb(0,127,0)} +span#textcolor1237{color:rgb(0,127,0)} span#textcolor1238{color:rgb(0,127,0)} span#textcolor1239{color:rgb(0,127,0)} -span#textcolor1240{color:rgb(0,127,0)} -span#textcolor1241{color:rgb(43,145,175)} -span#textcolor1242{color:rgb(0,127,0)} -span#textcolor1243{color:rgb(0,127,0)} -span#textcolor1244{color:rgb(0,127,0)} -span#textcolor1245{color:rgb(163,20,20)} +span#textcolor1240{color:rgb(0,0,255)} +span#textcolor1241{color:rgb(163,20,20)} +span#textcolor1242{color:rgb(163,20,20)} +span#textcolor1243{color:rgb(163,20,20)} +span#textcolor1244{color:rgb(163,20,20)} +span#textcolor1245{color:rgb(0,0,255)} +span#textcolor1246{color:rgb(163,20,20)} +span#textcolor1247{color:rgb(163,20,20)} +span#textcolor1248{color:rgb(163,20,20)} +span#textcolor1249{color:rgb(0,0,255)} +span#textcolor1250{color:rgb(0,127,0)} +span#textcolor1251{color:rgb(0,127,0)} +span#textcolor1252{color:rgb(0,127,0)} +span#textcolor1253{color:rgb(43,145,175)} +span#textcolor1254{color:rgb(0,127,0)} +span#textcolor1255{color:rgb(0,127,0)} +span#textcolor1256{color:rgb(0,127,0)} +span#textcolor1257{color:rgb(163,20,20)} pre#fancyvrb49{padding:5.69054pt;} pre#fancyvrb49{ border-top: solid 0.4pt; } pre#fancyvrb49{ border-left: solid 0.4pt; } pre#fancyvrb49{ border-bottom: solid 0.4pt; } pre#fancyvrb49{ border-right: solid 0.4pt; } -span#textcolor1246{color:rgb(0,127,0)} -span#textcolor1247{color:rgb(0,127,0)} -span#textcolor1248{color:rgb(0,127,0)} -span#textcolor1249{color:rgb(0,127,0)} -span#textcolor1250{color:rgb(0,127,0)} -span#textcolor1251{color:rgb(0,127,0)} -span#textcolor1252{color:rgb(0,127,0)} -span#textcolor1253{color:rgb(0,0,255)} -span#textcolor1254{color:rgb(0,0,255)} -span#textcolor1255{color:rgb(0,0,255)} -span#textcolor1256{color:rgb(0,127,0)} -span#textcolor1257{color:rgb(0,127,0)} span#textcolor1258{color:rgb(0,127,0)} span#textcolor1259{color:rgb(0,127,0)} span#textcolor1260{color:rgb(0,127,0)} span#textcolor1261{color:rgb(0,127,0)} -span#textcolor1262{color:rgb(0,0,255)} +span#textcolor1262{color:rgb(0,127,0)} span#textcolor1263{color:rgb(0,127,0)} span#textcolor1264{color:rgb(0,127,0)} -span#textcolor1265{color:rgb(0,127,0)} +span#textcolor1265{color:rgb(0,0,255)} span#textcolor1266{color:rgb(0,0,255)} -span#textcolor1267{color:rgb(0,127,0)} +span#textcolor1267{color:rgb(0,0,255)} span#textcolor1268{color:rgb(0,127,0)} span#textcolor1269{color:rgb(0,127,0)} span#textcolor1270{color:rgb(0,127,0)} span#textcolor1271{color:rgb(0,127,0)} span#textcolor1272{color:rgb(0,127,0)} span#textcolor1273{color:rgb(0,127,0)} -span#textcolor1274{color:rgb(0,127,0)} +span#textcolor1274{color:rgb(0,0,255)} span#textcolor1275{color:rgb(0,127,0)} span#textcolor1276{color:rgb(0,127,0)} span#textcolor1277{color:rgb(0,127,0)} -span#textcolor1278{color:rgb(0,127,0)} +span#textcolor1278{color:rgb(0,0,255)} span#textcolor1279{color:rgb(0,127,0)} span#textcolor1280{color:rgb(0,127,0)} span#textcolor1281{color:rgb(0,127,0)} span#textcolor1282{color:rgb(0,127,0)} span#textcolor1283{color:rgb(0,127,0)} -span#textcolor1284{color:rgb(0,0,255)} +span#textcolor1284{color:rgb(0,127,0)} span#textcolor1285{color:rgb(0,127,0)} span#textcolor1286{color:rgb(0,127,0)} span#textcolor1287{color:rgb(0,127,0)} @@ -1731,247 +1731,247 @@ span#textcolor1290{color:rgb(0,127,0)} span#textcolor1291{color:rgb(0,127,0)} span#textcolor1292{color:rgb(0,127,0)} span#textcolor1293{color:rgb(0,127,0)} -span#textcolor1294{color:rgb(0,0,255)} +span#textcolor1294{color:rgb(0,127,0)} span#textcolor1295{color:rgb(0,127,0)} -span#textcolor1296{color:rgb(0,127,0)} +span#textcolor1296{color:rgb(0,0,255)} span#textcolor1297{color:rgb(0,127,0)} span#textcolor1298{color:rgb(0,127,0)} span#textcolor1299{color:rgb(0,127,0)} span#textcolor1300{color:rgb(0,127,0)} span#textcolor1301{color:rgb(0,127,0)} span#textcolor1302{color:rgb(0,127,0)} -span#textcolor1303{color:rgb(0,0,255)} -span#textcolor1304{color:rgb(0,0,255)} +span#textcolor1303{color:rgb(0,127,0)} +span#textcolor1304{color:rgb(0,127,0)} +span#textcolor1305{color:rgb(0,127,0)} +span#textcolor1306{color:rgb(0,0,255)} +span#textcolor1307{color:rgb(0,127,0)} +span#textcolor1308{color:rgb(0,127,0)} +span#textcolor1309{color:rgb(0,127,0)} +span#textcolor1310{color:rgb(0,127,0)} +span#textcolor1311{color:rgb(0,127,0)} +span#textcolor1312{color:rgb(0,127,0)} +span#textcolor1313{color:rgb(0,127,0)} +span#textcolor1314{color:rgb(0,127,0)} +span#textcolor1315{color:rgb(0,0,255)} +span#textcolor1316{color:rgb(0,0,255)} pre#fancyvrb50{padding:5.69054pt;} pre#fancyvrb50{ border-top: solid 0.4pt; } pre#fancyvrb50{ border-left: solid 0.4pt; } pre#fancyvrb50{ border-bottom: solid 0.4pt; } pre#fancyvrb50{ border-right: solid 0.4pt; } -span#textcolor1305{color:rgb(0,127,0)} -span#textcolor1306{color:rgb(0,127,0)} -span#textcolor1307{color:rgb(0,127,0)} -span#textcolor1308{color:rgb(0,0,255)} -span#textcolor1309{color:rgb(0,127,0)} -span#textcolor1310{color:rgb(0,0,255)} -span#textcolor1311{color:rgb(0,127,0)} -span#textcolor1312{color:rgb(0,0,255)} -span#textcolor1313{color:rgb(0,127,0)} -span#textcolor1314{color:rgb(0,0,255)} -span#textcolor1315{color:rgb(0,127,0)} -span#textcolor1316{color:rgb(0,0,255)} span#textcolor1317{color:rgb(0,127,0)} -span#textcolor1318{color:rgb(0,0,255)} +span#textcolor1318{color:rgb(0,127,0)} span#textcolor1319{color:rgb(0,127,0)} span#textcolor1320{color:rgb(0,0,255)} span#textcolor1321{color:rgb(0,127,0)} span#textcolor1322{color:rgb(0,0,255)} -span#textcolor1323{color:rgb(43,145,175)} -span#textcolor1324{color:rgb(43,145,175)} -span#textcolor1325{color:rgb(43,145,175)} -span#textcolor1326{color:rgb(43,145,175)} +span#textcolor1323{color:rgb(0,127,0)} +span#textcolor1324{color:rgb(0,0,255)} +span#textcolor1325{color:rgb(0,127,0)} +span#textcolor1326{color:rgb(0,0,255)} span#textcolor1327{color:rgb(0,127,0)} span#textcolor1328{color:rgb(0,0,255)} -span#textcolor1329{color:rgb(0,0,255)} +span#textcolor1329{color:rgb(0,127,0)} span#textcolor1330{color:rgb(0,0,255)} -span#textcolor1331{color:rgb(0,0,255)} +span#textcolor1331{color:rgb(0,127,0)} span#textcolor1332{color:rgb(0,0,255)} -span#textcolor1333{color:rgb(0,0,255)} +span#textcolor1333{color:rgb(0,127,0)} span#textcolor1334{color:rgb(0,0,255)} -span#textcolor1335{color:rgb(0,0,255)} +span#textcolor1335{color:rgb(43,145,175)} span#textcolor1336{color:rgb(43,145,175)} span#textcolor1337{color:rgb(43,145,175)} -span#textcolor1338{color:rgb(0,0,255)} -span#textcolor1339{color:rgb(43,145,175)} -span#textcolor1340{color:rgb(43,145,175)} +span#textcolor1338{color:rgb(43,145,175)} +span#textcolor1339{color:rgb(0,127,0)} +span#textcolor1340{color:rgb(0,0,255)} span#textcolor1341{color:rgb(0,0,255)} span#textcolor1342{color:rgb(0,0,255)} span#textcolor1343{color:rgb(0,0,255)} -span#textcolor1344{color:rgb(43,145,175)} +span#textcolor1344{color:rgb(0,0,255)} span#textcolor1345{color:rgb(0,0,255)} -span#textcolor1346{color:rgb(43,145,175)} -span#textcolor1347{color:rgb(43,145,175)} -span#textcolor1348{color:rgb(0,0,255)} +span#textcolor1346{color:rgb(0,0,255)} +span#textcolor1347{color:rgb(0,0,255)} +span#textcolor1348{color:rgb(43,145,175)} span#textcolor1349{color:rgb(43,145,175)} span#textcolor1350{color:rgb(0,0,255)} span#textcolor1351{color:rgb(43,145,175)} span#textcolor1352{color:rgb(43,145,175)} -span#textcolor1353{color:rgb(43,145,175)} -span#textcolor1354{color:rgb(43,145,175)} +span#textcolor1353{color:rgb(0,0,255)} +span#textcolor1354{color:rgb(0,0,255)} span#textcolor1355{color:rgb(0,0,255)} span#textcolor1356{color:rgb(43,145,175)} -span#textcolor1357{color:rgb(43,145,175)} +span#textcolor1357{color:rgb(0,0,255)} span#textcolor1358{color:rgb(43,145,175)} -span#textcolor1359{color:rgb(0,0,255)} +span#textcolor1359{color:rgb(43,145,175)} span#textcolor1360{color:rgb(0,0,255)} -span#textcolor1361{color:rgb(0,0,255)} +span#textcolor1361{color:rgb(43,145,175)} span#textcolor1362{color:rgb(0,0,255)} -span#textcolor1363{color:rgb(0,127,0)} -span#textcolor1364{color:rgb(0,127,0)} -span#textcolor1365{color:rgb(0,127,0)} -span#textcolor1366{color:rgb(0,127,0)} -span#textcolor1367{color:rgb(0,127,0)} -span#textcolor1368{color:rgb(0,127,0)} -span#textcolor1369{color:rgb(0,127,0)} -span#textcolor1370{color:rgb(0,127,0)} -span#textcolor1371{color:rgb(0,127,0)} -span#textcolor1372{color:rgb(0,127,0)} +span#textcolor1363{color:rgb(43,145,175)} +span#textcolor1364{color:rgb(43,145,175)} +span#textcolor1365{color:rgb(43,145,175)} +span#textcolor1366{color:rgb(43,145,175)} +span#textcolor1367{color:rgb(0,0,255)} +span#textcolor1368{color:rgb(43,145,175)} +span#textcolor1369{color:rgb(43,145,175)} +span#textcolor1370{color:rgb(43,145,175)} +span#textcolor1371{color:rgb(0,0,255)} +span#textcolor1372{color:rgb(0,0,255)} span#textcolor1373{color:rgb(0,0,255)} -span#textcolor1374{color:rgb(43,145,175)} -span#textcolor1375{color:rgb(0,0,255)} -span#textcolor1376{color:rgb(0,0,255)} -span#textcolor1377{color:rgb(163,20,20)} -span#textcolor1378{color:rgb(163,20,20)} -span#textcolor1379{color:rgb(163,20,20)} -span#textcolor1380{color:rgb(0,0,255)} -span#textcolor1381{color:rgb(0,0,255)} +span#textcolor1374{color:rgb(0,0,255)} +span#textcolor1375{color:rgb(0,127,0)} +span#textcolor1376{color:rgb(0,127,0)} +span#textcolor1377{color:rgb(0,127,0)} +span#textcolor1378{color:rgb(0,127,0)} +span#textcolor1379{color:rgb(0,127,0)} +span#textcolor1380{color:rgb(0,127,0)} +span#textcolor1381{color:rgb(0,127,0)} span#textcolor1382{color:rgb(0,127,0)} span#textcolor1383{color:rgb(0,127,0)} span#textcolor1384{color:rgb(0,127,0)} -span#textcolor1385{color:rgb(0,127,0)} -span#textcolor1386{color:rgb(0,127,0)} -span#textcolor1387{color:rgb(0,127,0)} +span#textcolor1385{color:rgb(0,0,255)} +span#textcolor1386{color:rgb(43,145,175)} +span#textcolor1387{color:rgb(0,0,255)} span#textcolor1388{color:rgb(0,0,255)} -span#textcolor1389{color:rgb(43,145,175)} -span#textcolor1390{color:rgb(0,0,255)} -span#textcolor1391{color:rgb(0,0,255)} +span#textcolor1389{color:rgb(163,20,20)} +span#textcolor1390{color:rgb(163,20,20)} +span#textcolor1391{color:rgb(163,20,20)} span#textcolor1392{color:rgb(0,0,255)} span#textcolor1393{color:rgb(0,0,255)} -span#textcolor1394{color:rgb(43,145,175)} -span#textcolor1395{color:rgb(0,0,255)} -span#textcolor1396{color:rgb(0,0,255)} +span#textcolor1394{color:rgb(0,127,0)} +span#textcolor1395{color:rgb(0,127,0)} +span#textcolor1396{color:rgb(0,127,0)} span#textcolor1397{color:rgb(0,127,0)} span#textcolor1398{color:rgb(0,127,0)} span#textcolor1399{color:rgb(0,127,0)} -span#textcolor1400{color:rgb(0,127,0)} -span#textcolor1401{color:rgb(0,0,255)} +span#textcolor1400{color:rgb(0,0,255)} +span#textcolor1401{color:rgb(43,145,175)} span#textcolor1402{color:rgb(0,0,255)} span#textcolor1403{color:rgb(0,0,255)} -span#textcolor1404{color:rgb(43,145,175)} +span#textcolor1404{color:rgb(0,0,255)} span#textcolor1405{color:rgb(0,0,255)} span#textcolor1406{color:rgb(43,145,175)} -span#textcolor1407{color:rgb(43,145,175)} +span#textcolor1407{color:rgb(0,0,255)} span#textcolor1408{color:rgb(0,0,255)} -span#textcolor1409{color:rgb(43,145,175)} -span#textcolor1410{color:rgb(43,145,175)} -span#textcolor1411{color:rgb(43,145,175)} -span#textcolor1412{color:rgb(43,145,175)} +span#textcolor1409{color:rgb(0,127,0)} +span#textcolor1410{color:rgb(0,127,0)} +span#textcolor1411{color:rgb(0,127,0)} +span#textcolor1412{color:rgb(0,127,0)} span#textcolor1413{color:rgb(0,0,255)} span#textcolor1414{color:rgb(0,0,255)} span#textcolor1415{color:rgb(0,0,255)} -span#textcolor1416{color:rgb(0,0,255)} +span#textcolor1416{color:rgb(43,145,175)} span#textcolor1417{color:rgb(0,0,255)} span#textcolor1418{color:rgb(43,145,175)} -span#textcolor1419{color:rgb(0,0,255)} +span#textcolor1419{color:rgb(43,145,175)} span#textcolor1420{color:rgb(0,0,255)} -span#textcolor1421{color:rgb(163,20,20)} -span#textcolor1422{color:rgb(163,20,20)} -span#textcolor1423{color:rgb(163,20,20)} -span#textcolor1424{color:rgb(0,0,255)} +span#textcolor1421{color:rgb(43,145,175)} +span#textcolor1422{color:rgb(43,145,175)} +span#textcolor1423{color:rgb(43,145,175)} +span#textcolor1424{color:rgb(43,145,175)} span#textcolor1425{color:rgb(0,0,255)} span#textcolor1426{color:rgb(0,0,255)} -span#textcolor1427{color:rgb(43,145,175)} +span#textcolor1427{color:rgb(0,0,255)} span#textcolor1428{color:rgb(0,0,255)} span#textcolor1429{color:rgb(0,0,255)} -span#textcolor1430{color:rgb(0,0,255)} -span#textcolor1431{color:rgb(163,20,20)} -span#textcolor1432{color:rgb(163,20,20)} +span#textcolor1430{color:rgb(43,145,175)} +span#textcolor1431{color:rgb(0,0,255)} +span#textcolor1432{color:rgb(0,0,255)} span#textcolor1433{color:rgb(163,20,20)} -span#textcolor1434{color:rgb(0,0,255)} -span#textcolor1435{color:rgb(0,0,255)} +span#textcolor1434{color:rgb(163,20,20)} +span#textcolor1435{color:rgb(163,20,20)} span#textcolor1436{color:rgb(0,0,255)} span#textcolor1437{color:rgb(0,0,255)} span#textcolor1438{color:rgb(0,0,255)} -span#textcolor1439{color:rgb(0,0,255)} +span#textcolor1439{color:rgb(43,145,175)} span#textcolor1440{color:rgb(0,0,255)} -span#textcolor1441{color:rgb(43,145,175)} -span#textcolor1442{color:rgb(43,145,175)} -span#textcolor1443{color:rgb(43,145,175)} -span#textcolor1444{color:rgb(43,145,175)} -span#textcolor1445{color:rgb(43,145,175)} +span#textcolor1441{color:rgb(0,0,255)} +span#textcolor1442{color:rgb(0,0,255)} +span#textcolor1443{color:rgb(163,20,20)} +span#textcolor1444{color:rgb(163,20,20)} +span#textcolor1445{color:rgb(163,20,20)} span#textcolor1446{color:rgb(0,0,255)} span#textcolor1447{color:rgb(0,0,255)} span#textcolor1448{color:rgb(0,0,255)} span#textcolor1449{color:rgb(0,0,255)} -span#textcolor1450{color:rgb(163,20,20)} -span#textcolor1451{color:rgb(163,20,20)} -span#textcolor1452{color:rgb(163,20,20)} -span#textcolor1453{color:rgb(0,0,255)} -span#textcolor1454{color:rgb(0,0,255)} -span#textcolor1455{color:rgb(0,0,255)} -span#textcolor1456{color:rgb(0,0,255)} -span#textcolor1457{color:rgb(0,0,255)} -span#textcolor1458{color:rgb(43,145,175)} -span#textcolor1459{color:rgb(43,145,175)} -span#textcolor1460{color:rgb(43,145,175)} -span#textcolor1461{color:rgb(163,20,20)} +span#textcolor1450{color:rgb(0,0,255)} +span#textcolor1451{color:rgb(0,0,255)} +span#textcolor1452{color:rgb(0,0,255)} +span#textcolor1453{color:rgb(43,145,175)} +span#textcolor1454{color:rgb(43,145,175)} +span#textcolor1455{color:rgb(43,145,175)} +span#textcolor1456{color:rgb(43,145,175)} +span#textcolor1457{color:rgb(43,145,175)} +span#textcolor1458{color:rgb(0,0,255)} +span#textcolor1459{color:rgb(0,0,255)} +span#textcolor1460{color:rgb(0,0,255)} +span#textcolor1461{color:rgb(0,0,255)} span#textcolor1462{color:rgb(163,20,20)} span#textcolor1463{color:rgb(163,20,20)} span#textcolor1464{color:rgb(163,20,20)} -span#textcolor1465{color:rgb(163,20,20)} +span#textcolor1465{color:rgb(0,0,255)} +span#textcolor1466{color:rgb(0,0,255)} +span#textcolor1467{color:rgb(0,0,255)} +span#textcolor1468{color:rgb(0,0,255)} +span#textcolor1469{color:rgb(0,0,255)} +span#textcolor1470{color:rgb(43,145,175)} +span#textcolor1471{color:rgb(43,145,175)} +span#textcolor1472{color:rgb(43,145,175)} +span#textcolor1473{color:rgb(163,20,20)} +span#textcolor1474{color:rgb(163,20,20)} +span#textcolor1475{color:rgb(163,20,20)} +span#textcolor1476{color:rgb(163,20,20)} +span#textcolor1477{color:rgb(163,20,20)} pre#fancyvrb51{padding:5.69054pt;} pre#fancyvrb51{ border-top: solid 0.4pt; } pre#fancyvrb51{ border-left: solid 0.4pt; } pre#fancyvrb51{ border-bottom: solid 0.4pt; } pre#fancyvrb51{ border-right: solid 0.4pt; } -span#textcolor1466{color:rgb(0,127,0)} -span#textcolor1467{color:rgb(0,127,0)} -span#textcolor1468{color:rgb(0,127,0)} -span#textcolor1469{color:rgb(0,127,0)} -span#textcolor1470{color:rgb(0,127,0)} -span#textcolor1471{color:rgb(0,127,0)} -span#textcolor1472{color:rgb(0,127,0)} -span#textcolor1473{color:rgb(0,127,0)} -span#textcolor1474{color:rgb(0,127,0)} -span#textcolor1475{color:rgb(0,127,0)} -span#textcolor1476{color:rgb(0,127,0)} -span#textcolor1477{color:rgb(0,0,255)} span#textcolor1478{color:rgb(0,127,0)} -span#textcolor1479{color:rgb(0,0,255)} +span#textcolor1479{color:rgb(0,127,0)} span#textcolor1480{color:rgb(0,127,0)} -span#textcolor1481{color:rgb(0,0,255)} +span#textcolor1481{color:rgb(0,127,0)} span#textcolor1482{color:rgb(0,127,0)} -span#textcolor1483{color:rgb(0,0,255)} +span#textcolor1483{color:rgb(0,127,0)} span#textcolor1484{color:rgb(0,127,0)} -span#textcolor1485{color:rgb(0,0,255)} +span#textcolor1485{color:rgb(0,127,0)} span#textcolor1486{color:rgb(0,127,0)} -span#textcolor1487{color:rgb(0,0,255)} +span#textcolor1487{color:rgb(0,127,0)} span#textcolor1488{color:rgb(0,127,0)} -span#textcolor1489{color:rgb(0,127,0)} +span#textcolor1489{color:rgb(0,0,255)} span#textcolor1490{color:rgb(0,127,0)} -span#textcolor1491{color:rgb(0,127,0)} +span#textcolor1491{color:rgb(0,0,255)} span#textcolor1492{color:rgb(0,127,0)} span#textcolor1493{color:rgb(0,0,255)} span#textcolor1494{color:rgb(0,127,0)} span#textcolor1495{color:rgb(0,0,255)} span#textcolor1496{color:rgb(0,127,0)} -span#textcolor1497{color:rgb(43,145,175)} -span#textcolor1498{color:rgb(43,145,175)} -span#textcolor1499{color:rgb(43,145,175)} -span#textcolor1500{color:rgb(43,145,175)} +span#textcolor1497{color:rgb(0,0,255)} +span#textcolor1498{color:rgb(0,127,0)} +span#textcolor1499{color:rgb(0,0,255)} +span#textcolor1500{color:rgb(0,127,0)} span#textcolor1501{color:rgb(0,127,0)} span#textcolor1502{color:rgb(0,127,0)} span#textcolor1503{color:rgb(0,127,0)} span#textcolor1504{color:rgb(0,127,0)} span#textcolor1505{color:rgb(0,0,255)} -span#textcolor1506{color:rgb(43,145,175)} -span#textcolor1507{color:rgb(43,145,175)} +span#textcolor1506{color:rgb(0,127,0)} +span#textcolor1507{color:rgb(0,0,255)} span#textcolor1508{color:rgb(0,127,0)} -span#textcolor1509{color:rgb(0,127,0)} -span#textcolor1510{color:rgb(0,127,0)} -span#textcolor1511{color:rgb(0,127,0)} -span#textcolor1512{color:rgb(0,127,0)} +span#textcolor1509{color:rgb(43,145,175)} +span#textcolor1510{color:rgb(43,145,175)} +span#textcolor1511{color:rgb(43,145,175)} +span#textcolor1512{color:rgb(43,145,175)} span#textcolor1513{color:rgb(0,127,0)} span#textcolor1514{color:rgb(0,127,0)} span#textcolor1515{color:rgb(0,127,0)} span#textcolor1516{color:rgb(0,127,0)} -span#textcolor1517{color:rgb(0,127,0)} -span#textcolor1518{color:rgb(0,127,0)} -span#textcolor1519{color:rgb(0,127,0)} +span#textcolor1517{color:rgb(0,0,255)} +span#textcolor1518{color:rgb(43,145,175)} +span#textcolor1519{color:rgb(43,145,175)} span#textcolor1520{color:rgb(0,127,0)} -span#textcolor1521{color:rgb(0,0,255)} -span#textcolor1522{color:rgb(43,145,175)} -span#textcolor1523{color:rgb(43,145,175)} -span#textcolor1524{color:rgb(43,145,175)} +span#textcolor1521{color:rgb(0,127,0)} +span#textcolor1522{color:rgb(0,127,0)} +span#textcolor1523{color:rgb(0,127,0)} +span#textcolor1524{color:rgb(0,127,0)} span#textcolor1525{color:rgb(0,127,0)} span#textcolor1526{color:rgb(0,127,0)} span#textcolor1527{color:rgb(0,127,0)} @@ -1980,89 +1980,101 @@ span#textcolor1529{color:rgb(0,127,0)} span#textcolor1530{color:rgb(0,127,0)} span#textcolor1531{color:rgb(0,127,0)} span#textcolor1532{color:rgb(0,127,0)} -span#textcolor1533{color:rgb(0,127,0)} -span#textcolor1534{color:rgb(0,127,0)} -span#textcolor1535{color:rgb(0,127,0)} -span#textcolor1536{color:rgb(0,127,0)} +span#textcolor1533{color:rgb(0,0,255)} +span#textcolor1534{color:rgb(43,145,175)} +span#textcolor1535{color:rgb(43,145,175)} +span#textcolor1536{color:rgb(43,145,175)} span#textcolor1537{color:rgb(0,127,0)} span#textcolor1538{color:rgb(0,127,0)} -span#textcolor1539{color:rgb(43,145,175)} -span#textcolor1540{color:rgb(0,0,255)} -span#textcolor1541{color:rgb(43,145,175)} -span#textcolor1542{color:rgb(43,145,175)} -span#textcolor1543{color:rgb(43,145,175)} -span#textcolor1544{color:rgb(43,145,175)} -span#textcolor1545{color:rgb(43,145,175)} +span#textcolor1539{color:rgb(0,127,0)} +span#textcolor1540{color:rgb(0,127,0)} +span#textcolor1541{color:rgb(0,127,0)} +span#textcolor1542{color:rgb(0,127,0)} +span#textcolor1543{color:rgb(0,127,0)} +span#textcolor1544{color:rgb(0,127,0)} +span#textcolor1545{color:rgb(0,127,0)} span#textcolor1546{color:rgb(0,127,0)} span#textcolor1547{color:rgb(0,127,0)} span#textcolor1548{color:rgb(0,127,0)} -span#textcolor1549{color:rgb(163,20,20)} -span#textcolor1550{color:rgb(0,0,255)} -span#textcolor1551{color:rgb(163,20,20)} +span#textcolor1549{color:rgb(0,127,0)} +span#textcolor1550{color:rgb(0,127,0)} +span#textcolor1551{color:rgb(43,145,175)} span#textcolor1552{color:rgb(0,0,255)} -span#textcolor1553{color:rgb(163,20,20)} -span#textcolor1554{color:rgb(163,20,20)} -span#textcolor1555{color:rgb(163,20,20)} -span#textcolor1556{color:rgb(0,127,0)} -span#textcolor1557{color:rgb(0,127,0)} +span#textcolor1553{color:rgb(43,145,175)} +span#textcolor1554{color:rgb(43,145,175)} +span#textcolor1555{color:rgb(43,145,175)} +span#textcolor1556{color:rgb(43,145,175)} +span#textcolor1557{color:rgb(43,145,175)} span#textcolor1558{color:rgb(0,127,0)} span#textcolor1559{color:rgb(0,127,0)} -span#textcolor1560{color:rgb(0,0,255)} -span#textcolor1561{color:rgb(0,0,255)} -span#textcolor1562{color:rgb(43,145,175)} -span#textcolor1563{color:rgb(43,145,175)} -span#textcolor1564{color:rgb(43,145,175)} -span#textcolor1565{color:rgb(43,145,175)} -span#textcolor1566{color:rgb(43,145,175)} -span#textcolor1567{color:rgb(43,145,175)} -span#textcolor1568{color:rgb(43,145,175)} -span#textcolor1569{color:rgb(43,145,175)} -span#textcolor1570{color:rgb(0,0,255)} -span#textcolor1571{color:rgb(43,145,175)} -span#textcolor1572{color:rgb(43,145,175)} +span#textcolor1560{color:rgb(0,127,0)} +span#textcolor1561{color:rgb(163,20,20)} +span#textcolor1562{color:rgb(0,0,255)} +span#textcolor1563{color:rgb(163,20,20)} +span#textcolor1564{color:rgb(0,0,255)} +span#textcolor1565{color:rgb(163,20,20)} +span#textcolor1566{color:rgb(163,20,20)} +span#textcolor1567{color:rgb(163,20,20)} +span#textcolor1568{color:rgb(0,127,0)} +span#textcolor1569{color:rgb(0,127,0)} +span#textcolor1570{color:rgb(0,127,0)} +span#textcolor1571{color:rgb(0,127,0)} +span#textcolor1572{color:rgb(0,0,255)} span#textcolor1573{color:rgb(0,0,255)} span#textcolor1574{color:rgb(43,145,175)} span#textcolor1575{color:rgb(43,145,175)} -span#textcolor1576{color:rgb(0,0,255)} -span#textcolor1577{color:rgb(0,0,255)} +span#textcolor1576{color:rgb(43,145,175)} +span#textcolor1577{color:rgb(43,145,175)} span#textcolor1578{color:rgb(43,145,175)} -span#textcolor1579{color:rgb(0,0,255)} -span#textcolor1580{color:rgb(0,0,255)} +span#textcolor1579{color:rgb(43,145,175)} +span#textcolor1580{color:rgb(43,145,175)} span#textcolor1581{color:rgb(43,145,175)} -span#textcolor1582{color:rgb(43,145,175)} -span#textcolor1583{color:rgb(0,0,255)} -span#textcolor1584{color:rgb(0,0,255)} -span#textcolor1585{color:rgb(0,127,0)} +span#textcolor1582{color:rgb(0,0,255)} +span#textcolor1583{color:rgb(43,145,175)} +span#textcolor1584{color:rgb(43,145,175)} +span#textcolor1585{color:rgb(0,0,255)} span#textcolor1586{color:rgb(43,145,175)} -span#textcolor1587{color:rgb(0,127,0)} -span#textcolor1588{color:rgb(43,145,175)} -span#textcolor1589{color:rgb(43,145,175)} -span#textcolor1590{color:rgb(163,20,20)} -span#textcolor1591{color:rgb(163,20,20)} -span#textcolor1592{color:rgb(163,20,20)} -span#textcolor1593{color:rgb(0,0,255)} -span#textcolor1594{color:rgb(0,0,255)} -span#textcolor1595{color:rgb(43,145,175)} -span#textcolor1596{color:rgb(43,145,175)} -span#textcolor1597{color:rgb(0,0,255)} -span#textcolor1598{color:rgb(0,0,255)} +span#textcolor1587{color:rgb(43,145,175)} +span#textcolor1588{color:rgb(0,0,255)} +span#textcolor1589{color:rgb(0,0,255)} +span#textcolor1590{color:rgb(43,145,175)} +span#textcolor1591{color:rgb(0,0,255)} +span#textcolor1592{color:rgb(0,0,255)} +span#textcolor1593{color:rgb(43,145,175)} +span#textcolor1594{color:rgb(43,145,175)} +span#textcolor1595{color:rgb(0,0,255)} +span#textcolor1596{color:rgb(0,0,255)} +span#textcolor1597{color:rgb(0,127,0)} +span#textcolor1598{color:rgb(43,145,175)} span#textcolor1599{color:rgb(0,127,0)} -span#textcolor1600{color:rgb(0,127,0)} -span#textcolor1601{color:rgb(0,127,0)} -span#textcolor1602{color:rgb(0,0,255)} -span#textcolor1603{color:rgb(43,145,175)} -span#textcolor1604{color:rgb(43,145,175)} -span#textcolor1605{color:rgb(163,20,20)} -span#textcolor1606{color:rgb(163,20,20)} -span#textcolor1607{color:rgb(163,20,20)} -span#textcolor1608{color:rgb(163,20,20)} -span#textcolor1609{color:rgb(163,20,20)} -span#textcolor1610{color:rgb(163,20,20)} -span#textcolor1611{color:rgb(163,20,20)} -span#textcolor1612{color:rgb(163,20,20)} -span#textcolor1613{color:rgb(43,145,175)} -span#textcolor1614{color:rgb(43,145,175)} -span#textcolor1615{color:rgb(163,20,20)} +span#textcolor1600{color:rgb(43,145,175)} +span#textcolor1601{color:rgb(43,145,175)} +span#textcolor1602{color:rgb(163,20,20)} +span#textcolor1603{color:rgb(163,20,20)} +span#textcolor1604{color:rgb(163,20,20)} +span#textcolor1605{color:rgb(0,0,255)} +span#textcolor1606{color:rgb(0,0,255)} +span#textcolor1607{color:rgb(43,145,175)} +span#textcolor1608{color:rgb(43,145,175)} +span#textcolor1609{color:rgb(0,0,255)} +span#textcolor1610{color:rgb(0,0,255)} +span#textcolor1611{color:rgb(0,127,0)} +span#textcolor1612{color:rgb(0,127,0)} +span#textcolor1613{color:rgb(0,127,0)} +span#textcolor1614{color:rgb(0,0,255)} +span#textcolor1615{color:rgb(43,145,175)} +span#textcolor1616{color:rgb(43,145,175)} +span#textcolor1617{color:rgb(163,20,20)} +span#textcolor1618{color:rgb(163,20,20)} +span#textcolor1619{color:rgb(163,20,20)} +span#textcolor1620{color:rgb(163,20,20)} +span#textcolor1621{color:rgb(163,20,20)} +span#textcolor1622{color:rgb(163,20,20)} +span#textcolor1623{color:rgb(163,20,20)} +span#textcolor1624{color:rgb(163,20,20)} +span#textcolor1625{color:rgb(43,145,175)} +span#textcolor1626{color:rgb(43,145,175)} +span#textcolor1627{color:rgb(163,20,20)} pre#fancyvrb52{padding:5.69054pt;} pre#fancyvrb52{ border-top: solid 0.4pt; } pre#fancyvrb52{ border-left: solid 0.4pt; } @@ -2073,138 +2085,126 @@ pre#fancyvrb53{ border-top: solid 0.4pt; } pre#fancyvrb53{ border-left: solid 0.4pt; } pre#fancyvrb53{ border-bottom: solid 0.4pt; } pre#fancyvrb53{ border-right: solid 0.4pt; } -span#textcolor1616{color:rgb(0,127,0)} -span#textcolor1617{color:rgb(0,127,0)} -span#textcolor1618{color:rgb(0,127,0)} -span#textcolor1619{color:rgb(0,127,0)} -span#textcolor1620{color:rgb(0,0,255)} -span#textcolor1621{color:rgb(0,127,0)} -span#textcolor1622{color:rgb(0,0,255)} -span#textcolor1623{color:rgb(0,127,0)} -span#textcolor1624{color:rgb(0,0,255)} -span#textcolor1625{color:rgb(0,127,0)} -span#textcolor1626{color:rgb(0,0,255)} -span#textcolor1627{color:rgb(0,127,0)} -#colorbox1628{border: solid 1px rgb(255,0,0);} -#colorbox1628{background-color: rgb(255,255,255);} -span#textcolor1629{color:rgb(0,0,255)} +span#textcolor1628{color:rgb(0,127,0)} +span#textcolor1629{color:rgb(0,127,0)} span#textcolor1630{color:rgb(0,127,0)} -span#textcolor1631{color:rgb(0,0,255)} -span#textcolor1632{color:rgb(0,127,0)} -span#textcolor1633{color:rgb(0,0,255)} +span#textcolor1631{color:rgb(0,127,0)} +span#textcolor1632{color:rgb(0,0,255)} +span#textcolor1633{color:rgb(0,127,0)} span#textcolor1634{color:rgb(0,0,255)} -span#textcolor1635{color:rgb(0,0,255)} -span#textcolor1636{color:rgb(0,127,0)} +span#textcolor1635{color:rgb(0,127,0)} +span#textcolor1636{color:rgb(0,0,255)} span#textcolor1637{color:rgb(0,127,0)} -span#textcolor1638{color:rgb(0,127,0)} +span#textcolor1638{color:rgb(0,0,255)} span#textcolor1639{color:rgb(0,127,0)} -span#textcolor1640{color:rgb(0,127,0)} -span#textcolor1641{color:rgb(0,127,0)} +#colorbox1640{border: solid 1px rgb(255,0,0);} +#colorbox1640{background-color: rgb(255,255,255);} +span#textcolor1641{color:rgb(0,0,255)} span#textcolor1642{color:rgb(0,127,0)} span#textcolor1643{color:rgb(0,0,255)} -span#textcolor1644{color:rgb(0,0,255)} -span#textcolor1645{color:rgb(43,145,175)} +span#textcolor1644{color:rgb(0,127,0)} +span#textcolor1645{color:rgb(0,0,255)} span#textcolor1646{color:rgb(0,0,255)} span#textcolor1647{color:rgb(0,0,255)} -span#textcolor1648{color:rgb(0,0,255)} +span#textcolor1648{color:rgb(0,127,0)} span#textcolor1649{color:rgb(0,127,0)} span#textcolor1650{color:rgb(0,127,0)} span#textcolor1651{color:rgb(0,127,0)} span#textcolor1652{color:rgb(0,127,0)} span#textcolor1653{color:rgb(0,127,0)} -span#textcolor1654{color:rgb(0,0,255)} -span#textcolor1655{color:rgb(43,145,175)} +span#textcolor1654{color:rgb(0,127,0)} +span#textcolor1655{color:rgb(0,0,255)} span#textcolor1656{color:rgb(0,0,255)} -span#textcolor1657{color:rgb(0,127,0)} -span#textcolor1658{color:rgb(43,145,175)} -span#textcolor1659{color:rgb(0,127,0)} -span#textcolor1660{color:rgb(0,127,0)} -#colorbox1661{border: solid 1px rgb(255,0,0);} -#colorbox1661{background-color: rgb(255,255,255);} -span#textcolor1662{color:rgb(43,145,175)} +span#textcolor1657{color:rgb(43,145,175)} +span#textcolor1658{color:rgb(0,0,255)} +span#textcolor1659{color:rgb(0,0,255)} +span#textcolor1660{color:rgb(0,0,255)} +span#textcolor1661{color:rgb(0,127,0)} +span#textcolor1662{color:rgb(0,127,0)} span#textcolor1663{color:rgb(0,127,0)} -span#textcolor1664{color:rgb(0,0,255)} -span#textcolor1665{color:rgb(43,145,175)} -span#textcolor1666{color:rgb(43,145,175)} +span#textcolor1664{color:rgb(0,127,0)} +span#textcolor1665{color:rgb(0,127,0)} +span#textcolor1666{color:rgb(0,0,255)} span#textcolor1667{color:rgb(43,145,175)} -span#textcolor1668{color:rgb(0,127,0)} +span#textcolor1668{color:rgb(0,0,255)} span#textcolor1669{color:rgb(0,127,0)} -span#textcolor1670{color:rgb(0,127,0)} +span#textcolor1670{color:rgb(43,145,175)} span#textcolor1671{color:rgb(0,127,0)} -span#textcolor1672{color:rgb(0,0,255)} -span#textcolor1673{color:rgb(0,0,255)} -span#textcolor1674{color:rgb(0,127,0)} +span#textcolor1672{color:rgb(0,127,0)} +#colorbox1673{border: solid 1px rgb(255,0,0);} +#colorbox1673{background-color: rgb(255,255,255);} +span#textcolor1674{color:rgb(43,145,175)} span#textcolor1675{color:rgb(0,127,0)} -span#textcolor1676{color:rgb(0,127,0)} -span#textcolor1677{color:rgb(0,127,0)} -span#textcolor1678{color:rgb(163,20,20)} -span#textcolor1679{color:rgb(163,20,20)} -span#textcolor1680{color:rgb(163,20,20)} -span#textcolor1681{color:rgb(0,0,255)} -span#textcolor1682{color:rgb(0,0,255)} +span#textcolor1676{color:rgb(0,0,255)} +span#textcolor1677{color:rgb(43,145,175)} +span#textcolor1678{color:rgb(43,145,175)} +span#textcolor1679{color:rgb(43,145,175)} +span#textcolor1680{color:rgb(0,127,0)} +span#textcolor1681{color:rgb(0,127,0)} +span#textcolor1682{color:rgb(0,127,0)} span#textcolor1683{color:rgb(0,127,0)} -span#textcolor1684{color:rgb(0,127,0)} -span#textcolor1685{color:rgb(0,127,0)} +span#textcolor1684{color:rgb(0,0,255)} +span#textcolor1685{color:rgb(0,0,255)} span#textcolor1686{color:rgb(0,127,0)} span#textcolor1687{color:rgb(0,127,0)} -span#textcolor1688{color:rgb(0,0,255)} -span#textcolor1689{color:rgb(43,145,175)} -span#textcolor1690{color:rgb(0,0,255)} -span#textcolor1691{color:rgb(0,127,0)} -span#textcolor1692{color:rgb(0,0,255)} -span#textcolor1693{color:rgb(43,145,175)} -span#textcolor1694{color:rgb(0,127,0)} -span#textcolor1695{color:rgb(43,145,175)} +span#textcolor1688{color:rgb(0,127,0)} +span#textcolor1689{color:rgb(0,127,0)} +span#textcolor1690{color:rgb(163,20,20)} +span#textcolor1691{color:rgb(163,20,20)} +span#textcolor1692{color:rgb(163,20,20)} +span#textcolor1693{color:rgb(0,0,255)} +span#textcolor1694{color:rgb(0,0,255)} +span#textcolor1695{color:rgb(0,127,0)} span#textcolor1696{color:rgb(0,127,0)} span#textcolor1697{color:rgb(0,127,0)} -span#textcolor1698{color:rgb(43,145,175)} +span#textcolor1698{color:rgb(0,127,0)} span#textcolor1699{color:rgb(0,127,0)} -span#textcolor1700{color:rgb(0,127,0)} -span#textcolor1701{color:rgb(0,127,0)} -span#textcolor1702{color:rgb(0,127,0)} -span#textcolor1703{color:rgb(0,0,255)} -span#textcolor1704{color:rgb(0,127,0)} -span#textcolor1705{color:rgb(0,127,0)} +span#textcolor1700{color:rgb(0,0,255)} +span#textcolor1701{color:rgb(43,145,175)} +span#textcolor1702{color:rgb(0,0,255)} +span#textcolor1703{color:rgb(0,127,0)} +span#textcolor1704{color:rgb(0,0,255)} +span#textcolor1705{color:rgb(43,145,175)} span#textcolor1706{color:rgb(0,127,0)} -span#textcolor1707{color:rgb(163,20,20)} +span#textcolor1707{color:rgb(43,145,175)} span#textcolor1708{color:rgb(0,127,0)} span#textcolor1709{color:rgb(0,127,0)} -span#textcolor1710{color:rgb(0,127,0)} -span#textcolor1711{color:rgb(0,0,255)} +span#textcolor1710{color:rgb(43,145,175)} +span#textcolor1711{color:rgb(0,127,0)} span#textcolor1712{color:rgb(0,127,0)} span#textcolor1713{color:rgb(0,127,0)} span#textcolor1714{color:rgb(0,127,0)} -span#textcolor1715{color:rgb(43,145,175)} +span#textcolor1715{color:rgb(0,0,255)} span#textcolor1716{color:rgb(0,127,0)} span#textcolor1717{color:rgb(0,127,0)} span#textcolor1718{color:rgb(0,127,0)} -span#textcolor1719{color:rgb(0,127,0)} +span#textcolor1719{color:rgb(163,20,20)} span#textcolor1720{color:rgb(0,127,0)} span#textcolor1721{color:rgb(0,127,0)} -span#textcolor1722{color:rgb(0,0,255)} -span#textcolor1723{color:rgb(43,145,175)} -span#textcolor1724{color:rgb(0,0,255)} -span#textcolor1725{color:rgb(0,0,255)} +span#textcolor1722{color:rgb(0,127,0)} +span#textcolor1723{color:rgb(0,0,255)} +span#textcolor1724{color:rgb(0,127,0)} +span#textcolor1725{color:rgb(0,127,0)} span#textcolor1726{color:rgb(0,127,0)} -span#textcolor1727{color:rgb(0,127,0)} +span#textcolor1727{color:rgb(43,145,175)} span#textcolor1728{color:rgb(0,127,0)} span#textcolor1729{color:rgb(0,127,0)} span#textcolor1730{color:rgb(0,127,0)} span#textcolor1731{color:rgb(0,127,0)} -span#textcolor1732{color:rgb(0,0,255)} -span#textcolor1733{color:rgb(0,0,255)} -span#textcolor1734{color:rgb(0,127,0)} -span#textcolor1735{color:rgb(0,127,0)} -span#textcolor1736{color:rgb(0,127,0)} -span#textcolor1737{color:rgb(0,127,0)} +span#textcolor1732{color:rgb(0,127,0)} +span#textcolor1733{color:rgb(0,127,0)} +span#textcolor1734{color:rgb(0,0,255)} +span#textcolor1735{color:rgb(43,145,175)} +span#textcolor1736{color:rgb(0,0,255)} +span#textcolor1737{color:rgb(0,0,255)} span#textcolor1738{color:rgb(0,127,0)} span#textcolor1739{color:rgb(0,127,0)} span#textcolor1740{color:rgb(0,127,0)} span#textcolor1741{color:rgb(0,127,0)} -span#textcolor1742{color:rgb(0,0,255)} -span#textcolor1743{color:rgb(43,145,175)} -span#textcolor1744{color:rgb(0,127,0)} -span#textcolor1745{color:rgb(0,127,0)} +span#textcolor1742{color:rgb(0,127,0)} +span#textcolor1743{color:rgb(0,127,0)} +span#textcolor1744{color:rgb(0,0,255)} +span#textcolor1745{color:rgb(0,0,255)} span#textcolor1746{color:rgb(0,127,0)} span#textcolor1747{color:rgb(0,127,0)} span#textcolor1748{color:rgb(0,127,0)} @@ -2213,8 +2213,8 @@ span#textcolor1750{color:rgb(0,127,0)} span#textcolor1751{color:rgb(0,127,0)} span#textcolor1752{color:rgb(0,127,0)} span#textcolor1753{color:rgb(0,127,0)} -span#textcolor1754{color:rgb(0,127,0)} -span#textcolor1755{color:rgb(0,127,0)} +span#textcolor1754{color:rgb(0,0,255)} +span#textcolor1755{color:rgb(43,145,175)} span#textcolor1756{color:rgb(0,127,0)} span#textcolor1757{color:rgb(0,127,0)} span#textcolor1758{color:rgb(0,127,0)} @@ -2230,8 +2230,8 @@ span#textcolor1767{color:rgb(0,127,0)} span#textcolor1768{color:rgb(0,127,0)} span#textcolor1769{color:rgb(0,127,0)} span#textcolor1770{color:rgb(0,127,0)} -span#textcolor1771{color:rgb(0,0,255)} -span#textcolor1772{color:rgb(0,0,255)} +span#textcolor1771{color:rgb(0,127,0)} +span#textcolor1772{color:rgb(0,127,0)} span#textcolor1773{color:rgb(0,127,0)} span#textcolor1774{color:rgb(0,127,0)} span#textcolor1775{color:rgb(0,127,0)} @@ -2243,32 +2243,32 @@ span#textcolor1780{color:rgb(0,127,0)} span#textcolor1781{color:rgb(0,127,0)} span#textcolor1782{color:rgb(0,127,0)} span#textcolor1783{color:rgb(0,0,255)} -span#textcolor1784{color:rgb(0,127,0)} +span#textcolor1784{color:rgb(0,0,255)} span#textcolor1785{color:rgb(0,127,0)} span#textcolor1786{color:rgb(0,127,0)} span#textcolor1787{color:rgb(0,127,0)} span#textcolor1788{color:rgb(0,127,0)} span#textcolor1789{color:rgb(0,127,0)} -span#textcolor1790{color:rgb(0,0,255)} +span#textcolor1790{color:rgb(0,127,0)} span#textcolor1791{color:rgb(0,127,0)} span#textcolor1792{color:rgb(0,127,0)} span#textcolor1793{color:rgb(0,127,0)} span#textcolor1794{color:rgb(0,127,0)} -span#textcolor1795{color:rgb(43,145,175)} -span#textcolor1796{color:rgb(0,0,255)} -span#textcolor1797{color:rgb(0,0,255)} +span#textcolor1795{color:rgb(0,0,255)} +span#textcolor1796{color:rgb(0,127,0)} +span#textcolor1797{color:rgb(0,127,0)} span#textcolor1798{color:rgb(0,127,0)} span#textcolor1799{color:rgb(0,127,0)} span#textcolor1800{color:rgb(0,127,0)} span#textcolor1801{color:rgb(0,127,0)} -span#textcolor1802{color:rgb(0,127,0)} +span#textcolor1802{color:rgb(0,0,255)} span#textcolor1803{color:rgb(0,127,0)} span#textcolor1804{color:rgb(0,127,0)} span#textcolor1805{color:rgb(0,127,0)} span#textcolor1806{color:rgb(0,127,0)} -span#textcolor1807{color:rgb(0,127,0)} +span#textcolor1807{color:rgb(43,145,175)} span#textcolor1808{color:rgb(0,0,255)} -span#textcolor1809{color:rgb(0,127,0)} +span#textcolor1809{color:rgb(0,0,255)} span#textcolor1810{color:rgb(0,127,0)} span#textcolor1811{color:rgb(0,127,0)} span#textcolor1812{color:rgb(0,127,0)} @@ -2278,295 +2278,295 @@ span#textcolor1815{color:rgb(0,127,0)} span#textcolor1816{color:rgb(0,127,0)} span#textcolor1817{color:rgb(0,127,0)} span#textcolor1818{color:rgb(0,127,0)} -span#textcolor1819{color:rgb(0,0,255)} +span#textcolor1819{color:rgb(0,127,0)} span#textcolor1820{color:rgb(0,0,255)} -span#textcolor1821{color:rgb(0,0,255)} -span#textcolor1822{color:rgb(0,0,255)} +span#textcolor1821{color:rgb(0,127,0)} +span#textcolor1822{color:rgb(0,127,0)} span#textcolor1823{color:rgb(0,127,0)} span#textcolor1824{color:rgb(0,127,0)} span#textcolor1825{color:rgb(0,127,0)} span#textcolor1826{color:rgb(0,127,0)} -span#textcolor1827{color:rgb(0,0,255)} -span#textcolor1828{color:rgb(0,0,255)} -span#textcolor1829{color:rgb(0,0,255)} -span#textcolor1830{color:rgb(0,0,255)} +span#textcolor1827{color:rgb(0,127,0)} +span#textcolor1828{color:rgb(0,127,0)} +span#textcolor1829{color:rgb(0,127,0)} +span#textcolor1830{color:rgb(0,127,0)} span#textcolor1831{color:rgb(0,0,255)} -span#textcolor1832{color:rgb(0,127,0)} -span#textcolor1833{color:rgb(0,127,0)} -span#textcolor1834{color:rgb(0,127,0)} +span#textcolor1832{color:rgb(0,0,255)} +span#textcolor1833{color:rgb(0,0,255)} +span#textcolor1834{color:rgb(0,0,255)} span#textcolor1835{color:rgb(0,127,0)} span#textcolor1836{color:rgb(0,127,0)} span#textcolor1837{color:rgb(0,127,0)} -span#textcolor1838{color:rgb(43,145,175)} +span#textcolor1838{color:rgb(0,127,0)} span#textcolor1839{color:rgb(0,0,255)} -span#textcolor1840{color:rgb(163,20,20)} -span#textcolor1841{color:rgb(163,20,20)} -span#textcolor1842{color:rgb(163,20,20)} +span#textcolor1840{color:rgb(0,0,255)} +span#textcolor1841{color:rgb(0,0,255)} +span#textcolor1842{color:rgb(0,0,255)} span#textcolor1843{color:rgb(0,0,255)} -span#textcolor1844{color:rgb(163,20,20)} -span#textcolor1845{color:rgb(163,20,20)} -span#textcolor1846{color:rgb(163,20,20)} -span#textcolor1847{color:rgb(0,0,255)} +span#textcolor1844{color:rgb(0,127,0)} +span#textcolor1845{color:rgb(0,127,0)} +span#textcolor1846{color:rgb(0,127,0)} +span#textcolor1847{color:rgb(0,127,0)} span#textcolor1848{color:rgb(0,127,0)} span#textcolor1849{color:rgb(0,127,0)} -span#textcolor1850{color:rgb(0,127,0)} -span#textcolor1851{color:rgb(0,127,0)} -span#textcolor1852{color:rgb(0,127,0)} -span#textcolor1853{color:rgb(0,127,0)} -span#textcolor1854{color:rgb(43,145,175)} -span#textcolor1855{color:rgb(163,20,20)} +span#textcolor1850{color:rgb(43,145,175)} +span#textcolor1851{color:rgb(0,0,255)} +span#textcolor1852{color:rgb(163,20,20)} +span#textcolor1853{color:rgb(163,20,20)} +span#textcolor1854{color:rgb(163,20,20)} +span#textcolor1855{color:rgb(0,0,255)} span#textcolor1856{color:rgb(163,20,20)} span#textcolor1857{color:rgb(163,20,20)} span#textcolor1858{color:rgb(163,20,20)} +span#textcolor1859{color:rgb(0,0,255)} +span#textcolor1860{color:rgb(0,127,0)} +span#textcolor1861{color:rgb(0,127,0)} +span#textcolor1862{color:rgb(0,127,0)} +span#textcolor1863{color:rgb(0,127,0)} +span#textcolor1864{color:rgb(0,127,0)} +span#textcolor1865{color:rgb(0,127,0)} +span#textcolor1866{color:rgb(43,145,175)} +span#textcolor1867{color:rgb(163,20,20)} +span#textcolor1868{color:rgb(163,20,20)} +span#textcolor1869{color:rgb(163,20,20)} +span#textcolor1870{color:rgb(163,20,20)} pre#fancyvrb54{padding:5.69054pt;} pre#fancyvrb54{ border-top: solid 0.4pt; } pre#fancyvrb54{ border-left: solid 0.4pt; } pre#fancyvrb54{ border-bottom: solid 0.4pt; } pre#fancyvrb54{ border-right: solid 0.4pt; } -span#textcolor1859{color:rgb(0,127,0)} -span#textcolor1860{color:rgb(0,127,0)} -span#textcolor1861{color:rgb(0,127,0)} -span#textcolor1862{color:rgb(0,127,0)} -span#textcolor1863{color:rgb(0,0,255)} -span#textcolor1864{color:rgb(0,127,0)} -span#textcolor1865{color:rgb(0,0,255)} -span#textcolor1866{color:rgb(0,127,0)} -span#textcolor1867{color:rgb(0,0,255)} -span#textcolor1868{color:rgb(0,127,0)} -span#textcolor1869{color:rgb(0,0,255)} -span#textcolor1870{color:rgb(0,127,0)} -span#textcolor1871{color:rgb(0,0,255)} +span#textcolor1871{color:rgb(0,127,0)} span#textcolor1872{color:rgb(0,127,0)} -span#textcolor1873{color:rgb(0,0,255)} -span#textcolor1874{color:rgb(43,145,175)} -span#textcolor1875{color:rgb(43,145,175)} -span#textcolor1876{color:rgb(43,145,175)} -span#textcolor1877{color:rgb(43,145,175)} +span#textcolor1873{color:rgb(0,127,0)} +span#textcolor1874{color:rgb(0,127,0)} +span#textcolor1875{color:rgb(0,0,255)} +span#textcolor1876{color:rgb(0,127,0)} +span#textcolor1877{color:rgb(0,0,255)} span#textcolor1878{color:rgb(0,127,0)} -span#textcolor1879{color:rgb(43,145,175)} +span#textcolor1879{color:rgb(0,0,255)} span#textcolor1880{color:rgb(0,127,0)} -span#textcolor1881{color:rgb(43,145,175)} +span#textcolor1881{color:rgb(0,0,255)} span#textcolor1882{color:rgb(0,127,0)} -span#textcolor1883{color:rgb(0,127,0)} -span#textcolor1884{color:rgb(0,0,255)} -span#textcolor1885{color:rgb(163,20,20)} -span#textcolor1886{color:rgb(163,20,20)} -span#textcolor1887{color:rgb(163,20,20)} -span#textcolor1888{color:rgb(163,20,20)} -span#textcolor1889{color:rgb(0,127,0)} +span#textcolor1883{color:rgb(0,0,255)} +span#textcolor1884{color:rgb(0,127,0)} +span#textcolor1885{color:rgb(0,0,255)} +span#textcolor1886{color:rgb(43,145,175)} +span#textcolor1887{color:rgb(43,145,175)} +span#textcolor1888{color:rgb(43,145,175)} +span#textcolor1889{color:rgb(43,145,175)} span#textcolor1890{color:rgb(0,127,0)} -span#textcolor1891{color:rgb(0,0,255)} -span#textcolor1892{color:rgb(163,20,20)} -span#textcolor1893{color:rgb(163,20,20)} +span#textcolor1891{color:rgb(43,145,175)} +span#textcolor1892{color:rgb(0,127,0)} +span#textcolor1893{color:rgb(43,145,175)} span#textcolor1894{color:rgb(0,127,0)} -span#textcolor1895{color:rgb(0,0,255)} -span#textcolor1896{color:rgb(0,127,0)} -span#textcolor1897{color:rgb(0,127,0)} -span#textcolor1898{color:rgb(0,0,255)} -span#textcolor1899{color:rgb(0,0,255)} +span#textcolor1895{color:rgb(0,127,0)} +span#textcolor1896{color:rgb(0,0,255)} +span#textcolor1897{color:rgb(163,20,20)} +span#textcolor1898{color:rgb(163,20,20)} +span#textcolor1899{color:rgb(163,20,20)} span#textcolor1900{color:rgb(163,20,20)} -span#textcolor1901{color:rgb(0,0,255)} -span#textcolor1902{color:rgb(163,20,20)} -span#textcolor1903{color:rgb(0,127,0)} -span#textcolor1904{color:rgb(0,0,255)} -span#textcolor1905{color:rgb(0,0,255)} -span#textcolor1906{color:rgb(43,145,175)} -span#textcolor1907{color:rgb(0,127,0)} -span#textcolor1908{color:rgb(0,0,255)} -span#textcolor1909{color:rgb(0,0,255)} +span#textcolor1901{color:rgb(0,127,0)} +span#textcolor1902{color:rgb(0,127,0)} +span#textcolor1903{color:rgb(0,0,255)} +span#textcolor1904{color:rgb(163,20,20)} +span#textcolor1905{color:rgb(163,20,20)} +span#textcolor1906{color:rgb(0,127,0)} +span#textcolor1907{color:rgb(0,0,255)} +span#textcolor1908{color:rgb(0,127,0)} +span#textcolor1909{color:rgb(0,127,0)} +span#textcolor1910{color:rgb(0,0,255)} +span#textcolor1911{color:rgb(0,0,255)} +span#textcolor1912{color:rgb(163,20,20)} +span#textcolor1913{color:rgb(0,0,255)} +span#textcolor1914{color:rgb(163,20,20)} +span#textcolor1915{color:rgb(0,127,0)} +span#textcolor1916{color:rgb(0,0,255)} +span#textcolor1917{color:rgb(0,0,255)} +span#textcolor1918{color:rgb(43,145,175)} +span#textcolor1919{color:rgb(0,127,0)} +span#textcolor1920{color:rgb(0,0,255)} +span#textcolor1921{color:rgb(0,0,255)} pre#fancyvrb55{padding:5.69054pt;} pre#fancyvrb55{ border-top: solid 0.4pt; } pre#fancyvrb55{ border-left: solid 0.4pt; } pre#fancyvrb55{ border-bottom: solid 0.4pt; } pre#fancyvrb55{ border-right: solid 0.4pt; } -span#textcolor1910{color:rgb(0,127,0)} -span#textcolor1911{color:rgb(0,127,0)} -span#textcolor1912{color:rgb(0,127,0)} -span#textcolor1913{color:rgb(0,0,255)} -span#textcolor1914{color:rgb(0,127,0)} -span#textcolor1915{color:rgb(0,0,255)} -span#textcolor1916{color:rgb(0,127,0)} -span#textcolor1917{color:rgb(0,0,255)} -span#textcolor1918{color:rgb(0,127,0)} -span#textcolor1919{color:rgb(0,0,255)} -span#textcolor1920{color:rgb(0,127,0)} -span#textcolor1921{color:rgb(0,0,255)} span#textcolor1922{color:rgb(0,127,0)} -span#textcolor1923{color:rgb(0,0,255)} -span#textcolor1924{color:rgb(0,0,255)} +span#textcolor1923{color:rgb(0,127,0)} +span#textcolor1924{color:rgb(0,127,0)} span#textcolor1925{color:rgb(0,0,255)} -span#textcolor1926{color:rgb(0,0,255)} +span#textcolor1926{color:rgb(0,127,0)} span#textcolor1927{color:rgb(0,0,255)} -span#textcolor1928{color:rgb(43,145,175)} -span#textcolor1929{color:rgb(43,145,175)} -span#textcolor1930{color:rgb(163,20,20)} -span#textcolor1931{color:rgb(163,20,20)} -span#textcolor1932{color:rgb(163,20,20)} +span#textcolor1928{color:rgb(0,127,0)} +span#textcolor1929{color:rgb(0,0,255)} +span#textcolor1930{color:rgb(0,127,0)} +span#textcolor1931{color:rgb(0,0,255)} +span#textcolor1932{color:rgb(0,127,0)} span#textcolor1933{color:rgb(0,0,255)} -span#textcolor1934{color:rgb(43,145,175)} -span#textcolor1935{color:rgb(43,145,175)} -span#textcolor1936{color:rgb(163,20,20)} -span#textcolor1937{color:rgb(163,20,20)} -span#textcolor1938{color:rgb(163,20,20)} +span#textcolor1934{color:rgb(0,127,0)} +span#textcolor1935{color:rgb(0,0,255)} +span#textcolor1936{color:rgb(0,0,255)} +span#textcolor1937{color:rgb(0,0,255)} +span#textcolor1938{color:rgb(0,0,255)} span#textcolor1939{color:rgb(0,0,255)} span#textcolor1940{color:rgb(43,145,175)} span#textcolor1941{color:rgb(43,145,175)} -span#textcolor1942{color:rgb(0,0,255)} -span#textcolor1943{color:rgb(0,0,255)} +span#textcolor1942{color:rgb(163,20,20)} +span#textcolor1943{color:rgb(163,20,20)} span#textcolor1944{color:rgb(163,20,20)} -span#textcolor1945{color:rgb(163,20,20)} -span#textcolor1946{color:rgb(163,20,20)} -span#textcolor1947{color:rgb(163,20,20)} -span#textcolor1948{color:rgb(0,0,255)} -span#textcolor1949{color:rgb(0,0,255)} +span#textcolor1945{color:rgb(0,0,255)} +span#textcolor1946{color:rgb(43,145,175)} +span#textcolor1947{color:rgb(43,145,175)} +span#textcolor1948{color:rgb(163,20,20)} +span#textcolor1949{color:rgb(163,20,20)} span#textcolor1950{color:rgb(163,20,20)} span#textcolor1951{color:rgb(0,0,255)} -span#textcolor1952{color:rgb(0,0,255)} -span#textcolor1953{color:rgb(0,0,255)} +span#textcolor1952{color:rgb(43,145,175)} +span#textcolor1953{color:rgb(43,145,175)} span#textcolor1954{color:rgb(0,0,255)} -span#textcolor1955{color:rgb(43,145,175)} -span#textcolor1956{color:rgb(43,145,175)} +span#textcolor1955{color:rgb(0,0,255)} +span#textcolor1956{color:rgb(163,20,20)} span#textcolor1957{color:rgb(163,20,20)} span#textcolor1958{color:rgb(163,20,20)} span#textcolor1959{color:rgb(163,20,20)} -span#textcolor1960{color:rgb(163,20,20)} -span#textcolor1961{color:rgb(163,20,20)} +span#textcolor1960{color:rgb(0,0,255)} +span#textcolor1961{color:rgb(0,0,255)} +span#textcolor1962{color:rgb(163,20,20)} +span#textcolor1963{color:rgb(0,0,255)} +span#textcolor1964{color:rgb(0,0,255)} +span#textcolor1965{color:rgb(0,0,255)} +span#textcolor1966{color:rgb(0,0,255)} +span#textcolor1967{color:rgb(43,145,175)} +span#textcolor1968{color:rgb(43,145,175)} +span#textcolor1969{color:rgb(163,20,20)} +span#textcolor1970{color:rgb(163,20,20)} +span#textcolor1971{color:rgb(163,20,20)} +span#textcolor1972{color:rgb(163,20,20)} +span#textcolor1973{color:rgb(163,20,20)} pre#fancyvrb56{padding:5.69054pt;} pre#fancyvrb56{ border-top: solid 0.4pt; } pre#fancyvrb56{ border-left: solid 0.4pt; } pre#fancyvrb56{ border-bottom: solid 0.4pt; } pre#fancyvrb56{ border-right: solid 0.4pt; } -span#textcolor1962{color:rgb(0,127,0)} -span#textcolor1963{color:rgb(0,127,0)} -span#textcolor1964{color:rgb(0,127,0)} -span#textcolor1965{color:rgb(0,0,255)} -span#textcolor1966{color:rgb(0,127,0)} -span#textcolor1967{color:rgb(0,0,255)} -span#textcolor1968{color:rgb(0,127,0)} -span#textcolor1969{color:rgb(0,0,255)} -span#textcolor1970{color:rgb(0,127,0)} -span#textcolor1971{color:rgb(0,0,255)} -span#textcolor1972{color:rgb(0,127,0)} -span#textcolor1973{color:rgb(0,0,255)} -span#textcolor1974{color:rgb(43,145,175)} -span#textcolor1975{color:rgb(43,145,175)} -span#textcolor1976{color:rgb(43,145,175)} -span#textcolor1977{color:rgb(163,20,20)} -span#textcolor1978{color:rgb(163,20,20)} -span#textcolor1979{color:rgb(163,20,20)} -span#textcolor1980{color:rgb(0,0,255)} -span#textcolor1981{color:rgb(163,20,20)} -span#textcolor1982{color:rgb(163,20,20)} -span#textcolor1983{color:rgb(163,20,20)} -span#textcolor1984{color:rgb(0,0,255)} -span#textcolor1985{color:rgb(163,20,20)} -span#textcolor1986{color:rgb(163,20,20)} -span#textcolor1987{color:rgb(163,20,20)} -span#textcolor1988{color:rgb(163,20,20)} +span#textcolor1974{color:rgb(0,127,0)} +span#textcolor1975{color:rgb(0,127,0)} +span#textcolor1976{color:rgb(0,127,0)} +span#textcolor1977{color:rgb(0,0,255)} +span#textcolor1978{color:rgb(0,127,0)} +span#textcolor1979{color:rgb(0,0,255)} +span#textcolor1980{color:rgb(0,127,0)} +span#textcolor1981{color:rgb(0,0,255)} +span#textcolor1982{color:rgb(0,127,0)} +span#textcolor1983{color:rgb(0,0,255)} +span#textcolor1984{color:rgb(0,127,0)} +span#textcolor1985{color:rgb(0,0,255)} +span#textcolor1986{color:rgb(43,145,175)} +span#textcolor1987{color:rgb(43,145,175)} +span#textcolor1988{color:rgb(43,145,175)} span#textcolor1989{color:rgb(163,20,20)} span#textcolor1990{color:rgb(163,20,20)} -span#textcolor1991{color:rgb(0,0,255)} -span#textcolor1992{color:rgb(163,20,20)} +span#textcolor1991{color:rgb(163,20,20)} +span#textcolor1992{color:rgb(0,0,255)} span#textcolor1993{color:rgb(163,20,20)} span#textcolor1994{color:rgb(163,20,20)} -span#textcolor1995{color:rgb(0,0,255)} +span#textcolor1995{color:rgb(163,20,20)} span#textcolor1996{color:rgb(0,0,255)} -span#textcolor1997{color:rgb(43,145,175)} -span#textcolor1998{color:rgb(43,145,175)} +span#textcolor1997{color:rgb(163,20,20)} +span#textcolor1998{color:rgb(163,20,20)} span#textcolor1999{color:rgb(163,20,20)} span#textcolor2000{color:rgb(163,20,20)} span#textcolor2001{color:rgb(163,20,20)} span#textcolor2002{color:rgb(163,20,20)} -span#textcolor2003{color:rgb(163,20,20)} +span#textcolor2003{color:rgb(0,0,255)} +span#textcolor2004{color:rgb(163,20,20)} +span#textcolor2005{color:rgb(163,20,20)} +span#textcolor2006{color:rgb(163,20,20)} +span#textcolor2007{color:rgb(0,0,255)} +span#textcolor2008{color:rgb(0,0,255)} +span#textcolor2009{color:rgb(43,145,175)} +span#textcolor2010{color:rgb(43,145,175)} +span#textcolor2011{color:rgb(163,20,20)} +span#textcolor2012{color:rgb(163,20,20)} +span#textcolor2013{color:rgb(163,20,20)} +span#textcolor2014{color:rgb(163,20,20)} +span#textcolor2015{color:rgb(163,20,20)} pre#fancyvrb57{padding:5.69054pt;} pre#fancyvrb57{ border-top: solid 0.4pt; } pre#fancyvrb57{ border-left: solid 0.4pt; } pre#fancyvrb57{ border-bottom: solid 0.4pt; } pre#fancyvrb57{ border-right: solid 0.4pt; } -span#textcolor2004{color:rgb(0,127,0)} -span#textcolor2005{color:rgb(0,127,0)} -span#textcolor2006{color:rgb(0,127,0)} -span#textcolor2007{color:rgb(0,0,255)} -span#textcolor2008{color:rgb(0,127,0)} -span#textcolor2009{color:rgb(0,0,255)} -span#textcolor2010{color:rgb(0,127,0)} -span#textcolor2011{color:rgb(0,0,255)} -span#textcolor2012{color:rgb(0,127,0)} -span#textcolor2013{color:rgb(0,0,255)} -span#textcolor2014{color:rgb(0,127,0)} -span#textcolor2015{color:rgb(0,0,255)} span#textcolor2016{color:rgb(0,127,0)} -span#textcolor2017{color:rgb(0,0,255)} -span#textcolor2018{color:rgb(43,145,175)} -span#textcolor2019{color:rgb(43,145,175)} -span#textcolor2020{color:rgb(43,145,175)} -span#textcolor2021{color:rgb(43,145,175)} -span#textcolor2022{color:rgb(163,20,20)} -span#textcolor2023{color:rgb(163,20,20)} -span#textcolor2024{color:rgb(163,20,20)} -span#textcolor2025{color:rgb(0,127,0)} +span#textcolor2017{color:rgb(0,127,0)} +span#textcolor2018{color:rgb(0,127,0)} +span#textcolor2019{color:rgb(0,0,255)} +span#textcolor2020{color:rgb(0,127,0)} +span#textcolor2021{color:rgb(0,0,255)} +span#textcolor2022{color:rgb(0,127,0)} +span#textcolor2023{color:rgb(0,0,255)} +span#textcolor2024{color:rgb(0,127,0)} +span#textcolor2025{color:rgb(0,0,255)} span#textcolor2026{color:rgb(0,127,0)} -span#textcolor2027{color:rgb(0,127,0)} +span#textcolor2027{color:rgb(0,0,255)} span#textcolor2028{color:rgb(0,127,0)} -span#textcolor2029{color:rgb(163,20,20)} -span#textcolor2030{color:rgb(163,20,20)} -span#textcolor2031{color:rgb(163,20,20)} -span#textcolor2032{color:rgb(0,0,255)} +span#textcolor2029{color:rgb(0,0,255)} +span#textcolor2030{color:rgb(43,145,175)} +span#textcolor2031{color:rgb(43,145,175)} +span#textcolor2032{color:rgb(43,145,175)} span#textcolor2033{color:rgb(43,145,175)} -span#textcolor2034{color:rgb(43,145,175)} -span#textcolor2035{color:rgb(43,145,175)} -span#textcolor2036{color:rgb(43,145,175)} -span#textcolor2037{color:rgb(163,20,20)} -span#textcolor2038{color:rgb(163,20,20)} -span#textcolor2039{color:rgb(163,20,20)} +span#textcolor2034{color:rgb(163,20,20)} +span#textcolor2035{color:rgb(163,20,20)} +span#textcolor2036{color:rgb(163,20,20)} +span#textcolor2037{color:rgb(0,127,0)} +span#textcolor2038{color:rgb(0,127,0)} +span#textcolor2039{color:rgb(0,127,0)} span#textcolor2040{color:rgb(0,127,0)} -span#textcolor2041{color:rgb(0,127,0)} -span#textcolor2042{color:rgb(0,127,0)} -span#textcolor2043{color:rgb(0,127,0)} -span#textcolor2044{color:rgb(163,20,20)} -span#textcolor2045{color:rgb(163,20,20)} -span#textcolor2046{color:rgb(163,20,20)} -span#textcolor2047{color:rgb(0,0,255)} +span#textcolor2041{color:rgb(163,20,20)} +span#textcolor2042{color:rgb(163,20,20)} +span#textcolor2043{color:rgb(163,20,20)} +span#textcolor2044{color:rgb(0,0,255)} +span#textcolor2045{color:rgb(43,145,175)} +span#textcolor2046{color:rgb(43,145,175)} +span#textcolor2047{color:rgb(43,145,175)} span#textcolor2048{color:rgb(43,145,175)} -span#textcolor2049{color:rgb(43,145,175)} +span#textcolor2049{color:rgb(163,20,20)} span#textcolor2050{color:rgb(163,20,20)} span#textcolor2051{color:rgb(163,20,20)} -span#textcolor2052{color:rgb(163,20,20)} -span#textcolor2053{color:rgb(0,0,255)} -span#textcolor2054{color:rgb(0,0,255)} -span#textcolor2055{color:rgb(43,145,175)} -span#textcolor2056{color:rgb(43,145,175)} +span#textcolor2052{color:rgb(0,127,0)} +span#textcolor2053{color:rgb(0,127,0)} +span#textcolor2054{color:rgb(0,127,0)} +span#textcolor2055{color:rgb(0,127,0)} +span#textcolor2056{color:rgb(163,20,20)} span#textcolor2057{color:rgb(163,20,20)} span#textcolor2058{color:rgb(163,20,20)} -span#textcolor2059{color:rgb(163,20,20)} -span#textcolor2060{color:rgb(163,20,20)} -span#textcolor2061{color:rgb(163,20,20)} +span#textcolor2059{color:rgb(0,0,255)} +span#textcolor2060{color:rgb(43,145,175)} +span#textcolor2061{color:rgb(43,145,175)} +span#textcolor2062{color:rgb(163,20,20)} +span#textcolor2063{color:rgb(163,20,20)} +span#textcolor2064{color:rgb(163,20,20)} +span#textcolor2065{color:rgb(0,0,255)} +span#textcolor2066{color:rgb(0,0,255)} +span#textcolor2067{color:rgb(43,145,175)} +span#textcolor2068{color:rgb(43,145,175)} +span#textcolor2069{color:rgb(163,20,20)} +span#textcolor2070{color:rgb(163,20,20)} +span#textcolor2071{color:rgb(163,20,20)} +span#textcolor2072{color:rgb(163,20,20)} +span#textcolor2073{color:rgb(163,20,20)} pre#fancyvrb58{padding:5.69054pt;} pre#fancyvrb58{ border-top: solid 0.4pt; } pre#fancyvrb58{ border-left: solid 0.4pt; } pre#fancyvrb58{ border-bottom: solid 0.4pt; } pre#fancyvrb58{ border-right: solid 0.4pt; } -span#textcolor2062{color:rgb(0,127,0)} -span#textcolor2063{color:rgb(0,127,0)} -span#textcolor2064{color:rgb(0,127,0)} -span#textcolor2065{color:rgb(0,0,255)} -span#textcolor2066{color:rgb(0,127,0)} -span#textcolor2067{color:rgb(0,0,255)} -span#textcolor2068{color:rgb(0,127,0)} -span#textcolor2069{color:rgb(0,0,255)} -span#textcolor2070{color:rgb(0,127,0)} -span#textcolor2071{color:rgb(0,0,255)} -span#textcolor2072{color:rgb(43,145,175)} -span#textcolor2073{color:rgb(43,145,175)} -span#textcolor2074{color:rgb(43,145,175)} -span#textcolor2075{color:rgb(43,145,175)} -span#textcolor2076{color:rgb(163,20,20)} -span#textcolor2077{color:rgb(163,20,20)} -span#textcolor2078{color:rgb(163,20,20)} -span#textcolor2079{color:rgb(0,127,0)} -span#textcolor2080{color:rgb(163,20,20)} -span#textcolor2081{color:rgb(163,20,20)} -span#textcolor2082{color:rgb(163,20,20)} +span#textcolor2074{color:rgb(0,127,0)} +span#textcolor2075{color:rgb(0,127,0)} +span#textcolor2076{color:rgb(0,127,0)} +span#textcolor2077{color:rgb(0,0,255)} +span#textcolor2078{color:rgb(0,127,0)} +span#textcolor2079{color:rgb(0,0,255)} +span#textcolor2080{color:rgb(0,127,0)} +span#textcolor2081{color:rgb(0,0,255)} +span#textcolor2082{color:rgb(0,127,0)} span#textcolor2083{color:rgb(0,0,255)} span#textcolor2084{color:rgb(43,145,175)} span#textcolor2085{color:rgb(43,145,175)} @@ -2582,115 +2582,115 @@ span#textcolor2094{color:rgb(163,20,20)} span#textcolor2095{color:rgb(0,0,255)} span#textcolor2096{color:rgb(43,145,175)} span#textcolor2097{color:rgb(43,145,175)} -span#textcolor2098{color:rgb(163,20,20)} -span#textcolor2099{color:rgb(163,20,20)} +span#textcolor2098{color:rgb(43,145,175)} +span#textcolor2099{color:rgb(43,145,175)} span#textcolor2100{color:rgb(163,20,20)} -span#textcolor2101{color:rgb(0,0,255)} -span#textcolor2102{color:rgb(0,0,255)} -span#textcolor2103{color:rgb(43,145,175)} -span#textcolor2104{color:rgb(43,145,175)} +span#textcolor2101{color:rgb(163,20,20)} +span#textcolor2102{color:rgb(163,20,20)} +span#textcolor2103{color:rgb(0,127,0)} +span#textcolor2104{color:rgb(163,20,20)} span#textcolor2105{color:rgb(163,20,20)} span#textcolor2106{color:rgb(163,20,20)} -span#textcolor2107{color:rgb(163,20,20)} -span#textcolor2108{color:rgb(163,20,20)} -span#textcolor2109{color:rgb(163,20,20)} +span#textcolor2107{color:rgb(0,0,255)} +span#textcolor2108{color:rgb(43,145,175)} +span#textcolor2109{color:rgb(43,145,175)} +span#textcolor2110{color:rgb(163,20,20)} +span#textcolor2111{color:rgb(163,20,20)} +span#textcolor2112{color:rgb(163,20,20)} +span#textcolor2113{color:rgb(0,0,255)} +span#textcolor2114{color:rgb(0,0,255)} +span#textcolor2115{color:rgb(43,145,175)} +span#textcolor2116{color:rgb(43,145,175)} +span#textcolor2117{color:rgb(163,20,20)} +span#textcolor2118{color:rgb(163,20,20)} +span#textcolor2119{color:rgb(163,20,20)} +span#textcolor2120{color:rgb(163,20,20)} +span#textcolor2121{color:rgb(163,20,20)} pre#fancyvrb59{padding:5.69054pt;} pre#fancyvrb59{ border-top: solid 0.4pt; } pre#fancyvrb59{ border-left: solid 0.4pt; } pre#fancyvrb59{ border-bottom: solid 0.4pt; } pre#fancyvrb59{ border-right: solid 0.4pt; } -span#textcolor2110{color:rgb(0,127,0)} -span#textcolor2111{color:rgb(0,127,0)} -span#textcolor2112{color:rgb(0,127,0)} -span#textcolor2113{color:rgb(0,0,255)} -span#textcolor2114{color:rgb(0,127,0)} -span#textcolor2115{color:rgb(0,0,255)} -span#textcolor2116{color:rgb(0,127,0)} -span#textcolor2117{color:rgb(0,0,255)} -span#textcolor2118{color:rgb(0,127,0)} -span#textcolor2119{color:rgb(0,0,255)} -span#textcolor2120{color:rgb(0,0,255)} -span#textcolor2121{color:rgb(0,0,255)} -span#textcolor2122{color:rgb(0,0,255)} -span#textcolor2123{color:rgb(0,0,255)} -span#textcolor2124{color:rgb(0,0,255)} +span#textcolor2122{color:rgb(0,127,0)} +span#textcolor2123{color:rgb(0,127,0)} +span#textcolor2124{color:rgb(0,127,0)} span#textcolor2125{color:rgb(0,0,255)} -span#textcolor2126{color:rgb(43,145,175)} -span#textcolor2127{color:rgb(43,145,175)} +span#textcolor2126{color:rgb(0,127,0)} +span#textcolor2127{color:rgb(0,0,255)} span#textcolor2128{color:rgb(0,127,0)} -span#textcolor2129{color:rgb(0,127,0)} -span#textcolor2130{color:rgb(163,20,20)} -span#textcolor2131{color:rgb(163,20,20)} -span#textcolor2132{color:rgb(163,20,20)} +span#textcolor2129{color:rgb(0,0,255)} +span#textcolor2130{color:rgb(0,127,0)} +span#textcolor2131{color:rgb(0,0,255)} +span#textcolor2132{color:rgb(0,0,255)} span#textcolor2133{color:rgb(0,0,255)} -span#textcolor2134{color:rgb(43,145,175)} -span#textcolor2135{color:rgb(43,145,175)} -span#textcolor2136{color:rgb(43,145,175)} -span#textcolor2137{color:rgb(43,145,175)} -span#textcolor2138{color:rgb(163,20,20)} -span#textcolor2139{color:rgb(163,20,20)} -span#textcolor2140{color:rgb(163,20,20)} -span#textcolor2141{color:rgb(163,20,20)} -span#textcolor2142{color:rgb(0,0,255)} +span#textcolor2134{color:rgb(0,0,255)} +span#textcolor2135{color:rgb(0,0,255)} +span#textcolor2136{color:rgb(0,0,255)} +span#textcolor2137{color:rgb(0,0,255)} +span#textcolor2138{color:rgb(43,145,175)} +span#textcolor2139{color:rgb(43,145,175)} +span#textcolor2140{color:rgb(0,127,0)} +span#textcolor2141{color:rgb(0,127,0)} +span#textcolor2142{color:rgb(163,20,20)} span#textcolor2143{color:rgb(163,20,20)} span#textcolor2144{color:rgb(163,20,20)} -span#textcolor2145{color:rgb(163,20,20)} -span#textcolor2146{color:rgb(163,20,20)} -span#textcolor2147{color:rgb(163,20,20)} -span#textcolor2148{color:rgb(0,0,255)} +span#textcolor2145{color:rgb(0,0,255)} +span#textcolor2146{color:rgb(43,145,175)} +span#textcolor2147{color:rgb(43,145,175)} +span#textcolor2148{color:rgb(43,145,175)} span#textcolor2149{color:rgb(43,145,175)} -span#textcolor2150{color:rgb(43,145,175)} +span#textcolor2150{color:rgb(163,20,20)} span#textcolor2151{color:rgb(163,20,20)} span#textcolor2152{color:rgb(163,20,20)} span#textcolor2153{color:rgb(163,20,20)} span#textcolor2154{color:rgb(0,0,255)} -span#textcolor2155{color:rgb(0,0,255)} -span#textcolor2156{color:rgb(43,145,175)} -span#textcolor2157{color:rgb(43,145,175)} +span#textcolor2155{color:rgb(163,20,20)} +span#textcolor2156{color:rgb(163,20,20)} +span#textcolor2157{color:rgb(163,20,20)} span#textcolor2158{color:rgb(163,20,20)} span#textcolor2159{color:rgb(163,20,20)} -span#textcolor2160{color:rgb(163,20,20)} -span#textcolor2161{color:rgb(163,20,20)} -span#textcolor2162{color:rgb(163,20,20)} +span#textcolor2160{color:rgb(0,0,255)} +span#textcolor2161{color:rgb(43,145,175)} +span#textcolor2162{color:rgb(43,145,175)} +span#textcolor2163{color:rgb(163,20,20)} +span#textcolor2164{color:rgb(163,20,20)} +span#textcolor2165{color:rgb(163,20,20)} +span#textcolor2166{color:rgb(0,0,255)} +span#textcolor2167{color:rgb(0,0,255)} +span#textcolor2168{color:rgb(43,145,175)} +span#textcolor2169{color:rgb(43,145,175)} +span#textcolor2170{color:rgb(163,20,20)} +span#textcolor2171{color:rgb(163,20,20)} +span#textcolor2172{color:rgb(163,20,20)} +span#textcolor2173{color:rgb(163,20,20)} +span#textcolor2174{color:rgb(163,20,20)} pre#fancyvrb60{padding:5.69054pt;} pre#fancyvrb60{ border-top: solid 0.4pt; } pre#fancyvrb60{ border-left: solid 0.4pt; } pre#fancyvrb60{ border-bottom: solid 0.4pt; } pre#fancyvrb60{ border-right: solid 0.4pt; } -span#textcolor2163{color:rgb(0,127,0)} -span#textcolor2164{color:rgb(0,127,0)} -span#textcolor2165{color:rgb(0,127,0)} -span#textcolor2166{color:rgb(0,127,0)} -span#textcolor2167{color:rgb(0,127,0)} -span#textcolor2168{color:rgb(0,0,255)} -span#textcolor2169{color:rgb(0,127,0)} -span#textcolor2170{color:rgb(0,0,255)} -span#textcolor2171{color:rgb(0,127,0)} -span#textcolor2172{color:rgb(0,0,255)} -span#textcolor2173{color:rgb(0,127,0)} -span#textcolor2174{color:rgb(0,0,255)} span#textcolor2175{color:rgb(0,127,0)} -span#textcolor2176{color:rgb(0,0,255)} +span#textcolor2176{color:rgb(0,127,0)} span#textcolor2177{color:rgb(0,127,0)} -span#textcolor2178{color:rgb(163,20,20)} -span#textcolor2179{color:rgb(0,0,255)} -span#textcolor2180{color:rgb(43,145,175)} -span#textcolor2181{color:rgb(43,145,175)} +span#textcolor2178{color:rgb(0,127,0)} +span#textcolor2179{color:rgb(0,127,0)} +span#textcolor2180{color:rgb(0,0,255)} +span#textcolor2181{color:rgb(0,127,0)} span#textcolor2182{color:rgb(0,0,255)} -span#textcolor2183{color:rgb(0,0,255)} +span#textcolor2183{color:rgb(0,127,0)} span#textcolor2184{color:rgb(0,0,255)} span#textcolor2185{color:rgb(0,127,0)} -span#textcolor2186{color:rgb(0,127,0)} +span#textcolor2186{color:rgb(0,0,255)} span#textcolor2187{color:rgb(0,127,0)} -span#textcolor2188{color:rgb(0,127,0)} +span#textcolor2188{color:rgb(0,0,255)} span#textcolor2189{color:rgb(0,127,0)} -span#textcolor2190{color:rgb(0,127,0)} -span#textcolor2191{color:rgb(0,127,0)} -span#textcolor2192{color:rgb(0,0,255)} -span#textcolor2193{color:rgb(0,127,0)} -span#textcolor2194{color:rgb(0,127,0)} -span#textcolor2195{color:rgb(0,127,0)} -span#textcolor2196{color:rgb(0,127,0)} +span#textcolor2190{color:rgb(163,20,20)} +span#textcolor2191{color:rgb(0,0,255)} +span#textcolor2192{color:rgb(43,145,175)} +span#textcolor2193{color:rgb(43,145,175)} +span#textcolor2194{color:rgb(0,0,255)} +span#textcolor2195{color:rgb(0,0,255)} +span#textcolor2196{color:rgb(0,0,255)} span#textcolor2197{color:rgb(0,127,0)} span#textcolor2198{color:rgb(0,127,0)} span#textcolor2199{color:rgb(0,127,0)} @@ -2698,7 +2698,7 @@ span#textcolor2200{color:rgb(0,127,0)} span#textcolor2201{color:rgb(0,127,0)} span#textcolor2202{color:rgb(0,127,0)} span#textcolor2203{color:rgb(0,127,0)} -span#textcolor2204{color:rgb(0,127,0)} +span#textcolor2204{color:rgb(0,0,255)} span#textcolor2205{color:rgb(0,127,0)} span#textcolor2206{color:rgb(0,127,0)} span#textcolor2207{color:rgb(0,127,0)} @@ -2723,472 +2723,484 @@ span#textcolor2225{color:rgb(0,127,0)} span#textcolor2226{color:rgb(0,127,0)} span#textcolor2227{color:rgb(0,127,0)} span#textcolor2228{color:rgb(0,127,0)} -span#textcolor2229{color:rgb(163,20,20)} -span#textcolor2230{color:rgb(163,20,20)} -span#textcolor2231{color:rgb(163,20,20)} -span#textcolor2232{color:rgb(0,0,255)} -span#textcolor2233{color:rgb(43,145,175)} -span#textcolor2234{color:rgb(43,145,175)} -span#textcolor2235{color:rgb(163,20,20)} -span#textcolor2236{color:rgb(0,0,255)} -span#textcolor2237{color:rgb(0,0,255)} -span#textcolor2238{color:rgb(43,145,175)} -span#textcolor2239{color:rgb(43,145,175)} -span#textcolor2240{color:rgb(163,20,20)} +span#textcolor2229{color:rgb(0,127,0)} +span#textcolor2230{color:rgb(0,127,0)} +span#textcolor2231{color:rgb(0,127,0)} +span#textcolor2232{color:rgb(0,127,0)} +span#textcolor2233{color:rgb(0,127,0)} +span#textcolor2234{color:rgb(0,127,0)} +span#textcolor2235{color:rgb(0,127,0)} +span#textcolor2236{color:rgb(0,127,0)} +span#textcolor2237{color:rgb(0,127,0)} +span#textcolor2238{color:rgb(0,127,0)} +span#textcolor2239{color:rgb(0,127,0)} +span#textcolor2240{color:rgb(0,127,0)} +span#textcolor2241{color:rgb(163,20,20)} +span#textcolor2242{color:rgb(163,20,20)} +span#textcolor2243{color:rgb(163,20,20)} +span#textcolor2244{color:rgb(0,0,255)} +span#textcolor2245{color:rgb(43,145,175)} +span#textcolor2246{color:rgb(43,145,175)} +span#textcolor2247{color:rgb(163,20,20)} +span#textcolor2248{color:rgb(0,0,255)} +span#textcolor2249{color:rgb(0,0,255)} +span#textcolor2250{color:rgb(43,145,175)} +span#textcolor2251{color:rgb(43,145,175)} +span#textcolor2252{color:rgb(163,20,20)} pre#fancyvrb61{padding:5.69054pt;} pre#fancyvrb61{ border-top: solid 0.4pt; } pre#fancyvrb61{ border-left: solid 0.4pt; } pre#fancyvrb61{ border-bottom: solid 0.4pt; } pre#fancyvrb61{ border-right: solid 0.4pt; } -span#textcolor2241{color:rgb(0,127,0)} -span#textcolor2242{color:rgb(0,127,0)} -span#textcolor2243{color:rgb(0,127,0)} -span#textcolor2244{color:rgb(0,0,255)} -span#textcolor2245{color:rgb(0,127,0)} -span#textcolor2246{color:rgb(0,0,255)} -span#textcolor2247{color:rgb(0,127,0)} -span#textcolor2248{color:rgb(0,0,255)} -span#textcolor2249{color:rgb(0,127,0)} -span#textcolor2250{color:rgb(0,0,255)} -span#textcolor2251{color:rgb(0,127,0)} -span#textcolor2252{color:rgb(0,0,255)} span#textcolor2253{color:rgb(0,127,0)} -span#textcolor2254{color:rgb(0,0,255)} +span#textcolor2254{color:rgb(0,127,0)} span#textcolor2255{color:rgb(0,127,0)} span#textcolor2256{color:rgb(0,0,255)} span#textcolor2257{color:rgb(0,127,0)} -span#textcolor2258{color:rgb(163,20,20)} -span#textcolor2259{color:rgb(163,20,20)} +span#textcolor2258{color:rgb(0,0,255)} +span#textcolor2259{color:rgb(0,127,0)} span#textcolor2260{color:rgb(0,0,255)} -span#textcolor2261{color:rgb(0,0,255)} -span#textcolor2262{color:rgb(43,145,175)} -span#textcolor2263{color:rgb(0,0,255)} +span#textcolor2261{color:rgb(0,127,0)} +span#textcolor2262{color:rgb(0,0,255)} +span#textcolor2263{color:rgb(0,127,0)} span#textcolor2264{color:rgb(0,0,255)} -span#textcolor2265{color:rgb(0,0,255)} -span#textcolor2266{color:rgb(0,127,0)} +span#textcolor2265{color:rgb(0,127,0)} +span#textcolor2266{color:rgb(0,0,255)} span#textcolor2267{color:rgb(0,127,0)} -span#textcolor2268{color:rgb(0,127,0)} +span#textcolor2268{color:rgb(0,0,255)} span#textcolor2269{color:rgb(0,127,0)} -span#textcolor2270{color:rgb(0,127,0)} -span#textcolor2271{color:rgb(0,127,0)} -span#textcolor2272{color:rgb(0,127,0)} -span#textcolor2273{color:rgb(0,127,0)} -span#textcolor2274{color:rgb(0,127,0)} -span#textcolor2275{color:rgb(0,127,0)} -span#textcolor2276{color:rgb(0,127,0)} -span#textcolor2277{color:rgb(0,127,0)} +span#textcolor2270{color:rgb(163,20,20)} +span#textcolor2271{color:rgb(163,20,20)} +span#textcolor2272{color:rgb(0,0,255)} +span#textcolor2273{color:rgb(0,0,255)} +span#textcolor2274{color:rgb(43,145,175)} +span#textcolor2275{color:rgb(0,0,255)} +span#textcolor2276{color:rgb(0,0,255)} +span#textcolor2277{color:rgb(0,0,255)} span#textcolor2278{color:rgb(0,127,0)} span#textcolor2279{color:rgb(0,127,0)} -span#textcolor2280{color:rgb(0,0,255)} -span#textcolor2281{color:rgb(43,145,175)} -span#textcolor2282{color:rgb(43,145,175)} -span#textcolor2283{color:rgb(43,145,175)} -span#textcolor2284{color:rgb(43,145,175)} -span#textcolor2285{color:rgb(43,145,175)} -span#textcolor2286{color:rgb(43,145,175)} -span#textcolor2287{color:rgb(43,145,175)} -span#textcolor2288{color:rgb(0,0,255)} -span#textcolor2289{color:rgb(0,0,255)} -span#textcolor2290{color:rgb(0,0,255)} -span#textcolor2291{color:rgb(0,0,255)} -span#textcolor2292{color:rgb(43,145,175)} +span#textcolor2280{color:rgb(0,127,0)} +span#textcolor2281{color:rgb(0,127,0)} +span#textcolor2282{color:rgb(0,127,0)} +span#textcolor2283{color:rgb(0,127,0)} +span#textcolor2284{color:rgb(0,127,0)} +span#textcolor2285{color:rgb(0,127,0)} +span#textcolor2286{color:rgb(0,127,0)} +span#textcolor2287{color:rgb(0,127,0)} +span#textcolor2288{color:rgb(0,127,0)} +span#textcolor2289{color:rgb(0,127,0)} +span#textcolor2290{color:rgb(0,127,0)} +span#textcolor2291{color:rgb(0,127,0)} +span#textcolor2292{color:rgb(0,0,255)} span#textcolor2293{color:rgb(43,145,175)} span#textcolor2294{color:rgb(43,145,175)} -span#textcolor2295{color:rgb(163,20,20)} -span#textcolor2296{color:rgb(163,20,20)} -span#textcolor2297{color:rgb(163,20,20)} -span#textcolor2298{color:rgb(163,20,20)} -span#textcolor2299{color:rgb(163,20,20)} -span#textcolor2300{color:rgb(163,20,20)} +span#textcolor2295{color:rgb(43,145,175)} +span#textcolor2296{color:rgb(43,145,175)} +span#textcolor2297{color:rgb(43,145,175)} +span#textcolor2298{color:rgb(43,145,175)} +span#textcolor2299{color:rgb(43,145,175)} +span#textcolor2300{color:rgb(0,0,255)} span#textcolor2301{color:rgb(0,0,255)} span#textcolor2302{color:rgb(0,0,255)} span#textcolor2303{color:rgb(0,0,255)} -span#textcolor2304{color:rgb(163,20,20)} -span#textcolor2305{color:rgb(163,20,20)} -span#textcolor2306{color:rgb(163,20,20)} -span#textcolor2307{color:rgb(43,145,175)} -span#textcolor2308{color:rgb(43,145,175)} +span#textcolor2304{color:rgb(43,145,175)} +span#textcolor2305{color:rgb(43,145,175)} +span#textcolor2306{color:rgb(43,145,175)} +span#textcolor2307{color:rgb(163,20,20)} +span#textcolor2308{color:rgb(163,20,20)} span#textcolor2309{color:rgb(163,20,20)} span#textcolor2310{color:rgb(163,20,20)} span#textcolor2311{color:rgb(163,20,20)} span#textcolor2312{color:rgb(163,20,20)} -span#textcolor2313{color:rgb(163,20,20)} -span#textcolor2314{color:rgb(163,20,20)} -span#textcolor2315{color:rgb(0,127,0)} -span#textcolor2316{color:rgb(0,127,0)} -span#textcolor2317{color:rgb(0,127,0)} -span#textcolor2318{color:rgb(43,145,175)} +span#textcolor2313{color:rgb(0,0,255)} +span#textcolor2314{color:rgb(0,0,255)} +span#textcolor2315{color:rgb(0,0,255)} +span#textcolor2316{color:rgb(163,20,20)} +span#textcolor2317{color:rgb(163,20,20)} +span#textcolor2318{color:rgb(163,20,20)} span#textcolor2319{color:rgb(43,145,175)} span#textcolor2320{color:rgb(43,145,175)} -span#textcolor2321{color:rgb(0,0,255)} -span#textcolor2322{color:rgb(0,0,255)} -span#textcolor2323{color:rgb(43,145,175)} -span#textcolor2324{color:rgb(43,145,175)} +span#textcolor2321{color:rgb(163,20,20)} +span#textcolor2322{color:rgb(163,20,20)} +span#textcolor2323{color:rgb(163,20,20)} +span#textcolor2324{color:rgb(163,20,20)} span#textcolor2325{color:rgb(163,20,20)} span#textcolor2326{color:rgb(163,20,20)} -span#textcolor2327{color:rgb(163,20,20)} +span#textcolor2327{color:rgb(0,127,0)} +span#textcolor2328{color:rgb(0,127,0)} +span#textcolor2329{color:rgb(0,127,0)} +span#textcolor2330{color:rgb(43,145,175)} +span#textcolor2331{color:rgb(43,145,175)} +span#textcolor2332{color:rgb(43,145,175)} +span#textcolor2333{color:rgb(0,0,255)} +span#textcolor2334{color:rgb(0,0,255)} +span#textcolor2335{color:rgb(43,145,175)} +span#textcolor2336{color:rgb(43,145,175)} +span#textcolor2337{color:rgb(163,20,20)} +span#textcolor2338{color:rgb(163,20,20)} +span#textcolor2339{color:rgb(163,20,20)} pre#fancyvrb62{padding:5.69054pt;} pre#fancyvrb62{ border-top: solid 0.4pt; } pre#fancyvrb62{ border-left: solid 0.4pt; } pre#fancyvrb62{ border-bottom: solid 0.4pt; } pre#fancyvrb62{ border-right: solid 0.4pt; } -span#textcolor2328{color:rgb(0,127,0)} -span#textcolor2329{color:rgb(0,127,0)} -span#textcolor2330{color:rgb(0,127,0)} -span#textcolor2331{color:rgb(0,0,255)} -span#textcolor2332{color:rgb(0,127,0)} -span#textcolor2333{color:rgb(0,0,255)} -span#textcolor2334{color:rgb(0,127,0)} -span#textcolor2335{color:rgb(0,0,255)} -span#textcolor2336{color:rgb(0,127,0)} -span#textcolor2337{color:rgb(0,0,255)} -span#textcolor2338{color:rgb(0,127,0)} -span#textcolor2339{color:rgb(0,0,255)} -span#textcolor2340{color:rgb(43,145,175)} -span#textcolor2341{color:rgb(43,145,175)} -span#textcolor2342{color:rgb(43,145,175)} -span#textcolor2343{color:rgb(163,20,20)} -span#textcolor2344{color:rgb(163,20,20)} -span#textcolor2345{color:rgb(163,20,20)} -span#textcolor2346{color:rgb(163,20,20)} -span#textcolor2347{color:rgb(163,20,20)} -span#textcolor2348{color:rgb(163,20,20)} +span#textcolor2340{color:rgb(0,127,0)} +span#textcolor2341{color:rgb(0,127,0)} +span#textcolor2342{color:rgb(0,127,0)} +span#textcolor2343{color:rgb(0,0,255)} +span#textcolor2344{color:rgb(0,127,0)} +span#textcolor2345{color:rgb(0,0,255)} +span#textcolor2346{color:rgb(0,127,0)} +span#textcolor2347{color:rgb(0,0,255)} +span#textcolor2348{color:rgb(0,127,0)} span#textcolor2349{color:rgb(0,0,255)} -span#textcolor2350{color:rgb(43,145,175)} -span#textcolor2351{color:rgb(43,145,175)} -span#textcolor2352{color:rgb(163,20,20)} -span#textcolor2353{color:rgb(163,20,20)} -span#textcolor2354{color:rgb(163,20,20)} +span#textcolor2350{color:rgb(0,127,0)} +span#textcolor2351{color:rgb(0,0,255)} +span#textcolor2352{color:rgb(43,145,175)} +span#textcolor2353{color:rgb(43,145,175)} +span#textcolor2354{color:rgb(43,145,175)} span#textcolor2355{color:rgb(163,20,20)} span#textcolor2356{color:rgb(163,20,20)} span#textcolor2357{color:rgb(163,20,20)} -span#textcolor2358{color:rgb(0,0,255)} -span#textcolor2359{color:rgb(0,0,255)} -span#textcolor2360{color:rgb(43,145,175)} -span#textcolor2361{color:rgb(43,145,175)} -span#textcolor2362{color:rgb(163,20,20)} -span#textcolor2363{color:rgb(163,20,20)} +span#textcolor2358{color:rgb(163,20,20)} +span#textcolor2359{color:rgb(163,20,20)} +span#textcolor2360{color:rgb(163,20,20)} +span#textcolor2361{color:rgb(0,0,255)} +span#textcolor2362{color:rgb(43,145,175)} +span#textcolor2363{color:rgb(43,145,175)} span#textcolor2364{color:rgb(163,20,20)} span#textcolor2365{color:rgb(163,20,20)} span#textcolor2366{color:rgb(163,20,20)} +span#textcolor2367{color:rgb(163,20,20)} +span#textcolor2368{color:rgb(163,20,20)} +span#textcolor2369{color:rgb(163,20,20)} +span#textcolor2370{color:rgb(0,0,255)} +span#textcolor2371{color:rgb(0,0,255)} +span#textcolor2372{color:rgb(43,145,175)} +span#textcolor2373{color:rgb(43,145,175)} +span#textcolor2374{color:rgb(163,20,20)} +span#textcolor2375{color:rgb(163,20,20)} +span#textcolor2376{color:rgb(163,20,20)} +span#textcolor2377{color:rgb(163,20,20)} +span#textcolor2378{color:rgb(163,20,20)} pre#fancyvrb63{padding:5.69054pt;} pre#fancyvrb63{ border-top: solid 0.4pt; } pre#fancyvrb63{ border-left: solid 0.4pt; } pre#fancyvrb63{ border-bottom: solid 0.4pt; } pre#fancyvrb63{ border-right: solid 0.4pt; } -span#textcolor2367{color:rgb(0,127,0)} -span#textcolor2368{color:rgb(0,127,0)} -span#textcolor2369{color:rgb(0,127,0)} -span#textcolor2370{color:rgb(0,0,255)} -span#textcolor2371{color:rgb(0,127,0)} -span#textcolor2372{color:rgb(0,0,255)} -span#textcolor2373{color:rgb(0,127,0)} -span#textcolor2374{color:rgb(0,0,255)} -span#textcolor2375{color:rgb(0,127,0)} -span#textcolor2376{color:rgb(0,0,255)} -span#textcolor2377{color:rgb(0,0,255)} -span#textcolor2378{color:rgb(0,0,255)} -span#textcolor2379{color:rgb(0,0,255)} -span#textcolor2380{color:rgb(0,0,255)} -span#textcolor2381{color:rgb(43,145,175)} +span#textcolor2379{color:rgb(0,127,0)} +span#textcolor2380{color:rgb(0,127,0)} +span#textcolor2381{color:rgb(0,127,0)} span#textcolor2382{color:rgb(0,0,255)} -span#textcolor2383{color:rgb(163,20,20)} -span#textcolor2384{color:rgb(163,20,20)} -span#textcolor2385{color:rgb(163,20,20)} -span#textcolor2386{color:rgb(43,145,175)} -span#textcolor2387{color:rgb(163,20,20)} +span#textcolor2383{color:rgb(0,127,0)} +span#textcolor2384{color:rgb(0,0,255)} +span#textcolor2385{color:rgb(0,127,0)} +span#textcolor2386{color:rgb(0,0,255)} +span#textcolor2387{color:rgb(0,127,0)} span#textcolor2388{color:rgb(0,0,255)} -span#textcolor2389{color:rgb(43,145,175)} -span#textcolor2390{color:rgb(163,20,20)} -span#textcolor2391{color:rgb(163,20,20)} +span#textcolor2389{color:rgb(0,0,255)} +span#textcolor2390{color:rgb(0,0,255)} +span#textcolor2391{color:rgb(0,0,255)} +span#textcolor2392{color:rgb(0,0,255)} +span#textcolor2393{color:rgb(43,145,175)} +span#textcolor2394{color:rgb(0,0,255)} +span#textcolor2395{color:rgb(163,20,20)} +span#textcolor2396{color:rgb(163,20,20)} +span#textcolor2397{color:rgb(163,20,20)} +span#textcolor2398{color:rgb(43,145,175)} +span#textcolor2399{color:rgb(163,20,20)} +span#textcolor2400{color:rgb(0,0,255)} +span#textcolor2401{color:rgb(43,145,175)} +span#textcolor2402{color:rgb(163,20,20)} +span#textcolor2403{color:rgb(163,20,20)} pre#fancyvrb64{padding:5.69054pt;} pre#fancyvrb64{ border-top: solid 0.4pt; } pre#fancyvrb64{ border-left: solid 0.4pt; } pre#fancyvrb64{ border-bottom: solid 0.4pt; } pre#fancyvrb64{ border-right: solid 0.4pt; } -span#textcolor2392{color:rgb(0,127,0)} -span#textcolor2393{color:rgb(0,127,0)} -span#textcolor2394{color:rgb(0,127,0)} -span#textcolor2395{color:rgb(0,127,0)} -span#textcolor2396{color:rgb(0,127,0)} -span#textcolor2397{color:rgb(0,127,0)} -span#textcolor2398{color:rgb(0,127,0)} -span#textcolor2399{color:rgb(0,127,0)} -span#textcolor2400{color:rgb(0,127,0)} -span#textcolor2401{color:rgb(0,0,255)} -span#textcolor2402{color:rgb(0,127,0)} -span#textcolor2403{color:rgb(0,0,255)} span#textcolor2404{color:rgb(0,127,0)} -span#textcolor2405{color:rgb(0,0,255)} +span#textcolor2405{color:rgb(0,127,0)} span#textcolor2406{color:rgb(0,127,0)} -span#textcolor2407{color:rgb(0,0,255)} +span#textcolor2407{color:rgb(0,127,0)} span#textcolor2408{color:rgb(0,127,0)} -span#textcolor2409{color:rgb(0,0,255)} -span#textcolor2410{color:rgb(43,145,175)} +span#textcolor2409{color:rgb(0,127,0)} +span#textcolor2410{color:rgb(0,127,0)} span#textcolor2411{color:rgb(0,127,0)} span#textcolor2412{color:rgb(0,127,0)} span#textcolor2413{color:rgb(0,0,255)} -span#textcolor2414{color:rgb(0,0,255)} -span#textcolor2415{color:rgb(163,20,20)} +span#textcolor2414{color:rgb(0,127,0)} +span#textcolor2415{color:rgb(0,0,255)} span#textcolor2416{color:rgb(0,127,0)} -span#textcolor2417{color:rgb(0,127,0)} -span#textcolor2418{color:rgb(0,0,255)} +span#textcolor2417{color:rgb(0,0,255)} +span#textcolor2418{color:rgb(0,127,0)} span#textcolor2419{color:rgb(0,0,255)} -span#textcolor2420{color:rgb(163,20,20)} -span#textcolor2421{color:rgb(163,20,20)} -span#textcolor2422{color:rgb(0,127,0)} +span#textcolor2420{color:rgb(0,127,0)} +span#textcolor2421{color:rgb(0,0,255)} +span#textcolor2422{color:rgb(43,145,175)} span#textcolor2423{color:rgb(0,127,0)} span#textcolor2424{color:rgb(0,127,0)} span#textcolor2425{color:rgb(0,0,255)} -span#textcolor2426{color:rgb(43,145,175)} -span#textcolor2427{color:rgb(43,145,175)} +span#textcolor2426{color:rgb(0,0,255)} +span#textcolor2427{color:rgb(163,20,20)} span#textcolor2428{color:rgb(0,127,0)} -span#textcolor2429{color:rgb(0,0,255)} -span#textcolor2430{color:rgb(0,127,0)} +span#textcolor2429{color:rgb(0,127,0)} +span#textcolor2430{color:rgb(0,0,255)} span#textcolor2431{color:rgb(0,0,255)} -span#textcolor2432{color:rgb(0,0,255)} -span#textcolor2433{color:rgb(0,0,255)} -span#textcolor2434{color:rgb(43,145,175)} -span#textcolor2435{color:rgb(43,145,175)} -span#textcolor2436{color:rgb(163,20,20)} -span#textcolor2437{color:rgb(163,20,20)} -span#textcolor2438{color:rgb(163,20,20)} -span#textcolor2439{color:rgb(0,127,0)} -span#textcolor2440{color:rgb(0,0,255)} -span#textcolor2441{color:rgb(163,20,20)} -span#textcolor2442{color:rgb(163,20,20)} -span#textcolor2443{color:rgb(163,20,20)} +span#textcolor2432{color:rgb(163,20,20)} +span#textcolor2433{color:rgb(163,20,20)} +span#textcolor2434{color:rgb(0,127,0)} +span#textcolor2435{color:rgb(0,127,0)} +span#textcolor2436{color:rgb(0,127,0)} +span#textcolor2437{color:rgb(0,0,255)} +span#textcolor2438{color:rgb(43,145,175)} +span#textcolor2439{color:rgb(43,145,175)} +span#textcolor2440{color:rgb(0,127,0)} +span#textcolor2441{color:rgb(0,0,255)} +span#textcolor2442{color:rgb(0,127,0)} +span#textcolor2443{color:rgb(0,0,255)} span#textcolor2444{color:rgb(0,0,255)} -span#textcolor2445{color:rgb(0,127,0)} -span#textcolor2446{color:rgb(0,0,255)} -span#textcolor2447{color:rgb(163,20,20)} +span#textcolor2445{color:rgb(0,0,255)} +span#textcolor2446{color:rgb(43,145,175)} +span#textcolor2447{color:rgb(43,145,175)} span#textcolor2448{color:rgb(163,20,20)} span#textcolor2449{color:rgb(163,20,20)} -span#textcolor2450{color:rgb(0,0,255)} -span#textcolor2451{color:rgb(163,20,20)} -span#textcolor2452{color:rgb(163,20,20)} +span#textcolor2450{color:rgb(163,20,20)} +span#textcolor2451{color:rgb(0,127,0)} +span#textcolor2452{color:rgb(0,0,255)} span#textcolor2453{color:rgb(163,20,20)} -span#textcolor2454{color:rgb(0,0,255)} +span#textcolor2454{color:rgb(163,20,20)} span#textcolor2455{color:rgb(163,20,20)} -span#textcolor2456{color:rgb(163,20,20)} -span#textcolor2457{color:rgb(163,20,20)} +span#textcolor2456{color:rgb(0,0,255)} +span#textcolor2457{color:rgb(0,127,0)} span#textcolor2458{color:rgb(0,0,255)} span#textcolor2459{color:rgb(163,20,20)} span#textcolor2460{color:rgb(163,20,20)} span#textcolor2461{color:rgb(163,20,20)} -span#textcolor2462{color:rgb(163,20,20)} -span#textcolor2463{color:rgb(0,0,255)} +span#textcolor2462{color:rgb(0,0,255)} +span#textcolor2463{color:rgb(163,20,20)} span#textcolor2464{color:rgb(163,20,20)} span#textcolor2465{color:rgb(163,20,20)} -span#textcolor2466{color:rgb(163,20,20)} -span#textcolor2467{color:rgb(0,0,255)} -span#textcolor2468{color:rgb(0,0,255)} +span#textcolor2466{color:rgb(0,0,255)} +span#textcolor2467{color:rgb(163,20,20)} +span#textcolor2468{color:rgb(163,20,20)} span#textcolor2469{color:rgb(163,20,20)} -span#textcolor2470{color:rgb(163,20,20)} +span#textcolor2470{color:rgb(0,0,255)} span#textcolor2471{color:rgb(163,20,20)} -span#textcolor2472{color:rgb(0,0,255)} +span#textcolor2472{color:rgb(163,20,20)} span#textcolor2473{color:rgb(163,20,20)} span#textcolor2474{color:rgb(163,20,20)} -span#textcolor2475{color:rgb(163,20,20)} +span#textcolor2475{color:rgb(0,0,255)} span#textcolor2476{color:rgb(163,20,20)} -span#textcolor2477{color:rgb(0,0,255)} +span#textcolor2477{color:rgb(163,20,20)} span#textcolor2478{color:rgb(163,20,20)} -span#textcolor2479{color:rgb(163,20,20)} -span#textcolor2480{color:rgb(163,20,20)} -span#textcolor2481{color:rgb(0,0,255)} -span#textcolor2482{color:rgb(0,0,255)} -span#textcolor2483{color:rgb(0,127,0)} +span#textcolor2479{color:rgb(0,0,255)} +span#textcolor2480{color:rgb(0,0,255)} +span#textcolor2481{color:rgb(163,20,20)} +span#textcolor2482{color:rgb(163,20,20)} +span#textcolor2483{color:rgb(163,20,20)} span#textcolor2484{color:rgb(0,0,255)} -span#textcolor2485{color:rgb(43,145,175)} -span#textcolor2486{color:rgb(43,145,175)} +span#textcolor2485{color:rgb(163,20,20)} +span#textcolor2486{color:rgb(163,20,20)} span#textcolor2487{color:rgb(163,20,20)} span#textcolor2488{color:rgb(163,20,20)} -span#textcolor2489{color:rgb(163,20,20)} -span#textcolor2490{color:rgb(0,127,0)} -span#textcolor2491{color:rgb(0,127,0)} -span#textcolor2492{color:rgb(0,0,255)} -span#textcolor2493{color:rgb(0,127,0)} -span#textcolor2494{color:rgb(163,20,20)} -span#textcolor2495{color:rgb(163,20,20)} +span#textcolor2489{color:rgb(0,0,255)} +span#textcolor2490{color:rgb(163,20,20)} +span#textcolor2491{color:rgb(163,20,20)} +span#textcolor2492{color:rgb(163,20,20)} +span#textcolor2493{color:rgb(0,0,255)} +span#textcolor2494{color:rgb(0,0,255)} +span#textcolor2495{color:rgb(0,127,0)} +span#textcolor2496{color:rgb(0,0,255)} +span#textcolor2497{color:rgb(43,145,175)} +span#textcolor2498{color:rgb(43,145,175)} +span#textcolor2499{color:rgb(163,20,20)} +span#textcolor2500{color:rgb(163,20,20)} +span#textcolor2501{color:rgb(163,20,20)} +span#textcolor2502{color:rgb(0,127,0)} +span#textcolor2503{color:rgb(0,127,0)} +span#textcolor2504{color:rgb(0,0,255)} +span#textcolor2505{color:rgb(0,127,0)} +span#textcolor2506{color:rgb(163,20,20)} +span#textcolor2507{color:rgb(163,20,20)} pre#fancyvrb65{padding:5.69054pt;} pre#fancyvrb65{ border-top: solid 0.4pt; } pre#fancyvrb65{ border-left: solid 0.4pt; } pre#fancyvrb65{ border-bottom: solid 0.4pt; } pre#fancyvrb65{ border-right: solid 0.4pt; } -span#textcolor2496{color:rgb(0,127,0)} -span#textcolor2497{color:rgb(0,127,0)} -span#textcolor2498{color:rgb(0,127,0)} -span#textcolor2499{color:rgb(0,127,0)} -span#textcolor2500{color:rgb(0,127,0)} -span#textcolor2501{color:rgb(0,127,0)} -span#textcolor2502{color:rgb(0,127,0)} -span#textcolor2503{color:rgb(0,127,0)} -span#textcolor2504{color:rgb(0,127,0)} -span#textcolor2505{color:rgb(0,0,255)} -span#textcolor2506{color:rgb(0,127,0)} -span#textcolor2507{color:rgb(0,0,255)} span#textcolor2508{color:rgb(0,127,0)} -span#textcolor2509{color:rgb(0,0,255)} +span#textcolor2509{color:rgb(0,127,0)} span#textcolor2510{color:rgb(0,127,0)} -span#textcolor2511{color:rgb(0,0,255)} +span#textcolor2511{color:rgb(0,127,0)} span#textcolor2512{color:rgb(0,127,0)} -span#textcolor2513{color:rgb(0,0,255)} +span#textcolor2513{color:rgb(0,127,0)} span#textcolor2514{color:rgb(0,127,0)} -span#textcolor2515{color:rgb(0,0,255)} -span#textcolor2516{color:rgb(43,145,175)} -span#textcolor2517{color:rgb(0,127,0)} +span#textcolor2515{color:rgb(0,127,0)} +span#textcolor2516{color:rgb(0,127,0)} +span#textcolor2517{color:rgb(0,0,255)} span#textcolor2518{color:rgb(0,127,0)} span#textcolor2519{color:rgb(0,0,255)} -span#textcolor2520{color:rgb(0,0,255)} -span#textcolor2521{color:rgb(163,20,20)} +span#textcolor2520{color:rgb(0,127,0)} +span#textcolor2521{color:rgb(0,0,255)} span#textcolor2522{color:rgb(0,127,0)} -span#textcolor2523{color:rgb(0,127,0)} -span#textcolor2524{color:rgb(0,0,255)} +span#textcolor2523{color:rgb(0,0,255)} +span#textcolor2524{color:rgb(0,127,0)} span#textcolor2525{color:rgb(0,0,255)} -span#textcolor2526{color:rgb(163,20,20)} -span#textcolor2527{color:rgb(163,20,20)} -span#textcolor2528{color:rgb(0,127,0)} -span#textcolor2529{color:rgb(0,0,255)} -span#textcolor2530{color:rgb(43,145,175)} -span#textcolor2531{color:rgb(43,145,175)} -span#textcolor2532{color:rgb(43,145,175)} +span#textcolor2526{color:rgb(0,127,0)} +span#textcolor2527{color:rgb(0,0,255)} +span#textcolor2528{color:rgb(43,145,175)} +span#textcolor2529{color:rgb(0,127,0)} +span#textcolor2530{color:rgb(0,127,0)} +span#textcolor2531{color:rgb(0,0,255)} +span#textcolor2532{color:rgb(0,0,255)} span#textcolor2533{color:rgb(163,20,20)} -span#textcolor2534{color:rgb(163,20,20)} -span#textcolor2535{color:rgb(163,20,20)} -span#textcolor2536{color:rgb(0,127,0)} -span#textcolor2537{color:rgb(163,20,20)} +span#textcolor2534{color:rgb(0,127,0)} +span#textcolor2535{color:rgb(0,127,0)} +span#textcolor2536{color:rgb(0,0,255)} +span#textcolor2537{color:rgb(0,0,255)} span#textcolor2538{color:rgb(163,20,20)} span#textcolor2539{color:rgb(163,20,20)} span#textcolor2540{color:rgb(0,127,0)} -span#textcolor2541{color:rgb(0,127,0)} -span#textcolor2542{color:rgb(0,127,0)} -span#textcolor2543{color:rgb(0,0,255)} +span#textcolor2541{color:rgb(0,0,255)} +span#textcolor2542{color:rgb(43,145,175)} +span#textcolor2543{color:rgb(43,145,175)} span#textcolor2544{color:rgb(43,145,175)} -span#textcolor2545{color:rgb(43,145,175)} -span#textcolor2546{color:rgb(0,127,0)} -span#textcolor2547{color:rgb(0,0,255)} -span#textcolor2548{color:rgb(0,0,255)} -span#textcolor2549{color:rgb(0,0,255)} -span#textcolor2550{color:rgb(0,127,0)} -span#textcolor2551{color:rgb(0,0,255)} -span#textcolor2552{color:rgb(43,145,175)} -span#textcolor2553{color:rgb(43,145,175)} -span#textcolor2554{color:rgb(163,20,20)} -span#textcolor2555{color:rgb(163,20,20)} -span#textcolor2556{color:rgb(163,20,20)} -span#textcolor2557{color:rgb(0,127,0)} -span#textcolor2558{color:rgb(0,0,255)} -span#textcolor2559{color:rgb(163,20,20)} -span#textcolor2560{color:rgb(163,20,20)} -span#textcolor2561{color:rgb(163,20,20)} -span#textcolor2562{color:rgb(0,0,255)} -span#textcolor2563{color:rgb(0,127,0)} -span#textcolor2564{color:rgb(0,0,255)} -span#textcolor2565{color:rgb(163,20,20)} +span#textcolor2545{color:rgb(163,20,20)} +span#textcolor2546{color:rgb(163,20,20)} +span#textcolor2547{color:rgb(163,20,20)} +span#textcolor2548{color:rgb(0,127,0)} +span#textcolor2549{color:rgb(163,20,20)} +span#textcolor2550{color:rgb(163,20,20)} +span#textcolor2551{color:rgb(163,20,20)} +span#textcolor2552{color:rgb(0,127,0)} +span#textcolor2553{color:rgb(0,127,0)} +span#textcolor2554{color:rgb(0,127,0)} +span#textcolor2555{color:rgb(0,0,255)} +span#textcolor2556{color:rgb(43,145,175)} +span#textcolor2557{color:rgb(43,145,175)} +span#textcolor2558{color:rgb(0,127,0)} +span#textcolor2559{color:rgb(0,0,255)} +span#textcolor2560{color:rgb(0,0,255)} +span#textcolor2561{color:rgb(0,0,255)} +span#textcolor2562{color:rgb(0,127,0)} +span#textcolor2563{color:rgb(0,0,255)} +span#textcolor2564{color:rgb(43,145,175)} +span#textcolor2565{color:rgb(43,145,175)} span#textcolor2566{color:rgb(163,20,20)} span#textcolor2567{color:rgb(163,20,20)} -span#textcolor2568{color:rgb(0,0,255)} -span#textcolor2569{color:rgb(163,20,20)} -span#textcolor2570{color:rgb(163,20,20)} +span#textcolor2568{color:rgb(163,20,20)} +span#textcolor2569{color:rgb(0,127,0)} +span#textcolor2570{color:rgb(0,0,255)} span#textcolor2571{color:rgb(163,20,20)} -span#textcolor2572{color:rgb(0,0,255)} +span#textcolor2572{color:rgb(163,20,20)} span#textcolor2573{color:rgb(163,20,20)} -span#textcolor2574{color:rgb(163,20,20)} -span#textcolor2575{color:rgb(163,20,20)} +span#textcolor2574{color:rgb(0,0,255)} +span#textcolor2575{color:rgb(0,127,0)} span#textcolor2576{color:rgb(0,0,255)} span#textcolor2577{color:rgb(163,20,20)} span#textcolor2578{color:rgb(163,20,20)} span#textcolor2579{color:rgb(163,20,20)} -span#textcolor2580{color:rgb(163,20,20)} -span#textcolor2581{color:rgb(0,0,255)} +span#textcolor2580{color:rgb(0,0,255)} +span#textcolor2581{color:rgb(163,20,20)} span#textcolor2582{color:rgb(163,20,20)} span#textcolor2583{color:rgb(163,20,20)} -span#textcolor2584{color:rgb(163,20,20)} -span#textcolor2585{color:rgb(0,0,255)} -span#textcolor2586{color:rgb(0,0,255)} +span#textcolor2584{color:rgb(0,0,255)} +span#textcolor2585{color:rgb(163,20,20)} +span#textcolor2586{color:rgb(163,20,20)} span#textcolor2587{color:rgb(163,20,20)} -span#textcolor2588{color:rgb(163,20,20)} +span#textcolor2588{color:rgb(0,0,255)} span#textcolor2589{color:rgb(163,20,20)} -span#textcolor2590{color:rgb(0,0,255)} +span#textcolor2590{color:rgb(163,20,20)} span#textcolor2591{color:rgb(163,20,20)} span#textcolor2592{color:rgb(163,20,20)} -span#textcolor2593{color:rgb(163,20,20)} +span#textcolor2593{color:rgb(0,0,255)} span#textcolor2594{color:rgb(163,20,20)} -span#textcolor2595{color:rgb(0,0,255)} +span#textcolor2595{color:rgb(163,20,20)} span#textcolor2596{color:rgb(163,20,20)} -span#textcolor2597{color:rgb(163,20,20)} -span#textcolor2598{color:rgb(163,20,20)} -span#textcolor2599{color:rgb(0,0,255)} -span#textcolor2600{color:rgb(0,0,255)} -span#textcolor2601{color:rgb(0,127,0)} +span#textcolor2597{color:rgb(0,0,255)} +span#textcolor2598{color:rgb(0,0,255)} +span#textcolor2599{color:rgb(163,20,20)} +span#textcolor2600{color:rgb(163,20,20)} +span#textcolor2601{color:rgb(163,20,20)} span#textcolor2602{color:rgb(0,0,255)} -span#textcolor2603{color:rgb(43,145,175)} -span#textcolor2604{color:rgb(43,145,175)} +span#textcolor2603{color:rgb(163,20,20)} +span#textcolor2604{color:rgb(163,20,20)} span#textcolor2605{color:rgb(163,20,20)} span#textcolor2606{color:rgb(163,20,20)} -span#textcolor2607{color:rgb(163,20,20)} -span#textcolor2608{color:rgb(0,127,0)} -span#textcolor2609{color:rgb(0,127,0)} -span#textcolor2610{color:rgb(0,0,255)} -span#textcolor2611{color:rgb(0,127,0)} -span#textcolor2612{color:rgb(163,20,20)} -span#textcolor2613{color:rgb(163,20,20)} +span#textcolor2607{color:rgb(0,0,255)} +span#textcolor2608{color:rgb(163,20,20)} +span#textcolor2609{color:rgb(163,20,20)} +span#textcolor2610{color:rgb(163,20,20)} +span#textcolor2611{color:rgb(0,0,255)} +span#textcolor2612{color:rgb(0,0,255)} +span#textcolor2613{color:rgb(0,127,0)} +span#textcolor2614{color:rgb(0,0,255)} +span#textcolor2615{color:rgb(43,145,175)} +span#textcolor2616{color:rgb(43,145,175)} +span#textcolor2617{color:rgb(163,20,20)} +span#textcolor2618{color:rgb(163,20,20)} +span#textcolor2619{color:rgb(163,20,20)} +span#textcolor2620{color:rgb(0,127,0)} +span#textcolor2621{color:rgb(0,127,0)} +span#textcolor2622{color:rgb(0,0,255)} +span#textcolor2623{color:rgb(0,127,0)} +span#textcolor2624{color:rgb(163,20,20)} +span#textcolor2625{color:rgb(163,20,20)} pre#fancyvrb66{padding:5.69054pt;} pre#fancyvrb66{ border-top: solid 0.4pt; } pre#fancyvrb66{ border-left: solid 0.4pt; } pre#fancyvrb66{ border-bottom: solid 0.4pt; } pre#fancyvrb66{ border-right: solid 0.4pt; } -span#textcolor2614{color:rgb(0,127,0)} -span#textcolor2615{color:rgb(0,127,0)} -span#textcolor2616{color:rgb(0,127,0)} -span#textcolor2617{color:rgb(0,0,255)} -span#textcolor2618{color:rgb(0,127,0)} -span#textcolor2619{color:rgb(0,0,255)} -span#textcolor2620{color:rgb(0,127,0)} -span#textcolor2621{color:rgb(0,0,255)} -span#textcolor2622{color:rgb(0,0,255)} -span#textcolor2623{color:rgb(43,145,175)} -span#textcolor2624{color:rgb(43,145,175)} -span#textcolor2625{color:rgb(43,145,175)} -span#textcolor2626{color:rgb(43,145,175)} -span#textcolor2627{color:rgb(43,145,175)} -span#textcolor2628{color:rgb(163,20,20)} -span#textcolor2629{color:rgb(163,20,20)} -span#textcolor2630{color:rgb(163,20,20)} -span#textcolor2631{color:rgb(163,20,20)} -span#textcolor2632{color:rgb(163,20,20)} +span#textcolor2626{color:rgb(0,127,0)} +span#textcolor2627{color:rgb(0,127,0)} +span#textcolor2628{color:rgb(0,127,0)} +span#textcolor2629{color:rgb(0,0,255)} +span#textcolor2630{color:rgb(0,127,0)} +span#textcolor2631{color:rgb(0,0,255)} +span#textcolor2632{color:rgb(0,127,0)} span#textcolor2633{color:rgb(0,0,255)} -span#textcolor2634{color:rgb(163,20,20)} +span#textcolor2634{color:rgb(0,0,255)} span#textcolor2635{color:rgb(43,145,175)} span#textcolor2636{color:rgb(43,145,175)} -span#textcolor2637{color:rgb(163,20,20)} -span#textcolor2638{color:rgb(163,20,20)} -span#textcolor2639{color:rgb(163,20,20)} -span#textcolor2640{color:rgb(43,145,175)} -span#textcolor2641{color:rgb(43,145,175)} -span#textcolor2642{color:rgb(43,145,175)} +span#textcolor2637{color:rgb(43,145,175)} +span#textcolor2638{color:rgb(43,145,175)} +span#textcolor2639{color:rgb(43,145,175)} +span#textcolor2640{color:rgb(163,20,20)} +span#textcolor2641{color:rgb(163,20,20)} +span#textcolor2642{color:rgb(163,20,20)} span#textcolor2643{color:rgb(163,20,20)} -span#textcolor2644{color:rgb(43,145,175)} +span#textcolor2644{color:rgb(163,20,20)} span#textcolor2645{color:rgb(0,0,255)} -span#textcolor2646{color:rgb(0,0,255)} -span#textcolor2647{color:rgb(163,20,20)} -span#textcolor2648{color:rgb(0,0,255)} -span#textcolor2649{color:rgb(0,0,255)} -span#textcolor2650{color:rgb(0,0,255)} -span#textcolor2651{color:rgb(0,0,255)} -span#textcolor2652{color:rgb(0,0,255)} -span#textcolor2653{color:rgb(0,0,255)} -span#textcolor2654{color:rgb(0,0,255)} -span#textcolor2655{color:rgb(0,0,255)} -span#textcolor2656{color:rgb(0,0,255)} +span#textcolor2646{color:rgb(163,20,20)} +span#textcolor2647{color:rgb(43,145,175)} +span#textcolor2648{color:rgb(43,145,175)} +span#textcolor2649{color:rgb(163,20,20)} +span#textcolor2650{color:rgb(163,20,20)} +span#textcolor2651{color:rgb(163,20,20)} +span#textcolor2652{color:rgb(43,145,175)} +span#textcolor2653{color:rgb(43,145,175)} +span#textcolor2654{color:rgb(43,145,175)} +span#textcolor2655{color:rgb(163,20,20)} +span#textcolor2656{color:rgb(43,145,175)} span#textcolor2657{color:rgb(0,0,255)} span#textcolor2658{color:rgb(0,0,255)} -span#textcolor2659{color:rgb(0,0,255)} +span#textcolor2659{color:rgb(163,20,20)} span#textcolor2660{color:rgb(0,0,255)} -span#textcolor2661{color:rgb(43,145,175)} -span#textcolor2662{color:rgb(43,145,175)} -span#textcolor2663{color:rgb(163,20,20)} -span#textcolor2664{color:rgb(163,20,20)} +span#textcolor2661{color:rgb(0,0,255)} +span#textcolor2662{color:rgb(0,0,255)} +span#textcolor2663{color:rgb(0,0,255)} +span#textcolor2664{color:rgb(0,0,255)} +span#textcolor2665{color:rgb(0,0,255)} +span#textcolor2666{color:rgb(0,0,255)} +span#textcolor2667{color:rgb(0,0,255)} +span#textcolor2668{color:rgb(0,0,255)} +span#textcolor2669{color:rgb(0,0,255)} +span#textcolor2670{color:rgb(0,0,255)} +span#textcolor2671{color:rgb(0,0,255)} +span#textcolor2672{color:rgb(0,0,255)} +span#textcolor2673{color:rgb(43,145,175)} +span#textcolor2674{color:rgb(43,145,175)} +span#textcolor2675{color:rgb(163,20,20)} +span#textcolor2676{color:rgb(163,20,20)} pre#fancyvrb67{padding:5.69054pt;} pre#fancyvrb67{ border-top: solid 0.4pt; } pre#fancyvrb67{ border-left: solid 0.4pt; } @@ -3204,28 +3216,16 @@ pre#fancyvrb69{ border-top: solid 0.4pt; } pre#fancyvrb69{ border-left: solid 0.4pt; } pre#fancyvrb69{ border-bottom: solid 0.4pt; } pre#fancyvrb69{ border-right: solid 0.4pt; } -span#textcolor2665{color:rgb(0,127,0)} -span#textcolor2666{color:rgb(0,127,0)} -span#textcolor2667{color:rgb(0,127,0)} -span#textcolor2668{color:rgb(0,0,255)} -span#textcolor2669{color:rgb(0,127,0)} -span#textcolor2670{color:rgb(0,0,255)} -span#textcolor2671{color:rgb(0,127,0)} -span#textcolor2672{color:rgb(0,0,255)} -span#textcolor2673{color:rgb(0,127,0)} -span#textcolor2674{color:rgb(0,0,255)} -span#textcolor2675{color:rgb(0,0,255)} -span#textcolor2676{color:rgb(0,0,255)} -span#textcolor2677{color:rgb(0,0,255)} -span#textcolor2678{color:rgb(43,145,175)} -span#textcolor2679{color:rgb(0,0,255)} +span#textcolor2677{color:rgb(0,127,0)} +span#textcolor2678{color:rgb(0,127,0)} +span#textcolor2679{color:rgb(0,127,0)} span#textcolor2680{color:rgb(0,0,255)} -span#textcolor2681{color:rgb(0,0,255)} +span#textcolor2681{color:rgb(0,127,0)} span#textcolor2682{color:rgb(0,0,255)} -span#textcolor2683{color:rgb(0,0,255)} -span#textcolor2684{color:rgb(43,145,175)} -span#textcolor2685{color:rgb(43,145,175)} -span#textcolor2686{color:rgb(43,145,175)} +span#textcolor2683{color:rgb(0,127,0)} +span#textcolor2684{color:rgb(0,0,255)} +span#textcolor2685{color:rgb(0,127,0)} +span#textcolor2686{color:rgb(0,0,255)} span#textcolor2687{color:rgb(0,0,255)} span#textcolor2688{color:rgb(0,0,255)} span#textcolor2689{color:rgb(0,0,255)} @@ -3235,210 +3235,222 @@ span#textcolor2692{color:rgb(0,0,255)} span#textcolor2693{color:rgb(0,0,255)} span#textcolor2694{color:rgb(0,0,255)} span#textcolor2695{color:rgb(0,0,255)} -span#textcolor2696{color:rgb(0,0,255)} -span#textcolor2697{color:rgb(0,0,255)} +span#textcolor2696{color:rgb(43,145,175)} +span#textcolor2697{color:rgb(43,145,175)} span#textcolor2698{color:rgb(43,145,175)} span#textcolor2699{color:rgb(0,0,255)} -span#textcolor2700{color:rgb(43,145,175)} +span#textcolor2700{color:rgb(0,0,255)} span#textcolor2701{color:rgb(0,0,255)} -span#textcolor2702{color:rgb(0,0,255)} +span#textcolor2702{color:rgb(43,145,175)} span#textcolor2703{color:rgb(0,0,255)} span#textcolor2704{color:rgb(0,0,255)} span#textcolor2705{color:rgb(0,0,255)} span#textcolor2706{color:rgb(0,0,255)} -span#textcolor2707{color:rgb(163,20,20)} -span#textcolor2708{color:rgb(163,20,20)} -span#textcolor2709{color:rgb(163,20,20)} -span#textcolor2710{color:rgb(0,0,255)} +span#textcolor2707{color:rgb(0,0,255)} +span#textcolor2708{color:rgb(0,0,255)} +span#textcolor2709{color:rgb(0,0,255)} +span#textcolor2710{color:rgb(43,145,175)} span#textcolor2711{color:rgb(0,0,255)} -span#textcolor2712{color:rgb(163,20,20)} -span#textcolor2713{color:rgb(163,20,20)} -span#textcolor2714{color:rgb(163,20,20)} +span#textcolor2712{color:rgb(43,145,175)} +span#textcolor2713{color:rgb(0,0,255)} +span#textcolor2714{color:rgb(0,0,255)} span#textcolor2715{color:rgb(0,0,255)} span#textcolor2716{color:rgb(0,0,255)} span#textcolor2717{color:rgb(0,0,255)} -span#textcolor2718{color:rgb(43,145,175)} -span#textcolor2719{color:rgb(0,0,255)} -span#textcolor2720{color:rgb(43,145,175)} -span#textcolor2721{color:rgb(0,0,255)} -span#textcolor2722{color:rgb(0,127,0)} +span#textcolor2718{color:rgb(0,0,255)} +span#textcolor2719{color:rgb(163,20,20)} +span#textcolor2720{color:rgb(163,20,20)} +span#textcolor2721{color:rgb(163,20,20)} +span#textcolor2722{color:rgb(0,0,255)} span#textcolor2723{color:rgb(0,0,255)} -span#textcolor2724{color:rgb(0,0,255)} +span#textcolor2724{color:rgb(163,20,20)} span#textcolor2725{color:rgb(163,20,20)} span#textcolor2726{color:rgb(163,20,20)} -span#textcolor2727{color:rgb(163,20,20)} -span#textcolor2728{color:rgb(0,127,0)} -span#textcolor2729{color:rgb(0,127,0)} -span#textcolor2730{color:rgb(0,127,0)} -span#textcolor2731{color:rgb(0,127,0)} -span#textcolor2732{color:rgb(0,127,0)} -span#textcolor2733{color:rgb(0,127,0)} +span#textcolor2727{color:rgb(0,0,255)} +span#textcolor2728{color:rgb(0,0,255)} +span#textcolor2729{color:rgb(0,0,255)} +span#textcolor2730{color:rgb(43,145,175)} +span#textcolor2731{color:rgb(0,0,255)} +span#textcolor2732{color:rgb(43,145,175)} +span#textcolor2733{color:rgb(0,0,255)} span#textcolor2734{color:rgb(0,127,0)} -span#textcolor2735{color:rgb(0,127,0)} -span#textcolor2736{color:rgb(0,127,0)} -span#textcolor2737{color:rgb(0,127,0)} -span#textcolor2738{color:rgb(0,127,0)} -span#textcolor2739{color:rgb(0,127,0)} -span#textcolor2740{color:rgb(0,0,255)} -span#textcolor2741{color:rgb(43,145,175)} -span#textcolor2742{color:rgb(43,145,175)} -span#textcolor2743{color:rgb(43,145,175)} -span#textcolor2744{color:rgb(0,0,255)} -span#textcolor2745{color:rgb(43,145,175)} -span#textcolor2746{color:rgb(43,145,175)} -span#textcolor2747{color:rgb(43,145,175)} -span#textcolor2748{color:rgb(0,0,255)} -span#textcolor2749{color:rgb(163,20,20)} -span#textcolor2750{color:rgb(0,0,255)} -span#textcolor2751{color:rgb(163,20,20)} -span#textcolor2752{color:rgb(163,20,20)} -span#textcolor2753{color:rgb(163,20,20)} -span#textcolor2754{color:rgb(0,0,255)} -span#textcolor2755{color:rgb(0,0,255)} +span#textcolor2735{color:rgb(0,0,255)} +span#textcolor2736{color:rgb(0,0,255)} +span#textcolor2737{color:rgb(163,20,20)} +span#textcolor2738{color:rgb(163,20,20)} +span#textcolor2739{color:rgb(163,20,20)} +span#textcolor2740{color:rgb(0,127,0)} +span#textcolor2741{color:rgb(0,127,0)} +span#textcolor2742{color:rgb(0,127,0)} +span#textcolor2743{color:rgb(0,127,0)} +span#textcolor2744{color:rgb(0,127,0)} +span#textcolor2745{color:rgb(0,127,0)} +span#textcolor2746{color:rgb(0,127,0)} +span#textcolor2747{color:rgb(0,127,0)} +span#textcolor2748{color:rgb(0,127,0)} +span#textcolor2749{color:rgb(0,127,0)} +span#textcolor2750{color:rgb(0,127,0)} +span#textcolor2751{color:rgb(0,127,0)} +span#textcolor2752{color:rgb(0,0,255)} +span#textcolor2753{color:rgb(43,145,175)} +span#textcolor2754{color:rgb(43,145,175)} +span#textcolor2755{color:rgb(43,145,175)} span#textcolor2756{color:rgb(0,0,255)} -span#textcolor2757{color:rgb(163,20,20)} -span#textcolor2758{color:rgb(163,20,20)} -span#textcolor2759{color:rgb(163,20,20)} +span#textcolor2757{color:rgb(43,145,175)} +span#textcolor2758{color:rgb(43,145,175)} +span#textcolor2759{color:rgb(43,145,175)} span#textcolor2760{color:rgb(0,0,255)} -span#textcolor2761{color:rgb(0,127,0)} -span#textcolor2762{color:rgb(43,145,175)} +span#textcolor2761{color:rgb(163,20,20)} +span#textcolor2762{color:rgb(0,0,255)} span#textcolor2763{color:rgb(163,20,20)} -span#textcolor2764{color:rgb(0,127,0)} -span#textcolor2765{color:rgb(43,145,175)} -span#textcolor2766{color:rgb(163,20,20)} -span#textcolor2767{color:rgb(0,127,0)} +span#textcolor2764{color:rgb(163,20,20)} +span#textcolor2765{color:rgb(163,20,20)} +span#textcolor2766{color:rgb(0,0,255)} +span#textcolor2767{color:rgb(0,0,255)} span#textcolor2768{color:rgb(0,0,255)} span#textcolor2769{color:rgb(163,20,20)} span#textcolor2770{color:rgb(163,20,20)} span#textcolor2771{color:rgb(163,20,20)} span#textcolor2772{color:rgb(0,0,255)} -span#textcolor2773{color:rgb(163,20,20)} -span#textcolor2774{color:rgb(163,20,20)} +span#textcolor2773{color:rgb(0,127,0)} +span#textcolor2774{color:rgb(43,145,175)} span#textcolor2775{color:rgb(163,20,20)} -span#textcolor2776{color:rgb(163,20,20)} -span#textcolor2777{color:rgb(163,20,20)} +span#textcolor2776{color:rgb(0,127,0)} +span#textcolor2777{color:rgb(43,145,175)} span#textcolor2778{color:rgb(163,20,20)} -span#textcolor2779{color:rgb(0,0,255)} -span#textcolor2780{color:rgb(0,127,0)} -span#textcolor2781{color:rgb(0,0,255)} +span#textcolor2779{color:rgb(0,127,0)} +span#textcolor2780{color:rgb(0,0,255)} +span#textcolor2781{color:rgb(163,20,20)} span#textcolor2782{color:rgb(163,20,20)} span#textcolor2783{color:rgb(163,20,20)} -span#textcolor2784{color:rgb(163,20,20)} -span#textcolor2785{color:rgb(0,0,255)} -span#textcolor2786{color:rgb(0,0,255)} -span#textcolor2787{color:rgb(0,127,0)} -span#textcolor2788{color:rgb(0,0,255)} +span#textcolor2784{color:rgb(0,0,255)} +span#textcolor2785{color:rgb(163,20,20)} +span#textcolor2786{color:rgb(163,20,20)} +span#textcolor2787{color:rgb(163,20,20)} +span#textcolor2788{color:rgb(163,20,20)} span#textcolor2789{color:rgb(163,20,20)} span#textcolor2790{color:rgb(163,20,20)} -span#textcolor2791{color:rgb(163,20,20)} -span#textcolor2792{color:rgb(0,0,255)} -span#textcolor2793{color:rgb(43,145,175)} +span#textcolor2791{color:rgb(0,0,255)} +span#textcolor2792{color:rgb(0,127,0)} +span#textcolor2793{color:rgb(0,0,255)} span#textcolor2794{color:rgb(163,20,20)} -span#textcolor2795{color:rgb(0,127,0)} -span#textcolor2796{color:rgb(0,0,255)} +span#textcolor2795{color:rgb(163,20,20)} +span#textcolor2796{color:rgb(163,20,20)} span#textcolor2797{color:rgb(0,0,255)} -span#textcolor2798{color:rgb(163,20,20)} -span#textcolor2799{color:rgb(163,20,20)} -span#textcolor2800{color:rgb(163,20,20)} -span#textcolor2801{color:rgb(0,0,255)} -span#textcolor2802{color:rgb(43,145,175)} -span#textcolor2803{color:rgb(43,145,175)} -span#textcolor2804{color:rgb(0,127,0)} +span#textcolor2798{color:rgb(0,0,255)} +span#textcolor2799{color:rgb(0,127,0)} +span#textcolor2800{color:rgb(0,0,255)} +span#textcolor2801{color:rgb(163,20,20)} +span#textcolor2802{color:rgb(163,20,20)} +span#textcolor2803{color:rgb(163,20,20)} +span#textcolor2804{color:rgb(0,0,255)} span#textcolor2805{color:rgb(43,145,175)} span#textcolor2806{color:rgb(163,20,20)} -span#textcolor2807{color:rgb(163,20,20)} +span#textcolor2807{color:rgb(0,127,0)} span#textcolor2808{color:rgb(0,0,255)} -span#textcolor2809{color:rgb(43,145,175)} -span#textcolor2810{color:rgb(43,145,175)} +span#textcolor2809{color:rgb(0,0,255)} +span#textcolor2810{color:rgb(163,20,20)} span#textcolor2811{color:rgb(163,20,20)} span#textcolor2812{color:rgb(163,20,20)} +span#textcolor2813{color:rgb(0,0,255)} +span#textcolor2814{color:rgb(43,145,175)} +span#textcolor2815{color:rgb(43,145,175)} +span#textcolor2816{color:rgb(0,127,0)} +span#textcolor2817{color:rgb(43,145,175)} +span#textcolor2818{color:rgb(163,20,20)} +span#textcolor2819{color:rgb(163,20,20)} +span#textcolor2820{color:rgb(0,0,255)} +span#textcolor2821{color:rgb(43,145,175)} +span#textcolor2822{color:rgb(43,145,175)} +span#textcolor2823{color:rgb(163,20,20)} +span#textcolor2824{color:rgb(163,20,20)} pre#fancyvrb70{padding:5.69054pt;} pre#fancyvrb70{ border-top: solid 0.4pt; } pre#fancyvrb70{ border-left: solid 0.4pt; } pre#fancyvrb70{ border-bottom: solid 0.4pt; } pre#fancyvrb70{ border-right: solid 0.4pt; } -span#textcolor2813{color:rgb(0,127,0)} -span#textcolor2814{color:rgb(0,127,0)} -span#textcolor2815{color:rgb(0,127,0)} -span#textcolor2816{color:rgb(0,0,255)} -span#textcolor2817{color:rgb(0,127,0)} -span#textcolor2818{color:rgb(0,0,255)} -span#textcolor2819{color:rgb(0,127,0)} -span#textcolor2820{color:rgb(0,0,255)} -span#textcolor2821{color:rgb(0,127,0)} -span#textcolor2822{color:rgb(0,0,255)} -span#textcolor2823{color:rgb(43,145,175)} -span#textcolor2824{color:rgb(43,145,175)} -span#textcolor2825{color:rgb(0,0,255)} -span#textcolor2826{color:rgb(43,145,175)} -span#textcolor2827{color:rgb(0,0,255)} +span#textcolor2825{color:rgb(0,127,0)} +span#textcolor2826{color:rgb(0,127,0)} +span#textcolor2827{color:rgb(0,127,0)} span#textcolor2828{color:rgb(0,0,255)} -span#textcolor2829{color:rgb(0,0,255)} -span#textcolor2830{color:rgb(163,20,20)} -span#textcolor2831{color:rgb(163,20,20)} -span#textcolor2832{color:rgb(163,20,20)} -span#textcolor2833{color:rgb(163,20,20)} -span#textcolor2834{color:rgb(163,20,20)} -span#textcolor2835{color:rgb(163,20,20)} -span#textcolor2836{color:rgb(0,127,0)} +span#textcolor2829{color:rgb(0,127,0)} +span#textcolor2830{color:rgb(0,0,255)} +span#textcolor2831{color:rgb(0,127,0)} +span#textcolor2832{color:rgb(0,0,255)} +span#textcolor2833{color:rgb(0,127,0)} +span#textcolor2834{color:rgb(0,0,255)} +span#textcolor2835{color:rgb(43,145,175)} +span#textcolor2836{color:rgb(43,145,175)} span#textcolor2837{color:rgb(0,0,255)} -span#textcolor2838{color:rgb(0,0,255)} -span#textcolor2839{color:rgb(43,145,175)} +span#textcolor2838{color:rgb(43,145,175)} +span#textcolor2839{color:rgb(0,0,255)} span#textcolor2840{color:rgb(0,0,255)} -span#textcolor2841{color:rgb(163,20,20)} +span#textcolor2841{color:rgb(0,0,255)} span#textcolor2842{color:rgb(163,20,20)} span#textcolor2843{color:rgb(163,20,20)} -span#textcolor2844{color:rgb(0,127,0)} -span#textcolor2845{color:rgb(0,0,255)} -span#textcolor2846{color:rgb(0,0,255)} -span#textcolor2847{color:rgb(43,145,175)} -span#textcolor2848{color:rgb(0,0,255)} -span#textcolor2849{color:rgb(163,20,20)} -span#textcolor2850{color:rgb(163,20,20)} -span#textcolor2851{color:rgb(163,20,20)} -span#textcolor2852{color:rgb(0,127,0)} -span#textcolor2853{color:rgb(0,0,255)} -span#textcolor2854{color:rgb(0,0,255)} -span#textcolor2855{color:rgb(43,145,175)} -span#textcolor2856{color:rgb(0,0,255)} -span#textcolor2857{color:rgb(163,20,20)} -span#textcolor2858{color:rgb(163,20,20)} -span#textcolor2859{color:rgb(163,20,20)} -span#textcolor2860{color:rgb(0,127,0)} -span#textcolor2861{color:rgb(0,0,255)} -span#textcolor2862{color:rgb(0,0,255)} -span#textcolor2863{color:rgb(0,0,255)} -span#textcolor2864{color:rgb(0,0,255)} +span#textcolor2844{color:rgb(163,20,20)} +span#textcolor2845{color:rgb(163,20,20)} +span#textcolor2846{color:rgb(163,20,20)} +span#textcolor2847{color:rgb(163,20,20)} +span#textcolor2848{color:rgb(0,127,0)} +span#textcolor2849{color:rgb(0,0,255)} +span#textcolor2850{color:rgb(0,0,255)} +span#textcolor2851{color:rgb(43,145,175)} +span#textcolor2852{color:rgb(0,0,255)} +span#textcolor2853{color:rgb(163,20,20)} +span#textcolor2854{color:rgb(163,20,20)} +span#textcolor2855{color:rgb(163,20,20)} +span#textcolor2856{color:rgb(0,127,0)} +span#textcolor2857{color:rgb(0,0,255)} +span#textcolor2858{color:rgb(0,0,255)} +span#textcolor2859{color:rgb(43,145,175)} +span#textcolor2860{color:rgb(0,0,255)} +span#textcolor2861{color:rgb(163,20,20)} +span#textcolor2862{color:rgb(163,20,20)} +span#textcolor2863{color:rgb(163,20,20)} +span#textcolor2864{color:rgb(0,127,0)} span#textcolor2865{color:rgb(0,0,255)} span#textcolor2866{color:rgb(0,0,255)} -span#textcolor2867{color:rgb(163,20,20)} +span#textcolor2867{color:rgb(43,145,175)} span#textcolor2868{color:rgb(0,0,255)} -span#textcolor2869{color:rgb(43,145,175)} -span#textcolor2870{color:rgb(43,145,175)} -span#textcolor2871{color:rgb(43,145,175)} -span#textcolor2872{color:rgb(163,20,20)} -span#textcolor2873{color:rgb(163,20,20)} -span#textcolor2874{color:rgb(163,20,20)} +span#textcolor2869{color:rgb(163,20,20)} +span#textcolor2870{color:rgb(163,20,20)} +span#textcolor2871{color:rgb(163,20,20)} +span#textcolor2872{color:rgb(0,127,0)} +span#textcolor2873{color:rgb(0,0,255)} +span#textcolor2874{color:rgb(0,0,255)} span#textcolor2875{color:rgb(0,0,255)} -span#textcolor2876{color:rgb(163,20,20)} -span#textcolor2877{color:rgb(163,20,20)} -span#textcolor2878{color:rgb(163,20,20)} -span#textcolor2879{color:rgb(0,0,255)} +span#textcolor2876{color:rgb(0,0,255)} +span#textcolor2877{color:rgb(0,0,255)} +span#textcolor2878{color:rgb(0,0,255)} +span#textcolor2879{color:rgb(163,20,20)} span#textcolor2880{color:rgb(0,0,255)} -span#textcolor2881{color:rgb(0,0,255)} +span#textcolor2881{color:rgb(43,145,175)} span#textcolor2882{color:rgb(43,145,175)} span#textcolor2883{color:rgb(43,145,175)} span#textcolor2884{color:rgb(163,20,20)} span#textcolor2885{color:rgb(163,20,20)} span#textcolor2886{color:rgb(163,20,20)} -span#textcolor2887{color:rgb(163,20,20)} +span#textcolor2887{color:rgb(0,0,255)} span#textcolor2888{color:rgb(163,20,20)} +span#textcolor2889{color:rgb(163,20,20)} +span#textcolor2890{color:rgb(163,20,20)} +span#textcolor2891{color:rgb(0,0,255)} +span#textcolor2892{color:rgb(0,0,255)} +span#textcolor2893{color:rgb(0,0,255)} +span#textcolor2894{color:rgb(43,145,175)} +span#textcolor2895{color:rgb(43,145,175)} +span#textcolor2896{color:rgb(163,20,20)} +span#textcolor2897{color:rgb(163,20,20)} +span#textcolor2898{color:rgb(163,20,20)} +span#textcolor2899{color:rgb(163,20,20)} +span#textcolor2900{color:rgb(163,20,20)} pre#fancyvrb71{padding:5.69054pt;} pre#fancyvrb71{ border-top: solid 0.4pt; } pre#fancyvrb71{ border-left: solid 0.4pt; } pre#fancyvrb71{ border-bottom: solid 0.4pt; } pre#fancyvrb71{ border-right: solid 0.4pt; } -span#textcolor2889{color:rgb(0,0,255)} -span#textcolor2890{color:rgb(0,0,255)} +span#textcolor2901{color:rgb(0,0,255)} +span#textcolor2902{color:rgb(0,0,255)} /* end css.sty */ diff --git a/lkmpg.html b/lkmpg.html index 81dbc96..8ba660c 100644 --- a/lkmpg.html +++ b/lkmpg.html @@ -17,7 +17,7 @@

    The Linux Kernel Module Programming Guide

    Peter Jay Salzman, Michael Burian, Ori Pomerantz, Bob Mottram, Jim Huang

    -
    August 5, 2021
    +
    August 6, 2021
    @@ -49,51 +49,52 @@
      0.5.5 Code space
      0.5.6 Device Drivers
     0.6 Character Device drivers -
      0.6.1 The proc_ops Structure +
      0.6.1 The file_operations Structure
      0.6.2 The file structure
      0.6.3 Registering A Device
      0.6.4 Unregistering A Device
      0.6.5 chardev.c
      0.6.6 Writing Modules for Multiple Kernel Versions
     0.7 The /proc File System -
      0.7.1 Read and Write a /proc File -
      0.7.2 Manage /proc file with standard filesystem -
      0.7.3 Manage /proc file with seq_file -
     0.8 sysfs: Interacting with your module -
     0.9 Talking To Device Files -
     0.10 System Calls -
     0.11 Blocking Processes and threads -
      0.11.1 Sleep +
      0.7.1 The proc_ops Structure +
      0.7.2 Read and Write a /proc File +
      0.7.3 Manage /proc file with standard filesystem +
      0.7.4 Manage /proc file with seq_file +
     0.8 sysfs: Interacting with your module +
     0.9 Talking To Device Files +
     0.10 System Calls +
     0.11 Blocking Processes and threads -
      0.11.2 Completions -
     0.12 Avoiding Collisions and Deadlocks -
      0.12.1 Mutex -
      0.12.2 Spinlocks -
      0.12.3 Read and write locks -
      0.12.4 Atomic operations -
     0.13 Replacing Print Macros -
      0.13.1 Replacement -
      0.13.2 Flashing keyboard LEDs -
     0.14 Scheduling Tasks -
      0.14.1 Tasklets -
      0.14.2 Work queues -
     0.15 Interrupt Handlers -
      0.15.1 Interrupt Handlers -
      0.15.2 Detecting button presses -
      0.15.3 Bottom Half -
     0.16 Crypto -
      0.16.1 Hash functions -
      0.16.2 Symmetric key encryption -
     0.17 Standardising the interfaces: The Device Model -
     0.18 Optimizations -
      0.18.1 Likely and Unlikely conditions -
     0.19 Common Pitfalls -
      0.19.1 Using standard libraries -
      0.19.2 Disabling interrupts -
      0.19.3 Sticking your head inside a large carnivore -
     0.20 Where To Go From Here? +
      0.11.1 Sleep +
      0.11.2 Completions +
     0.12 Avoiding Collisions and Deadlocks +
      0.12.1 Mutex +
      0.12.2 Spinlocks +
      0.12.3 Read and write locks +
      0.12.4 Atomic operations +
     0.13 Replacing Print Macros +
      0.13.1 Replacement +
      0.13.2 Flashing keyboard LEDs +
     0.14 Scheduling Tasks +
      0.14.1 Tasklets +
      0.14.2 Work queues +
     0.15 Interrupt Handlers +
      0.15.1 Interrupt Handlers +
      0.15.2 Detecting button presses +
      0.15.3 Bottom Half +
     0.16 Crypto +
      0.16.1 Hash functions +
      0.16.2 Symmetric key encryption +
     0.17 Standardising the interfaces: The Device Model +
     0.18 Optimizations +
      0.18.1 Likely and Unlikely conditions +
     0.19 Common Pitfalls +
      0.19.1 Using standard libraries +
      0.19.2 Disabling interrupts +
      0.19.3 Sticking your head inside a large carnivore +
     0.20 Where To Go From Here?

    0.1 Introduction

    The Linux Kernel Module Programming Guide is a free book; you may reproduce @@ -111,10 +112,10 @@ electronic.

    Derivative works and translations of this document must be placed under the Open Software License, and the original copyright notice must remain intact. If you have contributed new material to this book, you must make the material and source -code available for your revisions. Please make revisions and updates available directly +code available for your revisions. Please make revisions and updates available directly to the document maintainer, Peter Jay Salzman <p@dirac.org>. This will allow for the merging of updates and provide consistent revisions to the Linux community. @@ -153,10 +154,10 @@ 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 +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.

    @@ -1135,129 +1136,141 @@ discussion can mean something very abstract.

    0.6 Character Device drivers

    -

    0.6.1 The proc_ops Structure

    -

    The proc_ops structure is defined in /usr/include/linux/fs.h, and holds pointers -to functions defined by the driver that perform various operations on the device. -Each field of the structure corresponds to the address of some function defined by the -driver to handle a requested operation. +

    0.6.1 The file_operations Structure

    +

    The file_operations structure is defined in /usr/include/linux/fs.h, and holds +pointers to functions defined by the driver that perform various operations on the +device. Each field of the structure corresponds to the address of some function +defined by the driver to handle a requested operation.

    For example, every character driver needs to define a function that reads from the -device. The proc_ops structure holds the address of the module’s function that -performs that operation. Here is what the definition looks like for kernel -3.0: +device. The file_operations structure holds the address of the module’s function +that performs that operation. Here is what the definition looks like for kernel +5.4:

    - 
    1struct proc_ops { 
-2    struct module *owner; 
-3    loff_t (*llseek) (struct file *, loff_t, int); 
-4    ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 
-5    ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 
-6    ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t); 
-7    ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t); 
-8    int (*iterate) (struct file *, struct dir_context *); 
-9    unsigned int (*poll) (struct file *, struct poll_table_struct *); 
-10    long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 
-11    long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 
-12    int (*mmap) (struct file *, struct vm_area_struct *); 
-13    int (*open) (struct inode *, struct file *); 
-14    int (*flush) (struct file *, fl_owner_t id); 
-15    int (*release) (struct inode *, struct file *); 
-16    int (*fsync) (struct file *, loff_t, loff_t, int datasync); 
-17    int (*aio_fsync) (struct kiocb *, int datasync); 
-18    int (*fasync) (int, struct file *, int); 
-19    int (*lock) (struct file *, int, struct file_lock *); 
-20    ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); 
-21    unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 
-22    int (*check_flags)(int); 
-23    int (*flock) (struct file *, int, struct file_lock *); 
-24    ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 
-25    ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 
-26    int (*setlease)(struct file *, long, struct file_lock **); 
-27    long (*fallocate)(struct file *file, int mode, loff_t offset, 
-28                    loff_t len); 
-29    int (*show_fdinfo)(struct seq_file *m, struct file *f); 
-30};
    -

    Some operations are not implemented by a driver. For example, a driver that + 

    1struct file_operations { 
+2    struct module *owner; 
+3    loff_t (*llseek) (struct file *, loff_t, int); 
+4    ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 
+5    ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 
+6    ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); 
+7    ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); 
+8    int (*iopoll)(struct kiocb *kiocb, bool spin); 
+9    int (*iterate) (struct file *, struct dir_context *); 
+10    int (*iterate_shared) (struct file *, struct dir_context *); 
+11    __poll_t (*poll) (struct file *, struct poll_table_struct *); 
+12    long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 
+13    long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 
+14    int (*mmap) (struct file *, struct vm_area_struct *); 
+15    unsigned long mmap_supported_flags; 
+16    int (*open) (struct inode *, struct file *); 
+17    int (*flush) (struct file *, fl_owner_t id); 
+18    int (*release) (struct inode *, struct file *); 
+19    int (*fsync) (struct file *, loff_t, loff_t, int datasync); 
+20    int (*fasync) (int, struct file *, int); 
+21    int (*lock) (struct file *, int, struct file_lock *); 
+22    ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); 
+23    unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 
+24    int (*check_flags)(int); 
+25    int (*flock) (struct file *, int, struct file_lock *); 
+26    ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 
+27    ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 
+28    int (*setlease)(struct file *, long, struct file_lock **, void **); 
+29    long (*fallocate)(struct file *file, int mode, loff_t offset, 
+30        loff_t len); 
+31    void (*show_fdinfo)(struct seq_file *m, struct file *f); 
+32    ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, 
+33        loff_t, size_t, unsigned int); 
+34    loff_t (*remap_file_range)(struct file *file_in, loff_t pos_in, 
+35             struct file *file_out, loff_t pos_out, 
+36             loff_t len, unsigned int remap_flags); 
+37    int (*fadvise)(struct file *, loff_t, loff_t, int); 
+38} __randomize_layout;
    +

    Some operations are not implemented by a driver. For example, a driver that handles a video card will not need to read from a directory structure. The -corresponding entries in the proc_ops structure should be set to NULL. -

    There is a gcc extension that makes assigning to this structure more convenient. +corresponding entries in the file_operations structure should be set to +NULL. +

    There is a gcc extension that makes assigning to this structure more convenient. You will see it in modern drivers, and may catch you by surprise. This is what the new way of assigning to the structure looks like:

    -

    - 
    1struct proc_ops fops = { 
-2              proc_read: device_read, 
-3              proc_write: device_write, 
-4              proc_open: device_open, 
-5              proc_release: device_release 
-6};
    -

    However, there is also a C99 way of assigning to elements of a structure, +

    + 
    1struct file_operations fops = { 
+2              read: device_read, 
+3              write: device_write, 
+4              open: device_open, 
+5              release: device_release 
+6};
    +

    However, there is also a C99 way of assigning to elements of a structure, designated initializers, and this is definitely preferred over using the GNU extension. You should use this syntax in case someone wants to port your driver. It will help with compatibility:

    - 
    1struct proc_ops fops = { 
-2              .proc_read = device_read, 
-3              .proc_write = device_write, 
-4              .proc_open = device_open, 
-5              .proc_release = device_release 
-6};
    -

    The meaning is clear, and you should be aware that any member of the + 

    1struct file_operations fops = { 
+2              .read = device_read, 
+3              .write = device_write, 
+4              .open = device_open, 
+5              .release = device_release 
+6};
    +

    The meaning is clear, and you should be aware that any member of the structure which you do not explicitly assign will be initialized to NULL by gcc. -

    An instance of struct proc_ops containing pointers to functions that are used to -implement read, write, open, … syscalls is commonly named fops. -

    +

    An instance of struct file_operations containing pointers to functions +that are used to implement read, write, open, … syscalls is commonly named +fops. +

    Sin Linux v5.6, the proc_ops structure was introduced to replace the use of the +file_operations structure when registering proc handlers. +

    0.6.2 The file structure

    -

    Each device is represented in the kernel by a file structure, which is defined in +

    Each device is represented in the kernel by a file structure, which is defined in linux/fs.h. Be aware that a file is a kernel level structure and never appears in a user space program. It is not the same thing as a FILE, which is defined by glibc and would never appear in a kernel space function. Also, its name is a bit misleading; it represents an abstract open ‘file’, not a file on a disk, which is represented by a structure named inode. -

    An instance of struct file is commonly named filp. You’ll also see it refered to as +

    An instance of struct file is commonly named filp. You’ll also see it refered to as struct file file. Resist the temptation. -

    Go ahead and look at the definition of file. Most of the entries you see, like struct +

    Go ahead and look at the definition of file. Most of the entries you see, like struct dentry are not used by device drivers, and you can ignore them. This is because drivers do not fill file directly; they only use structures contained in file which are created elsewhere. -

    +

    0.6.3 Registering A Device

    -

    As discussed earlier, char devices are accessed through device files, usually located in +

    As discussed earlier, char devices are accessed through device files, usually located in /dev. This is by convention. When writing a driver, it is OK to put the device file in your current directory. Just make sure you place it in /dev for a production driver. The major number tells you which driver handles which device file. The minor number is used only by the driver itself to differentiate -which device it is operating on, just in case the driver handles more than one -device. -

    Adding a driver to your system means registering it with the kernel. -This is synonymous with assigning it a major number during the module’s -initialization. You do this by using the register_chrdev function, defined by +which device it is operating on, just in case the driver handles more than one +device. +

    Adding a driver to your system means registering it with the kernel. +This is synonymous with assigning it a major number during the module’s +initialization. You do this by using the register_chrdev function, defined by linux/fs.h.

    - 
    1int register_chrdev(unsigned int major, const char *name, struct proc_ops *fops);
    -

    where unsigned int major is the major number you want to request, const char + 

    1int register_chrdev(unsigned int major, const char *name, struct file_operations *fops);
    +

    where unsigned int major is the major number you want to request, const char *name is the name of the device as it will appear in /proc/devices and struct -proc_ops *fops is a pointer to the proc_ops table for your driver. A negative return -value means the registration failed. Note that we didn’t pass the minor number to -register_chrdev. That is because the kernel doesn’t care about the minor number; -only our driver uses it. -

    Now the question is, how do you get a major number without hijacking one that’s +file_operations *fops is a pointer to the file_operations table for your driver. A +negative return value means the registration failed. Note that we didn’t pass the +minor number to register_chrdev. That is because the kernel doesn’t care about the +minor number; only our driver uses it. +

    Now the question is, how do you get a major number without hijacking one that’s already in use? The easiest way would be to look through Documentation /devices.txt and pick an unused one. That is a bad way of doing things because you will never be sure if the number you picked will be assigned later. The answer is that you can ask the kernel to assign you a dynamic major number. -

    If you pass a major number of 0 to register_chrdev, the return value will be the +

    If you pass a major number of 0 to register_chrdev, the return value will be the dynamically allocated major number. The downside is that you can not make a device file in advance, since you don’t know what the major number will be. There are a couple of ways to do this. First, the driver itself can print the newly assigned @@ -1267,10 +1280,10 @@ or write a shell script to read the file in and make the device file. The third we can have our driver make the the device file using the device_create function after a successful registration and device_destroy during the call to cleanup_module. -

    +

    0.6.4 Unregistering A Device

    -

    We can not allow the kernel module to be rmmod’ed whenever root feels like it. If the +

    We can not allow the kernel module to be rmmod’ed whenever root feels like it. If the device file is opened by a process and then we remove the kernel module, using the file would cause a call to the memory location where the appropriate function (read/write) used to be. If we are lucky, no other code was loaded there, and we’ll get @@ -1278,15 +1291,15 @@ an ugly error message. If we are unlucky, another kernel module was loaded into same location, which means a jump into the middle of another function within the kernel. The results of this would be impossible to predict, but they can not be very positive. -

    Normally, when you don’t want to allow something, you return an error code + + + +

    Normally, when you don’t want to allow something, you return an error code (a negative number) from the function which is supposed to do it. With cleanup_module that’s impossible because it is a void function. However, there is a counter which keeps track of how many processes are using your module. You can see what its value is by looking at the 3rd field of /proc/modules. If this number isn’t zero, rmmod will fail. Note that you don’t have to check the counter from within - - - cleanup_module because the check will be performed for you by the system call sys_delete_module, defined in linux/module.c. You should not use this counter directly, but there are functions defined in linux/module.h which let you increase, @@ -1296,19 +1309,19 @@ decrease and display this counter:

  • try_module_get(THIS_MODULE): Increment the use count.
  • module_put(THIS_MODULE): Decrement the use count.
    • -

    It is important to keep the counter accurate; if you ever do lose track of the +

    It is important to keep the counter accurate; if you ever do lose track of the correct usage count, you will never be able to unload the module; it’s now reboot time, boys and girls. This is bound to happen to you sooner or later during a module’s development. -

    +

    0.6.5 chardev.c

    -

    The next code sample creates a char driver named chardev. You can cat its device +

    The next code sample creates a char driver named chardev. You can cat its device file. 

    1cat /proc/devices
    -

    (or open the file with a program) and the driver will put the number of times the +

    (or open the file with a program) and the driver will put the number of times the device file has been read from into the file. We do not support writing to the file (like echo "hi" > /dev/hello), but catch these attempts and tell the user that the operation is not supported. Don’t worry if you don’t see what we @@ -1317,243 +1330,244 @@ simply read in the data and print a message acknowledging that we received it.

    - 
    1/* 
-2 *  chardev.c: Creates a read-only char device that says how many times 
-3 *  you've read from the dev file 
-4 */ 
+   
    1/* 
+2 *  chardev.c: Creates a read-only char device that says how many times 
+3 *  you've read from the dev file 
+4 */ 
 5 
-6#include <linux/cdev.h> 
-7#include <linux/delay.h> 
-8#include <linux/device.h> 
-9#include <linux/fs.h> 
-10#include <linux/init.h> 
-11#include <linux/irq.h> 
-12#include <linux/kernel.h> 
-13#include <linux/module.h> 
-14#include <linux/poll.h> 
+6#include <linux/cdev.h> 
+7#include <linux/delay.h> 
+8#include <linux/device.h> 
+9#include <linux/fs.h> 
+10#include <linux/init.h> 
+11#include <linux/irq.h> 
+12#include <linux/kernel.h> 
+13#include <linux/module.h> 
+14#include <linux/poll.h> 
 15 
-16/* 
-17 *  Prototypes - this would normally go in a .h file 
-18 */ 
-19int init_module(void); 
-20void cleanup_module(void); 
-21static int device_open(struct inode *, struct file *); 
-22static int device_release(struct inode *, struct file *); 
-23static ssize_t device_read(struct file *, char *, size_t, loff_t *); 
-24static ssize_t device_write(struct file *, const char *, size_t, loff_t *); 
+16/* 
+17 *  Prototypes - this would normally go in a .h file 
+18 */ 
+19int init_module(void); 
+20void cleanup_module(void); 
+21static int device_open(struct inode *, struct file *); 
+22static int device_release(struct inode *, struct file *); 
+23static ssize_t device_read(struct file *, char *, size_t, loff_t *); 
+24static ssize_t device_write(struct file *, const char *, size_t, loff_t *); 
 25 
-26#define SUCCESS 0 
-27#define DEVICE_NAME "chardev" /* Dev name as it appears in /proc/devices   */ 
-28#define BUF_LEN 80            /* Max length of the message from the device */ 
+26#define SUCCESS 0 
+27#define DEVICE_NAME "chardev" /* Dev name as it appears in /proc/devices   */ 
+28#define BUF_LEN 80            /* Max length of the message from the device */ 
 29 
-30/* 
-31 * Global variables are declared as static, so are global within the file. 
-32 */ 
+30/* 
+31 * Global variables are declared as static, so are global within the file. 
+32 */ 
 33 
-34static int Major;           /* Major number assigned to our device driver */ 
-35static int Device_Open = 0; /* Is device open? 
-36                             * Used to prevent multiple access to device */ 
-37static char msg[BUF_LEN];   /* The msg the device will give when asked */ 
-38static char *msg_Ptr; 
+34static int Major;           /* Major number assigned to our device driver */ 
+35static int Device_Open = 0; /* Is device open? 
+36                             * Used to prevent multiple access to device */ 
+37static char msg[BUF_LEN];   /* The msg the device will give when asked */ 
+38static char *msg_Ptr; 
 39 
-40static struct class *cls; 
+40static struct class *cls; 
 41 
-42static struct file_operations chardev_fops = {.read = device_read, 
+42static struct file_operations chardev_fops = {.read = device_read, 
 43                                              .write = device_write, 
 44                                              .open = device_open, 
 45                                              .release = device_release}; 
 46 
-47/* 
-48 * This function is called when the module is loaded 
-49 */ 
-50int init_module(void) 
+47/* 
+48 * This function is called when the module is loaded 
+49 */ 
+50int init_module(void) 
 51{ 
 52    Major = register_chrdev(0, DEVICE_NAME, &chardev_fops); 
 53 
-54    if (Major < 0) { 
-55        pr_alert("Registering char device failed with %d\n", Major); 
-56        return Major; 
+54    if (Major < 0) { 
+55        pr_alert("Registering char device failed with %d\n", Major); 
+56        return Major; 
 57    } 
 58 
-59    pr_info("I was assigned major number %d.\n", Major); 
+59    pr_info("I was assigned major number %d.\n", Major); 
 60 
 61    cls = class_create(THIS_MODULE, DEVICE_NAME); 
 62    device_create(cls, NULL, MKDEV(Major, 0), NULL, DEVICE_NAME); 
 63 
-64    pr_info("Device created on /dev/%s\n", DEVICE_NAME); 
+64    pr_info("Device created on /dev/%s\n", DEVICE_NAME); 
 65 
-66    return SUCCESS; 
+66    return SUCCESS; 
 67} 
 68 
-69/* 
-70 * This function is called when the module is unloaded 
-71 */ 
-72void cleanup_module(void) 
+69/* 
+70 * This function is called when the module is unloaded 
+71 */ 
+72void cleanup_module(void) 
 73{ 
 74    device_destroy(cls, MKDEV(Major, 0)); 
 75    class_destroy(cls); 
 76 
-77    /* 
-78     * Unregister the device 
-79     */ 
+77    /* 
+78     * Unregister the device 
+79     */ 
 80    unregister_chrdev(Major, DEVICE_NAME); 
 81} 
 82 
-83/* 
-84 * Methods 
-85 */ 
+83/* 
+84 * Methods 
+85 */ 
 86 
-87/* 
-88 * Called when a process tries to open the device file, like 
-89 * "sudo cat /dev/chardev" 
-90 */ 
-91static int device_open(struct inode *inode, struct file *file) 
+87/* 
+88 * Called when a process tries to open the device file, like 
+89 * "sudo cat /dev/chardev" 
+90 */ 
+91static int device_open(struct inode *inode, struct file *file) 
 92{ 
-93    static int counter = 0; 
+93    static int counter = 0; 
 94 
-95    if (Device_Open) 
-96        return -EBUSY; 
+95    if (Device_Open) 
+96        return -EBUSY; 
 97 
 98    Device_Open++; 
-99    sprintf(msg, "I already told you %d times Hello world!\n", counter++); 
+99    sprintf(msg, "I already told you %d times Hello world!\n", counter++); 
 100    msg_Ptr = msg; 
 101    try_module_get(THIS_MODULE); 
 102 
-103    return SUCCESS; 
+103    return SUCCESS; 
 104} 
 105 
-106/* 
-107 * Called when a process closes the device file. 
-108 */ 
-109static int device_release(struct inode *inode, struct file *file) 
+106/* 
+107 * Called when a process closes the device file. 
+108 */ 
+109static int device_release(struct inode *inode, struct file *file) 
 110{ 
-111    Device_Open--; /* We're now ready for our next caller */ 
+111    Device_Open--; /* We're now ready for our next caller */ 
 112 
-113    /* 
-114     * Decrement the usage count, or else once you opened the file, you'll 
-115     * never get get rid of the module. 
-116     */ 
+113    /* 
+114     * Decrement the usage count, or else once you opened the file, you'll 
+115     * never get get rid of the module. 
+116     */ 
 117    module_put(THIS_MODULE); 
 118 
-119    return SUCCESS; 
+119    return SUCCESS; 
 120} 
 121 
-122/* 
-123 * Called when a process, which already opened the dev file, attempts to 
-124 * read from it. 
-125 */ 
-126static ssize_t device_read(struct file *filp, /* see include/linux/fs.h   */ 
-127                           char *buffer,      /* buffer to fill with data */ 
-128                           size_t length,     /* length of the buffer     */ 
+122/* 
+123 * Called when a process, which already opened the dev file, attempts to 
+124 * read from it. 
+125 */ 
+126static ssize_t device_read(struct file *filp, /* see include/linux/fs.h   */ 
+127                           char *buffer,      /* buffer to fill with data */ 
+128                           size_t length,     /* length of the buffer     */ 
 129                           loff_t *offset) 
 130{ 
-131    /* 
-132     * Number of bytes actually written to the buffer 
-133     */ 
-134    int bytes_read = 0; 
+131    /* 
+132     * Number of bytes actually written to the buffer 
+133     */ 
+134    int bytes_read = 0; 
 135 
-136    /* 
-137     * If we're at the end of the message, 
-138     * return 0 signifying end of file 
-139     */ 
-140    if (*msg_Ptr == 0) 
-141        return 0; 
+136    /* 
+137     * If we're at the end of the message, 
+138     * return 0 signifying end of file 
+139     */ 
+140    if (*msg_Ptr == 0) 
+141        return 0; 
 142 
-143    /* 
-144     * Actually put the data into the buffer 
-145     */ 
-146    while (length && *msg_Ptr) { 
-147        /* 
-148         * The buffer is in the user data segment, not the kernel 
-149         * segment so "*" assignment won't work.  We have to use 
-150         * put_user which copies data from the kernel data segment to 
-151         * the user data segment. 
-152         */ 
+143    /* 
+144     * Actually put the data into the buffer 
+145     */ 
+146    while (length && *msg_Ptr) { 
+147        /* 
+148         * The buffer is in the user data segment, not the kernel 
+149         * segment so "*" assignment won't work.  We have to use 
+150         * put_user which copies data from the kernel data segment to 
+151         * the user data segment. 
+152         */ 
 153        put_user(*(msg_Ptr++), buffer++); 
 154 
 155        length--; 
 156        bytes_read++; 
 157    } 
 158 
-159    /* 
-160     * Most read functions return the number of bytes put into the buffer 
-161     */ 
-162    return bytes_read; 
+159    /* 
+160     * Most read functions return the number of bytes put into the buffer 
+161     */ 
+162    return bytes_read; 
 163} 
 164 
-165/* 
-166 * Called when a process writes to dev file: echo "hi" > /dev/hello 
-167 */ 
-168static ssize_t device_write(struct file *filp, 
-169                            const char *buff, 
-170                            size_t len, 
+165/* 
+166 * Called when a process writes to dev file: echo "hi" > /dev/hello 
+167 */ 
+168static ssize_t device_write(struct file *filp, 
+169                            const char *buff, 
+170                            size_t len, 
 171                            loff_t *off) 
 172{ 
-173    pr_alert("Sorry, this operation isn't supported.\n"); 
-174    return -EINVAL; 
+173    pr_alert("Sorry, this operation isn't supported.\n"); 
+174    return -EINVAL; 
 175} 
 176 
-177MODULE_LICENSE("GPL");
    
+177MODULE_LICENSE("GPL");
    -

    +

    0.6.6 Writing Modules for Multiple Kernel Versions

    -

    The system calls, which are the major interface the kernel shows to the processes, +

    The system calls, which are the major interface the kernel shows to the processes, generally stay the same across versions. A new system call may be added, but usually the old ones will behave exactly like they used to. This is necessary for backward compatibility – a new kernel version is not supposed to break regular processes. In most cases, the device files will also remain the same. On the other hand, the internal interfaces within the kernel can and do change between versions. -

    There are differences between different kernel versions, and if you want to support +

    There are differences between different kernel versions, and if you want to support multiple kernel versions, you will find yourself having to code conditional compilation directives. The way to do this to compare the macro LINUX_VERSION_CODE to the macro KERNEL_VERSION. In version a.b.c of the kernel, the value of this macro would -be 216a + 28b+ c . -

    While previous versions of this guide showed how you can write backward +be 16 8 +2 a + 2 b+ c . +

    While previous versions of this guide showed how you can write backward compatible code with such constructs in great detail, we decided to break with this tradition for the better. People interested in doing such might now use a LKMPG with a version matching to their kernel. -

    +

    0.7 The /proc File System

    -

    In Linux, there is an additional mechanism for the kernel and kernel modules to send +

    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 provides the list of modules and /proc/meminfo which stats memory usage statistics. -

    The method to use the proc file system is very similar to the one used with device +

    The method to use the proc file system is very similar to the one used with device drivers — a structure is created 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. -

    Normal file systems are located on a disk, rather than just in memory (which is +

    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 nowhere 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 +

    Because we don’t get called when the file is opened or closed, there’s nowhere 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. -

    Here a simple example showing how to use a /proc file. This is the HelloWorld +

    Here a simple example showing how to use a /proc file. This is the HelloWorld for the /proc filesystem. There are three parts: create the file /proc/helloworld in the function init_module, return a value (and a buffer) when the file /proc/helloworld is read in the callback function procfile_read, and delete the file /proc/helloworld in the function cleanup_module. -

    The /proc/helloworld is created when the module is loaded with the function +

    The /proc/helloworld is created when the module is loaded with the function proc_create. The return value is a struct proc_dir_entry , and it will be used to configure the file /proc/helloworld (for example, the owner of this file). A null return value means that the creation has failed. -

    Each time, everytime the file /proc/helloworld is read, the function +

    Each time, everytime the file /proc/helloworld is read, the function procfile_read is called. Two parameters of this function are very important: the buffer (the second parameter) and the offset (the fourth one). The content of the buffer will be returned to the application which read it (for example the @@ -1569,97 +1583,111 @@ endlessly. $ cat /proc/helloworld HelloWorld! -

    +

    - 
    1/* 
-2 *  procfs1.c 
-3 */ 
+   
    1/* 
+2 *  procfs1.c 
+3 */ 
 4 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/proc_fs.h> 
-8#include <linux/uaccess.h> 
-9#include <linux/version.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/proc_fs.h> 
+8#include <linux/uaccess.h> 
+9#include <linux/version.h> 
 10 
-11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-12#define HAVE_PROC_OPS 
-13#endif 
+11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+12#define HAVE_PROC_OPS 
+13#endif 
 14 
-15#define procfs_name "helloworld" 
+15#define procfs_name "helloworld" 
 16 
-17struct proc_dir_entry *Our_Proc_File; 
+17struct proc_dir_entry *Our_Proc_File; 
 18 
 19 
-20ssize_t procfile_read(struct file *filePointer, 
-21                      char *buffer, 
-22                      size_t buffer_length, 
+20ssize_t procfile_read(struct file *filePointer, 
+21                      char *buffer, 
+22                      size_t buffer_length, 
 23                      loff_t *offset) 
 24{ 
-25    char s[13] = "HelloWorld!\n"; 
-26    int len = sizeof(s); 
-27    ssize_t ret = len; 
+25    char s[13] = "HelloWorld!\n"; 
+26    int len = sizeof(s); 
+27    ssize_t ret = len; 
 28 
-29    if (*offset >= len || copy_to_user(buffer, s, len)) { 
-30                  pr_info("copy_to_user failed\n"); 
+29    if (*offset >= len || copy_to_user(buffer, s, len)) { 
+30                  pr_info("copy_to_user failed\n"); 
 31                  ret = 0; 
 32    } 
-33    else { 
-34        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
+33    else { 
+34        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
 35        *offset += len; 
 36    } 
 37 
-38    return ret; 
+38    return ret; 
 39} 
 40 
-41#ifdef HAVE_PROC_OPS 
-42static const struct proc_ops proc_file_fops = { 
+41#ifdef HAVE_PROC_OPS 
+42static const struct proc_ops proc_file_fops = { 
 43    .proc_read = procfile_read, 
 44}; 
-45#else 
-46static const struct file_operations proc_file_fops = { 
+45#else 
+46static const struct file_operations proc_file_fops = { 
 47    .read = procfile_read, 
 48}; 
-49#endif 
+49#endif 
 50 
-51int init_module() 
+51int init_module() 
 52{ 
 53    Our_Proc_File = proc_create(procfs_name, 0644, NULL, &proc_file_fops); 
-54    if (NULL == Our_Proc_File) { 
+54    if (NULL == Our_Proc_File) { 
 55        proc_remove(Our_Proc_File); 
-56        pr_alert("Error:Could not initialize /proc/%s\n", procfs_name); 
-57        return -ENOMEM; 
+56        pr_alert("Error:Could not initialize /proc/%s\n", procfs_name); 
+57        return -ENOMEM; 
 58    } 
 59 
-60    pr_info("/proc/%s created\n", procfs_name); 
-61    return 0; 
+60    pr_info("/proc/%s created\n", procfs_name); 
+61    return 0; 
 62} 
 63 
-64void cleanup_module() 
+64void cleanup_module() 
 65{ 
 66    proc_remove(Our_Proc_File); 
-67    pr_info("/proc/%s removed\n", procfs_name); 
+67    pr_info("/proc/%s removed\n", procfs_name); 
 68} 
 69 
-70MODULE_LICENSE("GPL");
    
-
    
+70MODULE_LICENSE("GPL");
    +

    -

    0.7.1 Read and Write a /proc File

    -

    We have seen a very simple example for a /proc file where we only read the +

    0.7.1 The proc_ops Structure

    +

    The proc_ops structure is defined in /usr/include/linux/proc_fs.h in Linux +v5.6+. In older kernels, it used file_operations for custom hooks in /proc file +system, but it contains some members that are unnecessary in VFS, and every time +VFS expands file_operations set, /proc code comes bloated. On the other hand, +not only the space, but also some operations were saved by this structure to improve +its performance. For example, the file which never disappears in /proc can set the +proc_flag as PROC_ENTRY_PERMANENT to save 2 atomic ops, 1 +allocation, 1 free in per open/read/close sequence. +

    +

    +

    0.7.2 Read and Write a /proc File

    +

    We have seen a very simple example for a /proc file where we only read the file /proc/helloworld. It is also possible to write in a /proc file. It works the same way as read, a function is called when the /proc file is written. But there is a little difference with read, data comes from user, so you have to import data from user space to kernel space (with copy_from_user or get_user) -

    The reason for copy_from_user or get_user is that Linux memory (on Intel +

    The reason for copy_from_user or get_user is that Linux memory (on 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 for each of the processes. -

    The only memory segment accessible to a process is its own, so when +

    The only memory segment accessible to a process is its own, so when writing regular programs to run as processes, there is 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 @@ -1672,148 +1700,145 @@ user space, but not for the read function because data is already in kernel space.

    - 
    1/* 
-2 *  procfs2.c -  create a "file" in /proc 
-3 */ 
-4 
-5#include <linux/kernel.h>  /* We're doing kernel work */ 
-6#include <linux/module.h>  /* Specifically, a module */ 
-7#include <linux/proc_fs.h> /* Necessary because we use the proc fs */ 
-8#include <linux/uaccess.h> /* for copy_from_user */ 
-9#include <linux/version.h> 
-10 
-11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-12#define HAVE_PROC_OPS 
-13#endif 
-14 
-15#define PROCFS_MAX_SIZE 1024 
-16#define PROCFS_NAME "buffer1k" 
-17 
-18/** 
-19 * This structure hold information about the /proc file 
-20 * 
-21 */ 
-22static struct proc_dir_entry *Our_Proc_File; 
-23 
-24/** 
-25 * The buffer used to store character for this module 
-26 * 
-27 */ 
-28static char procfs_buffer[PROCFS_MAX_SIZE]; 
-29 
-30/** 
-31 * The size of the buffer 
-32 * 
-33 */ 
-34static unsigned long procfs_buffer_size = 0; 
-35 
-36/** 
-37 * This function is called then the /proc file is read 
-38 * 
-39 */ 
-40ssize_t procfile_read(struct file *filePointer, 
-41                      char *buffer, 
-42                      size_t buffer_length, 
-43                      loff_t *offset) 
-44{ 
-45    char s[13] = "HelloWorld!\n"; 
-46    int len = sizeof(s); 
-47    ssize_t ret = len; 
-48 
-49    if (*offset >= len || copy_to_user(buffer, s, len)) { 
-50                  pr_info("copy_to_user failed\n"); 
-51                  ret = 0; 
-52    } 
-53    else { 
-54        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
-55        *offset += len; 
-56    } 
-57 
-58    return ret; 
-59} 
-60 
-61 
-62/** 
-63 * This function is called with the /proc file is written 
-64 * 
-65 */ 
-66static ssize_t procfile_write(struct file *file, 
-67                              const char *buff, 
-68                              size_t len, 
-69                              loff_t *off) 
-70{ 
-71    procfs_buffer_size = len; 
-72    if (procfs_buffer_size > PROCFS_MAX_SIZE) 
-73        procfs_buffer_size = PROCFS_MAX_SIZE; 
-74 
-75    if (copy_from_user(procfs_buffer, buff, procfs_buffer_size)) 
-76        return -EFAULT; 
-77 
-78    procfs_buffer[procfs_buffer_size] = '\0'; 
-79    return procfs_buffer_size; 
-80} 
-81 
-82#ifdef HAVE_PROC_OPS 
-83static const struct proc_ops proc_file_fops = { 
-84    .proc_read = procfile_read, 
-85    .proc_write = procfile_write, 
-86}; 
-87#else 
-88static const struct file_operations proc_file_fops = { 
-89    .read = procfile_read, 
-90    .write = procfile_write, 
-91}; 
-92#endif 
-93 
-94/** 
-95 *This function is called when the module is loaded 
-96 * 
-97 */ 
-98int init_module() 
-99{ 
-100    Our_Proc_File = proc_create(PROCFS_NAME, 0644, NULL, &proc_file_fops); 
-101    if (NULL == Our_Proc_File) { 
-102        proc_remove(Our_Proc_File); 
-103        pr_alert("Error:Could not initialize /proc/%s\n", PROCFS_NAME); 
-104        return -ENOMEM; 
-105    } 
-106 
-107    pr_info("/proc/%s created\n", PROCFS_NAME); 
-108    return 0; 
-109} 
-110 
-111/** 
-112 *This function is called when the module is unloaded 
-113 * 
-114 */ 
-115void cleanup_module() 
-116{ 
-117    proc_remove(Our_Proc_File); 
-118    pr_info("/proc/%s removed\n", PROCFS_NAME); 
-119} 
-120 
-121MODULE_LICENSE("GPL");
    - - - -

    + 

    1/* 
+2 *  procfs2.c -  create a "file" in /proc 
+3 */ 
+4 
+5#include <linux/kernel.h>  /* We're doing kernel work */ 
+6#include <linux/module.h>  /* Specifically, a module */ 
+7#include <linux/proc_fs.h> /* Necessary because we use the proc fs */ 
+8#include <linux/uaccess.h> /* for copy_from_user */ 
+9#include <linux/version.h> 
+10 
+11#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+12#define HAVE_PROC_OPS 
+13#endif 
+14 
+15#define PROCFS_MAX_SIZE 1024 
+16#define PROCFS_NAME "buffer1k" 
+17 
+18/** 
+19 * This structure hold information about the /proc file 
+20 * 
+21 */ 
+22static struct proc_dir_entry *Our_Proc_File; 
+23 
+24/** 
+25 * The buffer used to store character for this module 
+26 * 
+27 */ 
+28static char procfs_buffer[PROCFS_MAX_SIZE]; 
+29 
+30/** 
+31 * The size of the buffer 
+32 * 
+33 */ 
+34static unsigned long procfs_buffer_size = 0; 
+35 
+36/** 
+37 * This function is called then the /proc file is read 
+38 * 
+39 */ 
+40ssize_t procfile_read(struct file *filePointer, 
+41                      char *buffer, 
+42                      size_t buffer_length, 
+43                      loff_t *offset) 
+44{ 
+45    char s[13] = "HelloWorld!\n"; 
+46    int len = sizeof(s); 
+47    ssize_t ret = len; 
+48 
+49    if (*offset >= len || copy_to_user(buffer, s, len)) { 
+50                  pr_info("copy_to_user failed\n"); 
+51                  ret = 0; 
+52    } 
+53    else { 
+54        pr_info("procfile read %s\n", filePointer->f_path.dentry->d_name.name); 
+55        *offset += len; 
+56    } 
+57 
+58    return ret; 
+59} 
+60 
+61 
+62/** 
+63 * This function is called with the /proc file is written 
+64 * 
+65 */ 
+66static ssize_t procfile_write(struct file *file, 
+67                              const char *buff, 
+68                              size_t len, 
+69                              loff_t *off) 
+70{ 
+71    procfs_buffer_size = len; 
+72    if (procfs_buffer_size > PROCFS_MAX_SIZE) 
+73        procfs_buffer_size = PROCFS_MAX_SIZE; 
+74 
+75    if (copy_from_user(procfs_buffer, buff, procfs_buffer_size)) 
+76        return -EFAULT; 
+77 
+78    procfs_buffer[procfs_buffer_size] = '\0'; 
+79    return procfs_buffer_size; 
+80} 
+81 
+82#ifdef HAVE_PROC_OPS 
+83static const struct proc_ops proc_file_fops = { 
+84    .proc_read = procfile_read, 
+85    .proc_write = procfile_write, 
+86}; 
+87#else 
+88static const struct file_operations proc_file_fops = { 
+89    .read = procfile_read, 
+90    .write = procfile_write, 
+91}; 
+92#endif 
+93 
+94/** 
+95 *This function is called when the module is loaded 
+96 * 
+97 */ 
+98int init_module() 
+99{ 
+100    Our_Proc_File = proc_create(PROCFS_NAME, 0644, NULL, &proc_file_fops); 
+101    if (NULL == Our_Proc_File) { 
+102        proc_remove(Our_Proc_File); 
+103        pr_alert("Error:Could not initialize /proc/%s\n", PROCFS_NAME); 
+104        return -ENOMEM; 
+105    } 
+106 
+107    pr_info("/proc/%s created\n", PROCFS_NAME); 
+108    return 0; 
+109} 
+110 
+111/** 
+112 *This function is called when the module is unloaded 
+113 * 
+114 */ 
+115void cleanup_module() 
+116{ 
+117    proc_remove(Our_Proc_File); 
+118    pr_info("/proc/%s removed\n", PROCFS_NAME); 
+119} 
+120 
+121MODULE_LICENSE("GPL");
    +

    -

    0.7.2 Manage /proc file with standard filesystem

    -

    We have seen how to read and write a /proc file with the /proc interface. But it is +

    0.7.3 Manage /proc file with standard filesystem

    +

    We have seen how to read and write a /proc file with the /proc interface. But it is also possible to manage /proc file with inodes. The main concern is to use advanced functions, like permissions. -

    In Linux, there is a standard mechanism for file system registration. Since every +

    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, there is a special structure to hold pointers to all those functions, struct inode_operations, which includes a pointer to struct proc_ops. -

    The difference between file and inode operations is that file operations deal with +

    The difference between file and inode operations is that file operations deal with the file itself whereas inode operations deal with ways of referencing the file, such as creating links to it. -

    In /proc, whenever we register a new file, we’re allowed to specify which struct +

    In /proc, whenever we register a new file, we’re allowed to specify which struct inode_operations will be used to access to it. This is the mechanism we use, a struct inode_operations which includes a pointer to a struct proc_ops which includes pointers to our procfs_read and procfs_write functions. -

    Another interesting point here is the module_permission function. This function +

    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 @@ -1821,139 +1846,139 @@ 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. -

    It is important to note that the standard roles of read and write are reversed in +

    It is 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. -

    -

    - 
    1/* 
-2 *  procfs3.c 
-3 */ 
-4 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/proc_fs.h> 
-8#include <linux/sched.h> 
-9#include <linux/uaccess.h> 
-10#include <linux/version.h> 
-11 
-12#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-13#define HAVE_PROC_OPS 
-14#endif 
-15 
-16#define PROCFS_MAX_SIZE 2048 
-17#define PROCFS_ENTRY_FILENAME "buffer2k" 
-18 
-19struct proc_dir_entry *Our_Proc_File; 
-20static char procfs_buffer[PROCFS_MAX_SIZE]; 
-21static unsigned long procfs_buffer_size = 0; 
-22 
-23static ssize_t procfs_read(struct file *filp, 
-24                           char *buffer, 
-25                           size_t length, 
-26                           loff_t *offset) 
-27{ 
-28    static int finished = 0; 
-29    if (finished) { 
-30        pr_debug("procfs_read: END\n"); 
-31        finished = 0; 
-32        return 0; 
-33    } 
-34    finished = 1; 
-35    if (copy_to_user(buffer, procfs_buffer, procfs_buffer_size)) 
-36        return -EFAULT; 
-37    pr_debug("procfs_read: read %lu bytes\n", procfs_buffer_size); 
-38    return procfs_buffer_size; 
-39} 
-40static ssize_t procfs_write(struct file *file, 
-41                            const char *buffer, 
-42                            size_t len, 
-43                            loff_t *off) 
-44{ 
-45    if (len > PROCFS_MAX_SIZE) 
-46        procfs_buffer_size = PROCFS_MAX_SIZE; 
-47    else 
-48        procfs_buffer_size = len; 
-49    if (copy_from_user(procfs_buffer, buffer, procfs_buffer_size)) 
-50        return -EFAULT; 
-51    pr_debug("procfs_write: write %lu bytes\n", procfs_buffer_size); 
-52    return procfs_buffer_size; 
-53} 
-54int procfs_open(struct inode *inode, struct file *file) 
-55{ 
-56    try_module_get(THIS_MODULE); 
-57    return 0; 
-58} 
-59int procfs_close(struct inode *inode, struct file *file) 
-60{ 
-61    module_put(THIS_MODULE); 
-62    return 0; 
-63} 
-64 
-65#ifdef HAVE_PROC_OPS 
-66static struct proc_ops File_Ops_4_Our_Proc_File = { 
-67    .proc_read = procfs_read, 
-68    .proc_write = procfs_write, 
-69    .proc_open = procfs_open, 
-70    .proc_release = procfs_close, 
-71}; 
-72#else 
-73static const struct file_operations File_Ops_4_Our_Proc_File = { 
-74    .read = procfs_read, 
-75    .write = procfs_write, 
-76    .open = procfs_open, 
-77    .release = procfs_close, 
-78}; 
-79#endif 
-80 
-81int init_module() 
-82{ 
-83    Our_Proc_File = proc_create(PROCFS_ENTRY_FILENAME, 0644, NULL, 
-84                                &File_Ops_4_Our_Proc_File); 
-85    if (Our_Proc_File == NULL) { 
-86        remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
-87        pr_debug("Error: Could not initialize /proc/%s\n", 
-88                 PROCFS_ENTRY_FILENAME); 
-89        return -ENOMEM; 
-90    } 
-91    proc_set_size(Our_Proc_File, 80); 
-92    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
-93 
-94    pr_debug("/proc/%s created\n", PROCFS_ENTRY_FILENAME); 
-95    return 0; 
-96} 
-97void cleanup_module() 
-98{ 
-99    remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
-100    pr_debug("/proc/%s removed\n", PROCFS_ENTRY_FILENAME); 
-101} 
-102 
-103MODULE_LICENSE("GPL");
    -

    Still hungry for procfs examples? Well, first of all keep in mind, there are rumors -around, claiming that procfs is on its way out, consider using sysfs instead. Consider -using this mechanism, in case you want to document something kernel related -yourself. -

    -

    -

    0.7.3 Manage /proc file with seq_file

    -

    As we have seen, writing a /proc file may be quite “complex”. So to help people -writting /proc file, there is an API named seq_file that helps formating a /proc file +input. +

    +

    + 
    1/* 
+2 *  procfs3.c 
+3 */ 
+4 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/proc_fs.h> 
+8#include <linux/sched.h> 
+9#include <linux/uaccess.h> 
+10#include <linux/version.h> 
+11 
+12#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+13#define HAVE_PROC_OPS 
+14#endif 
+15 
+16#define PROCFS_MAX_SIZE 2048 
+17#define PROCFS_ENTRY_FILENAME "buffer2k" 
+18 
+19struct proc_dir_entry *Our_Proc_File; 
+20static char procfs_buffer[PROCFS_MAX_SIZE]; 
+21static unsigned long procfs_buffer_size = 0; 
+22 
+23static ssize_t procfs_read(struct file *filp, 
+24                           char *buffer, 
+25                           size_t length, 
+26                           loff_t *offset) 
+27{ 
+28    static int finished = 0; 
+29    if (finished) { 
+30        pr_debug("procfs_read: END\n"); 
+31        finished = 0; 
+32        return 0; 
+33    } 
+34    finished = 1; 
+35    if (copy_to_user(buffer, procfs_buffer, procfs_buffer_size)) 
+36        return -EFAULT; 
+37    pr_debug("procfs_read: read %lu bytes\n", procfs_buffer_size); 
+38    return procfs_buffer_size; 
+39} 
+40static ssize_t procfs_write(struct file *file, 
+41                            const char *buffer, 
+42                            size_t len, 
+43                            loff_t *off) 
+44{ 
+45    if (len > PROCFS_MAX_SIZE) 
+46        procfs_buffer_size = PROCFS_MAX_SIZE; 
+47    else 
+48        procfs_buffer_size = len; 
+49    if (copy_from_user(procfs_buffer, buffer, procfs_buffer_size)) 
+50        return -EFAULT; 
+51    pr_debug("procfs_write: write %lu bytes\n", procfs_buffer_size); 
+52    return procfs_buffer_size; 
+53} 
+54int procfs_open(struct inode *inode, struct file *file) 
+55{ 
+56    try_module_get(THIS_MODULE); 
+57    return 0; 
+58} 
+59int procfs_close(struct inode *inode, struct file *file) 
+60{ 
+61    module_put(THIS_MODULE); 
+62    return 0; 
+63} 
+64 
+65#ifdef HAVE_PROC_OPS 
+66static struct proc_ops File_Ops_4_Our_Proc_File = { 
+67    .proc_read = procfs_read, 
+68    .proc_write = procfs_write, 
+69    .proc_open = procfs_open, 
+70    .proc_release = procfs_close, 
+71}; 
+72#else 
+73static const struct file_operations File_Ops_4_Our_Proc_File = { 
+74    .read = procfs_read, 
+75    .write = procfs_write, 
+76    .open = procfs_open, 
+77    .release = procfs_close, 
+78}; 
+79#endif 
+80 
+81int init_module() 
+82{ 
+83    Our_Proc_File = proc_create(PROCFS_ENTRY_FILENAME, 0644, NULL, 
+84                                &File_Ops_4_Our_Proc_File); 
+85    if (Our_Proc_File == NULL) { 
+86        remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
+87        pr_debug("Error: Could not initialize /proc/%s\n", 
+88                 PROCFS_ENTRY_FILENAME); 
+89        return -ENOMEM; 
+90    } 
+91    proc_set_size(Our_Proc_File, 80); 
+92    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
+93 
+94    pr_debug("/proc/%s created\n", PROCFS_ENTRY_FILENAME); 
+95    return 0; 
+96} 
+97void cleanup_module() 
+98{ 
+99    remove_proc_entry(PROCFS_ENTRY_FILENAME, NULL); 
+100    pr_debug("/proc/%s removed\n", PROCFS_ENTRY_FILENAME); 
+101} 
+102 
+103MODULE_LICENSE("GPL");
    +

    Still hungry for procfs examples? Well, first of all keep in mind, there are rumors +around, claiming that procfs is on its way out, consider using sysfs instead. Consider +using this mechanism, in case you want to document something kernel related +yourself. +

    +

    +

    0.7.4 Manage /proc file with seq_file

    +

    As we have seen, writing a /proc file may be quite “complex”. So to help people +writting /proc file, there is an API named seq_file that helps formating a /proc file for output. It is based on sequence, which is composed of 3 functions: start(), next(), and stop(). The seq_file API starts a sequence when a user read the /proc file. -

    A sequence begins with the call of the function start(). If the return is a non +

    A sequence begins with the call of the function start(). If the return is a non NULL value, the function next() is called. This function is an iterator, the goal is to go through all the data. Each time next() is called, the function show() is also called. It writes data values in the buffer read by the user. The function next() is called until it returns NULL. The sequence ends when next() returns NULL, then the function stop() is called. -

    BE CAREFUL: when a sequence is finished, another one starts. That means that +

    BE CAREFUL: when a sequence is finished, another one starts. That means that at the end of function stop(), the function start() is called again. This loop finishes when the function start() returns NULL. You can see a scheme of this in the Figure 1. @@ -1963,282 +1988,282 @@ Figure 1

    srYrsNNYtaeenetoooertusetupstrxr((ntn))( tis)istrr teeaNreNatUaUtmLtLmeLmLen?e?ntntt +

    srYrsNNYtaeenetoooertusetupstrxr((ntn))( tis)istrr teeaNreNatUaUtmLtLmeLmLen?e?ntntt

    -
    Figure 1:How seq_file works
    +
    Figure 1:How seq_file works
    -

    The seq_file provides basic functions for proc_ops, such as seq_read, +

    The seq_file provides basic functions for proc_ops, such as seq_read, seq_lseek, and some others. But nothing to write in the /proc file. Of course, you can still use the same way as in the previous example.

    - 
    1/* 
-2 *  procfs4.c -  create a "file" in /proc 
-3 *  This program uses the seq_file library to manage the /proc file. 
-4 * 
-5 */ 
-6 
-7#include <linux/kernel.h>   /* We're doing kernel work */ 
-8#include <linux/module.h>   /* Specifically, a module */ 
-9#include <linux/proc_fs.h>  /* Necessary because we use proc fs */ 
-10#include <linux/seq_file.h> /* for seq_file */ 
-11#include <linux/version.h> 
-12 
-13#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-14#define HAVE_PROC_OPS 
-15#endif 
-16 
-17#define PROC_NAME "iter" 
-18 
-19MODULE_LICENSE("GPL"); 
-20 
-21/** 
-22 * This function is called at the beginning of a sequence. 
-23 * ie, when: 
-24 *      - the /proc file is read (first time) 
-25 *      - after the function stop (end of sequence) 
-26 * 
-27 */ 
-28static void *my_seq_start(struct seq_file *s, loff_t *pos) 
-29{ 
-30    static unsigned long counter = 0; 
-31 
-32    /* beginning a new sequence ? */ 
-33    if (*pos == 0) { 
-34        /* yes => return a non null value to begin the sequence */ 
-35        return &counter; 
-36    } else { 
-37        /* no => it's the end of the sequence, return end to stop reading */ 
-38        *pos = 0; 
-39        return NULL; 
-40    } 
-41} 
-42 
-43/** 
-44 * This function is called after the beginning of a sequence. 
-45 * It's called untill the return is NULL (this ends the sequence). 
-46 * 
-47 */ 
-48static void *my_seq_next(struct seq_file *s, void *v, loff_t *pos) 
-49{ 
-50    unsigned long *tmp_v = (unsigned long *) v; 
-51    (*tmp_v)++; 
-52    (*pos)++; 
-53    return NULL; 
-54} 
-55 
-56/** 
-57 * This function is called at the end of a sequence 
-58 * 
-59 */ 
-60static void my_seq_stop(struct seq_file *s, void *v) 
-61{ 
-62    /* nothing to do, we use a static value in start() */ 
-63} 
-64 
-65/** 
-66 * This function is called for each "step" of a sequence 
-67 * 
-68 */ 
-69static int my_seq_show(struct seq_file *s, void *v) 
-70{ 
-71    loff_t *spos = (loff_t *) v; 
-72 
-73    seq_printf(s, "%Ld\n", *spos); 
-74    return 0; 
-75} 
-76 
-77/** 
-78 * This structure gather "function" to manage the sequence 
-79 * 
-80 */ 
-81static struct seq_operations my_seq_ops = {.start = my_seq_start, 
-82                                           .next = my_seq_next, 
-83                                           .stop = my_seq_stop, 
-84                                           .show = my_seq_show}; 
-85 
-86/** 
-87 * This function is called when the /proc file is open. 
-88 * 
-89 */ 
-90static int my_open(struct inode *inode, struct file *file) 
-91{ 
-92    return seq_open(file, &my_seq_ops); 
-93}; 
-94 
-95/** 
-96 * This structure gather "function" that manage the /proc file 
-97 * 
-98 */ 
-99#ifdef HAVE_PROC_OPS 
-100static const struct proc_ops my_file_ops = { 
-101    .proc_open = my_open, 
-102    .proc_read = seq_read, 
-103    .proc_lseek = seq_lseek, 
-104    .proc_release = seq_release, 
-105}; 
-106#else 
-107static const struct file_operations my_file_ops = { 
-108    .open = my_open, 
-109    .read = seq_read, 
-110    .llseek = seq_lseek, 
-111    .release = seq_release, 
-112}; 
-113#endif 
-114 
-115/** 
-116 * This function is called when the module is loaded 
-117 * 
-118 */ 
-119int init_module(void) 
-120{ 
-121    struct proc_dir_entry *entry; 
-122 
-123    entry = proc_create(PROC_NAME, 0, NULL, &my_file_ops); 
-124    if (entry == NULL) { 
-125        remove_proc_entry(PROC_NAME, NULL); 
-126        pr_debug("Error: Could not initialize /proc/%s\n", PROC_NAME); 
-127        return -ENOMEM; 
-128    } 
-129 
-130    return 0; 
-131} 
-132 
-133/** 
-134 * This function is called when the module is unloaded. 
-135 * 
-136 */ 
-137void cleanup_module(void) 
-138{ 
-139    remove_proc_entry(PROC_NAME, NULL); 
-140    pr_debug("/proc/%s removed\n", PROC_NAME); 
-141}
    -

    If you want more information, you can read this web page: + 

    1/* 
+2 *  procfs4.c -  create a "file" in /proc 
+3 *  This program uses the seq_file library to manage the /proc file. 
+4 * 
+5 */ 
+6 
+7#include <linux/kernel.h>   /* We're doing kernel work */ 
+8#include <linux/module.h>   /* Specifically, a module */ 
+9#include <linux/proc_fs.h>  /* Necessary because we use proc fs */ 
+10#include <linux/seq_file.h> /* for seq_file */ 
+11#include <linux/version.h> 
+12 
+13#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+14#define HAVE_PROC_OPS 
+15#endif 
+16 
+17#define PROC_NAME "iter" 
+18 
+19MODULE_LICENSE("GPL"); 
+20 
+21/** 
+22 * This function is called at the beginning of a sequence. 
+23 * ie, when: 
+24 *      - the /proc file is read (first time) 
+25 *      - after the function stop (end of sequence) 
+26 * 
+27 */ 
+28static void *my_seq_start(struct seq_file *s, loff_t *pos) 
+29{ 
+30    static unsigned long counter = 0; 
+31 
+32    /* beginning a new sequence ? */ 
+33    if (*pos == 0) { 
+34        /* yes => return a non null value to begin the sequence */ 
+35        return &counter; 
+36    } else { 
+37        /* no => it's the end of the sequence, return end to stop reading */ 
+38        *pos = 0; 
+39        return NULL; 
+40    } 
+41} 
+42 
+43/** 
+44 * This function is called after the beginning of a sequence. 
+45 * It's called untill the return is NULL (this ends the sequence). 
+46 * 
+47 */ 
+48static void *my_seq_next(struct seq_file *s, void *v, loff_t *pos) 
+49{ 
+50    unsigned long *tmp_v = (unsigned long *) v; 
+51    (*tmp_v)++; 
+52    (*pos)++; 
+53    return NULL; 
+54} 
+55 
+56/** 
+57 * This function is called at the end of a sequence 
+58 * 
+59 */ 
+60static void my_seq_stop(struct seq_file *s, void *v) 
+61{ 
+62    /* nothing to do, we use a static value in start() */ 
+63} 
+64 
+65/** 
+66 * This function is called for each "step" of a sequence 
+67 * 
+68 */ 
+69static int my_seq_show(struct seq_file *s, void *v) 
+70{ 
+71    loff_t *spos = (loff_t *) v; 
+72 
+73    seq_printf(s, "%Ld\n", *spos); 
+74    return 0; 
+75} 
+76 
+77/** 
+78 * This structure gather "function" to manage the sequence 
+79 * 
+80 */ 
+81static struct seq_operations my_seq_ops = {.start = my_seq_start, 
+82                                           .next = my_seq_next, 
+83                                           .stop = my_seq_stop, 
+84                                           .show = my_seq_show}; 
+85 
+86/** 
+87 * This function is called when the /proc file is open. 
+88 * 
+89 */ 
+90static int my_open(struct inode *inode, struct file *file) 
+91{ 
+92    return seq_open(file, &my_seq_ops); 
+93}; 
+94 
+95/** 
+96 * This structure gather "function" that manage the /proc file 
+97 * 
+98 */ 
+99#ifdef HAVE_PROC_OPS 
+100static const struct proc_ops my_file_ops = { 
+101    .proc_open = my_open, 
+102    .proc_read = seq_read, 
+103    .proc_lseek = seq_lseek, 
+104    .proc_release = seq_release, 
+105}; 
+106#else 
+107static const struct file_operations my_file_ops = { 
+108    .open = my_open, 
+109    .read = seq_read, 
+110    .llseek = seq_lseek, 
+111    .release = seq_release, 
+112}; 
+113#endif 
+114 
+115/** 
+116 * This function is called when the module is loaded 
+117 * 
+118 */ 
+119int init_module(void) 
+120{ 
+121    struct proc_dir_entry *entry; 
+122 
+123    entry = proc_create(PROC_NAME, 0, NULL, &my_file_ops); 
+124    if (entry == NULL) { 
+125        remove_proc_entry(PROC_NAME, NULL); 
+126        pr_debug("Error: Could not initialize /proc/%s\n", PROC_NAME); 
+127        return -ENOMEM; 
+128    } 
+129 
+130    return 0; 
+131} 
+132 
+133/** 
+134 * This function is called when the module is unloaded. 
+135 * 
+136 */ 
+137void cleanup_module(void) 
+138{ 
+139    remove_proc_entry(PROC_NAME, NULL); 
+140    pr_debug("/proc/%s removed\n", PROC_NAME); 
+141}
    +

    If you want more information, you can read this web page:

    -

    You can also read the code of fs/seq_file.c in the linux kernel. +

    You can also read the code of fs/seq_file.c in the linux kernel.

    -

    0.8 sysfs: Interacting with your module

    -

    sysfs allows you to interact with the running kernel from userspace by reading or +

    0.8 sysfs: Interacting with your module

    +

    sysfs allows you to interact with the running kernel from userspace by reading or setting variables inside of modules. This can be useful for debugging purposes, or just as an interface for applications or scripts. You can find sysfs directories and files under the sys directory on your system.

    - 
    1ls -l /sys
    -

    An example of a hello world module which includes the creation of a variable + 

    1ls -l /sys
    +

    An example of a hello world module which includes the creation of a variable accessible via sysfs is given below.

    - 
    1/* 
-2 *  hello-sysfs.c sysfs example 
-3 */ 
-4#include <linux/fs.h> 
-5#include <linux/init.h> 
-6#include <linux/kobject.h> 
-7#include <linux/module.h> 
-8#include <linux/string.h> 
-9#include <linux/sysfs.h> 
-10 
-11MODULE_LICENSE("GPL"); 
-12 
-13static struct kobject *mymodule; 
-14 
-15/* the variable you want to be able to change */ 
-16static int myvariable = 0; 
-17 
-18static ssize_t myvariable_show(struct kobject *kobj, 
-19                               struct kobj_attribute *attr, 
-20                               char *buf) 
-21{ 
-22    return sprintf(buf, "%d\n", myvariable); 
-23} 
-24 
-25static ssize_t myvariable_store(struct kobject *kobj, 
-26                                struct kobj_attribute *attr, 
-27                                char *buf, 
-28                                size_t count) 
-29{ 
-30    sscanf(buf, "%du", &myvariable); 
-31    return count; 
-32} 
-33 
-34 
-35static struct kobj_attribute myvariable_attribute = 
-36    __ATTR(myvariable, 0660, myvariable_show, (void *) myvariable_store); 
-37 
-38static int __init mymodule_init(void) 
-39{ 
-40    int error = 0; 
-41 
-42    pr_info("mymodule: initialised\n"); 
-43 
-44    mymodule = kobject_create_and_add("mymodule", kernel_kobj); 
-45    if (!mymodule) 
-46        return -ENOMEM; 
-47 
-48    error = sysfs_create_file(mymodule, &myvariable_attribute.attr); 
-49    if (error) { 
-50        pr_info( 
-51            "failed to create the myvariable file " 
-52            "in /sys/kernel/mymodule\n"); 
-53    } 
-54 
-55    return error; 
-56} 
-57 
-58static void __exit mymodule_exit(void) 
-59{ 
-60    pr_info("mymodule: Exit success\n"); 
-61    kobject_put(mymodule); 
-62} 
-63 
-64module_init(mymodule_init); 
-65module_exit(mymodule_exit);
    -

    Make and install the module: + 

    1/* 
+2 *  hello-sysfs.c sysfs example 
+3 */ 
+4#include <linux/fs.h> 
+5#include <linux/init.h> 
+6#include <linux/kobject.h> 
+7#include <linux/module.h> 
+8#include <linux/string.h> 
+9#include <linux/sysfs.h> 
+10 
+11MODULE_LICENSE("GPL"); 
+12 
+13static struct kobject *mymodule; 
+14 
+15/* the variable you want to be able to change */ 
+16static int myvariable = 0; 
+17 
+18static ssize_t myvariable_show(struct kobject *kobj, 
+19                               struct kobj_attribute *attr, 
+20                               char *buf) 
+21{ 
+22    return sprintf(buf, "%d\n", myvariable); 
+23} 
+24 
+25static ssize_t myvariable_store(struct kobject *kobj, 
+26                                struct kobj_attribute *attr, 
+27                                char *buf, 
+28                                size_t count) 
+29{ 
+30    sscanf(buf, "%du", &myvariable); 
+31    return count; 
+32} 
+33 
+34 
+35static struct kobj_attribute myvariable_attribute = 
+36    __ATTR(myvariable, 0660, myvariable_show, (void *) myvariable_store); 
+37 
+38static int __init mymodule_init(void) 
+39{ 
+40    int error = 0; 
+41 
+42    pr_info("mymodule: initialised\n"); 
+43 
+44    mymodule = kobject_create_and_add("mymodule", kernel_kobj); 
+45    if (!mymodule) 
+46        return -ENOMEM; 
+47 
+48    error = sysfs_create_file(mymodule, &myvariable_attribute.attr); 
+49    if (error) { 
+50        pr_info( 
+51            "failed to create the myvariable file " 
+52            "in /sys/kernel/mymodule\n"); 
+53    } 
+54 
+55    return error; 
+56} 
+57 
+58static void __exit mymodule_exit(void) 
+59{ 
+60    pr_info("mymodule: Exit success\n"); 
+61    kobject_put(mymodule); 
+62} 
+63 
+64module_init(mymodule_init); 
+65module_exit(mymodule_exit);
    +

    Make and install the module:

    - 
    1make 
-2sudo insmod hello-sysfs.ko
    -

    Check that it exists: + 

    1make 
+2sudo insmod hello-sysfs.ko
    +

    Check that it exists:

    - 
    1sudo lsmod | grep hello_sysfs
    -

    What is the current value of myvariable ? + 

    1sudo lsmod | grep hello_sysfs
    +

    What is the current value of myvariable ?

    - 
    1cat /sys/kernel/mymodule/myvariable
    -

    Set the value of myvariable and check that it changed. + 

    1cat /sys/kernel/mymodule/myvariable
    +

    Set the value of myvariable and check that it changed.

    - 
    1echo "32" > /sys/kernel/mymodule/myvariable 
-2cat /sys/kernel/mymodule/myvariable
    -

    Finally, remove the test module: + 

    1echo "32" > /sys/kernel/mymodule/myvariable 
+2cat /sys/kernel/mymodule/myvariable
    +

    Finally, remove the test module:

    - 
    1sudo rmmod hello_sysfs
    -

    + 

    1sudo rmmod hello_sysfs
    +

    -

    0.9 Talking To Device Files

    -

    Device files are supposed to represent physical devices. Most physical devices are +

    0.9 Talking To Device Files

    +

    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 +

    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 @@ -2248,7 +2273,7 @@ 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 +

    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), both or neither. Notice @@ -2258,600 +2283,600 @@ write ioctl’s (to return information to a process), both or neither. Notice here the roles of read and write are reversed again, so in ioctl’s read is to send information to the kernel and write is to receive information from the kernel. -

    The ioctl function is called with three parameters: the file descriptor of the +

    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. You will not be able to pass a structure this way, but you will be able to pass a pointer to the structure. -

    The ioctl number encodes the major device number, the type of the ioctl, the +

    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 +

    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.

    - 
    1/* 
-2 *  chardev2.c - Create an input/output character device 
-3 */ 
-4 
-5#include <linux/cdev.h> 
-6#include <linux/delay.h> 
-7#include <linux/device.h> 
-8#include <linux/fs.h> 
-9#include <linux/init.h> 
-10#include <linux/irq.h> 
-11#include <linux/kernel.h> /* We're doing kernel work */ 
-12#include <linux/module.h> /* Specifically, a module */ 
-13#include <linux/poll.h> 
-14 
-15#include "chardev.h" 
-16#define SUCCESS 0 
-17#define DEVICE_NAME "char_dev" 
-18#define BUF_LEN 80 
-19 
-20/* 
-21 * Is the device open right now? Used to prevent 
-22 * concurent access into the same device 
-23 */ 
-24static int Device_Open = 0; 
-25 
-26/* 
-27 * The message the device will give when asked 
-28 */ 
-29static char Message[BUF_LEN]; 
-30 
-31/* 
-32 * How far did the process reading the message get? 
-33 * Useful if the message is larger than the size of the 
-34 * buffer we get to fill in device_read. 
-35 */ 
-36static char *Message_Ptr; 
-37 
-38static int Major; /* Major number assigned to our device driver */ 
-39static struct class *cls; 
-40 
-41/* 
-42 * This is called whenever a process attempts to open the device file 
-43 */ 
-44static int device_open(struct inode *inode, struct file *file) 
-45{ 
-46    pr_info("device_open(%p)\n", file); 
-47 
-48    /* 
-49     * We don't want to talk to two processes at the same time 
-50     */ 
-51    if (Device_Open) 
-52        return -EBUSY; 
-53 
-54    Device_Open++; 
-55    /* 
-56     * Initialize the message 
-57     */ 
-58    Message_Ptr = Message; 
-59    try_module_get(THIS_MODULE); 
-60    return SUCCESS; 
-61} 
-62 
-63static int device_release(struct inode *inode, struct file *file) 
-64{ 
-65    pr_info("device_release(%p,%p)\n", inode, file); 
-66 
-67    /* 
-68     * We're now ready for our next caller 
-69     */ 
-70    Device_Open--; 
-71 
-72    module_put(THIS_MODULE); 
-73    return SUCCESS; 
-74} 
-75 
-76/* 
-77 * This function is called whenever a process which has already opened the 
-78 * device file attempts to read from it. 
-79 */ 
-80static ssize_t device_read(struct file *file,   /* see include/linux/fs.h   */ 
-81                           char __user *buffer, /* buffer to be 
-82                                                 * filled with data */ 
-83                           size_t length,       /* length of the buffer     */ 
-84                           loff_t *offset) 
-85{ 
-86    /* 
-87     * Number of bytes actually written to the buffer 
-88     */ 
-89    int bytes_read = 0; 
-90 
-91    pr_info("device_read(%p,%p,%ld)\n", file, buffer, length); 
-92 
-93    /* 
-94     * If we're at the end of the message, return 0 
-95     * (which signifies end of file) 
-96     */ 
-97    if (*Message_Ptr == 0) 
-98        return 0; 
-99 
-100    /* 
-101     * Actually put the data into the buffer 
-102     */ 
-103    while (length && *Message_Ptr) { 
-104        /* 
-105         * Because the buffer is in the user data segment, 
-106         * not the kernel data segment, assignment wouldn't 
-107         * work. Instead, we have to use put_user which 
-108         * copies data from the kernel data segment to the 
-109         * user data segment. 
-110         */ 
-111        put_user(*(Message_Ptr++), buffer++); 
-112        length--; 
-113        bytes_read++; 
-114    } 
-115 
-116    pr_info("Read %d bytes, %ld left\n", bytes_read, length); 
-117 
-118    /* 
-119     * Read functions are supposed to return the number 
-120     * of bytes actually inserted into the buffer 
-121     */ 
-122    return bytes_read; 
-123} 
-124 
-125/* 
-126 * This function is called when somebody tries to 
-127 * write into our device file. 
-128 */ 
-129static ssize_t device_write(struct file *file, 
-130                            const char __user *buffer, 
-131                            size_t length, 
-132                            loff_t *offset) 
-133{ 
-134    int i; 
-135 
-136    pr_info("device_write(%p,%s,%ld)", file, buffer, length); 
-137 
-138    for (i = 0; i < length && i < BUF_LEN; i++) 
-139        get_user(Message[i], buffer + i); 
-140 
-141    Message_Ptr = Message; 
-142 
-143    /* 
-144     * Again, return the number of input characters used 
-145     */ 
-146    return i; 
-147} 
-148 
-149/* 
-150 * This function is called whenever a process tries to do an ioctl on our 
-151 * device file. We get two extra parameters (additional to the inode and file 
-152 * structures, which all device functions get): the number of the ioctl called 
-153 * and the parameter given to the ioctl function. 
-154 * 
-155 * If the ioctl is write or read/write (meaning output is returned to the 
-156 * calling process), the ioctl call returns the output of this function. 
-157 * 
-158 */ 
-159long device_ioctl(struct file *file,      /* ditto */ 
-160                  unsigned int ioctl_num, /* number and param for ioctl */ 
-161                  unsigned long ioctl_param) 
-162{ 
-163    int i; 
-164    char *temp; 
-165    char ch; 
-166 
-167    /* 
-168     * Switch according to the ioctl called 
-169     */ 
-170    switch (ioctl_num) { 
-171    case IOCTL_SET_MSG: 
-172        /* 
-173         * Receive a pointer to a message (in user space) and set that 
-174         * to be the device's message.  Get the parameter given to 
-175         * ioctl by the process. 
-176         */ 
-177        temp = (char *) ioctl_param; 
-178 
-179        /* 
-180         * Find the length of the message 
-181         */ 
-182        get_user(ch, temp); 
-183        for (i = 0; ch && i < BUF_LEN; i++, temp++) 
-184            get_user(ch, temp); 
-185 
-186        device_write(file, (char *) ioctl_param, i, 0); 
-187        break; 
-188 
-189    case IOCTL_GET_MSG: 
-190        /* 
-191         * Give the current message to the calling process - 
-192         * the parameter we got is a pointer, fill it. 
-193         */ 
-194        i = device_read(file, (char *) ioctl_param, 99, 0); 
-195 
-196        /* 
-197         * Put a zero at the end of the buffer, so it will be 
-198         * properly terminated 
-199         */ 
-200        put_user('\0', (char *) ioctl_param + i); 
-201        break; 
-202 
-203    case IOCTL_GET_NTH_BYTE: 
-204        /* 
-205         * This ioctl is both input (ioctl_param) and 
-206         * output (the return value of this function) 
-207         */ 
-208        return Message[ioctl_param]; 
-209        break; 
-210    } 
-211 
-212    return SUCCESS; 
-213} 
-214 
-215/* Module Declarations */ 
-216 
-217/* 
-218 * This structure will hold the functions to be called 
-219 * when a process does something to the device we 
-220 * created. Since a pointer to this structure is kept in 
-221 * the devices table, it can't be local to 
-222 * init_module. NULL is for unimplemented functions. 
-223 */ 
-224struct file_operations Fops = { 
-225    .read = device_read, 
-226    .write = device_write, 
-227    .unlocked_ioctl = device_ioctl, 
-228    .open = device_open, 
-229    .release = device_release, /* a.k.a. close */ 
-230}; 
-231 
-232/* 
-233 * Initialize the module - Register the character device 
-234 */ 
-235int init_module() 
-236{ 
-237    int ret_val; 
-238    /* 
-239     * Register the character device (atleast try) 
-240     */ 
-241    ret_val = register_chrdev(MAJOR_NUM, DEVICE_NAME, &Fops); 
-242 
-243    /* 
-244     * Negative values signify an error 
-245     */ 
-246    if (ret_val < 0) { 
-247        pr_alert("%s failed with %d\n", 
-248                 "Sorry, registering the character device ", ret_val); 
-249        return ret_val; 
-250    } 
-251 
-252    Major = ret_val; 
-253 
-254    cls = class_create(THIS_MODULE, DEVICE_FILE_NAME); 
-255    device_create(cls, NULL, MKDEV(Major, MAJOR_NUM), NULL, DEVICE_FILE_NAME); 
-256 
-257    pr_info("Device created on /dev/%s\n", DEVICE_FILE_NAME); 
-258 
-259    return 0; 
-260} 
-261 
-262/* 
-263 * Cleanup - unregister the appropriate file from /proc 
-264 */ 
-265void cleanup_module() 
-266{ 
-267    device_destroy(cls, MKDEV(Major, 0)); 
-268    class_destroy(cls); 
-269 
-270    /* 
-271     * Unregister the device 
-272     */ 
-273    unregister_chrdev(Major, DEVICE_NAME); 
-274} 
-275 
-276MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  chardev2.c - Create an input/output character device 
+3 */ 
+4 
+5#include <linux/cdev.h> 
+6#include <linux/delay.h> 
+7#include <linux/device.h> 
+8#include <linux/fs.h> 
+9#include <linux/init.h> 
+10#include <linux/irq.h> 
+11#include <linux/kernel.h> /* We're doing kernel work */ 
+12#include <linux/module.h> /* Specifically, a module */ 
+13#include <linux/poll.h> 
+14 
+15#include "chardev.h" 
+16#define SUCCESS 0 
+17#define DEVICE_NAME "char_dev" 
+18#define BUF_LEN 80 
+19 
+20/* 
+21 * Is the device open right now? Used to prevent 
+22 * concurent access into the same device 
+23 */ 
+24static int Device_Open = 0; 
+25 
+26/* 
+27 * The message the device will give when asked 
+28 */ 
+29static char Message[BUF_LEN]; 
+30 
+31/* 
+32 * How far did the process reading the message get? 
+33 * Useful if the message is larger than the size of the 
+34 * buffer we get to fill in device_read. 
+35 */ 
+36static char *Message_Ptr; 
+37 
+38static int Major; /* Major number assigned to our device driver */ 
+39static struct class *cls; 
+40 
+41/* 
+42 * This is called whenever a process attempts to open the device file 
+43 */ 
+44static int device_open(struct inode *inode, struct file *file) 
+45{ 
+46    pr_info("device_open(%p)\n", file); 
+47 
+48    /* 
+49     * We don't want to talk to two processes at the same time 
+50     */ 
+51    if (Device_Open) 
+52        return -EBUSY; 
+53 
+54    Device_Open++; 
+55    /* 
+56     * Initialize the message 
+57     */ 
+58    Message_Ptr = Message; 
+59    try_module_get(THIS_MODULE); 
+60    return SUCCESS; 
+61} 
+62 
+63static int device_release(struct inode *inode, struct file *file) 
+64{ 
+65    pr_info("device_release(%p,%p)\n", inode, file); 
+66 
+67    /* 
+68     * We're now ready for our next caller 
+69     */ 
+70    Device_Open--; 
+71 
+72    module_put(THIS_MODULE); 
+73    return SUCCESS; 
+74} 
+75 
+76/* 
+77 * This function is called whenever a process which has already opened the 
+78 * device file attempts to read from it. 
+79 */ 
+80static ssize_t device_read(struct file *file,   /* see include/linux/fs.h   */ 
+81                           char __user *buffer, /* buffer to be 
+82                                                 * filled with data */ 
+83                           size_t length,       /* length of the buffer     */ 
+84                           loff_t *offset) 
+85{ 
+86    /* 
+87     * Number of bytes actually written to the buffer 
+88     */ 
+89    int bytes_read = 0; 
+90 
+91    pr_info("device_read(%p,%p,%ld)\n", file, buffer, length); 
+92 
+93    /* 
+94     * If we're at the end of the message, return 0 
+95     * (which signifies end of file) 
+96     */ 
+97    if (*Message_Ptr == 0) 
+98        return 0; 
+99 
+100    /* 
+101     * Actually put the data into the buffer 
+102     */ 
+103    while (length && *Message_Ptr) { 
+104        /* 
+105         * Because the buffer is in the user data segment, 
+106         * not the kernel data segment, assignment wouldn't 
+107         * work. Instead, we have to use put_user which 
+108         * copies data from the kernel data segment to the 
+109         * user data segment. 
+110         */ 
+111        put_user(*(Message_Ptr++), buffer++); 
+112        length--; 
+113        bytes_read++; 
+114    } 
+115 
+116    pr_info("Read %d bytes, %ld left\n", bytes_read, length); 
+117 
+118    /* 
+119     * Read functions are supposed to return the number 
+120     * of bytes actually inserted into the buffer 
+121     */ 
+122    return bytes_read; 
+123} 
+124 
+125/* 
+126 * This function is called when somebody tries to 
+127 * write into our device file. 
+128 */ 
+129static ssize_t device_write(struct file *file, 
+130                            const char __user *buffer, 
+131                            size_t length, 
+132                            loff_t *offset) 
+133{ 
+134    int i; 
+135 
+136    pr_info("device_write(%p,%s,%ld)", file, buffer, length); 
+137 
+138    for (i = 0; i < length && i < BUF_LEN; i++) 
+139        get_user(Message[i], buffer + i); 
+140 
+141    Message_Ptr = Message; 
+142 
+143    /* 
+144     * Again, return the number of input characters used 
+145     */ 
+146    return i; 
+147} 
+148 
+149/* 
+150 * This function is called whenever a process tries to do an ioctl on our 
+151 * device file. We get two extra parameters (additional to the inode and file 
+152 * structures, which all device functions get): the number of the ioctl called 
+153 * and the parameter given to the ioctl function. 
+154 * 
+155 * If the ioctl is write or read/write (meaning output is returned to the 
+156 * calling process), the ioctl call returns the output of this function. 
+157 * 
+158 */ 
+159long device_ioctl(struct file *file,      /* ditto */ 
+160                  unsigned int ioctl_num, /* number and param for ioctl */ 
+161                  unsigned long ioctl_param) 
+162{ 
+163    int i; 
+164    char *temp; 
+165    char ch; 
+166 
+167    /* 
+168     * Switch according to the ioctl called 
+169     */ 
+170    switch (ioctl_num) { 
+171    case IOCTL_SET_MSG: 
+172        /* 
+173         * Receive a pointer to a message (in user space) and set that 
+174         * to be the device's message.  Get the parameter given to 
+175         * ioctl by the process. 
+176         */ 
+177        temp = (char *) ioctl_param; 
+178 
+179        /* 
+180         * Find the length of the message 
+181         */ 
+182        get_user(ch, temp); 
+183        for (i = 0; ch && i < BUF_LEN; i++, temp++) 
+184            get_user(ch, temp); 
+185 
+186        device_write(file, (char *) ioctl_param, i, 0); 
+187        break; 
+188 
+189    case IOCTL_GET_MSG: 
+190        /* 
+191         * Give the current message to the calling process - 
+192         * the parameter we got is a pointer, fill it. 
+193         */ 
+194        i = device_read(file, (char *) ioctl_param, 99, 0); 
+195 
+196        /* 
+197         * Put a zero at the end of the buffer, so it will be 
+198         * properly terminated 
+199         */ 
+200        put_user('\0', (char *) ioctl_param + i); 
+201        break; 
+202 
+203    case IOCTL_GET_NTH_BYTE: 
+204        /* 
+205         * This ioctl is both input (ioctl_param) and 
+206         * output (the return value of this function) 
+207         */ 
+208        return Message[ioctl_param]; 
+209        break; 
+210    } 
+211 
+212    return SUCCESS; 
+213} 
+214 
+215/* Module Declarations */ 
+216 
+217/* 
+218 * This structure will hold the functions to be called 
+219 * when a process does something to the device we 
+220 * created. Since a pointer to this structure is kept in 
+221 * the devices table, it can't be local to 
+222 * init_module. NULL is for unimplemented functions. 
+223 */ 
+224struct file_operations Fops = { 
+225    .read = device_read, 
+226    .write = device_write, 
+227    .unlocked_ioctl = device_ioctl, 
+228    .open = device_open, 
+229    .release = device_release, /* a.k.a. close */ 
+230}; 
+231 
+232/* 
+233 * Initialize the module - Register the character device 
+234 */ 
+235int init_module() 
+236{ 
+237    int ret_val; 
+238    /* 
+239     * Register the character device (atleast try) 
+240     */ 
+241    ret_val = register_chrdev(MAJOR_NUM, DEVICE_NAME, &Fops); 
+242 
+243    /* 
+244     * Negative values signify an error 
+245     */ 
+246    if (ret_val < 0) { 
+247        pr_alert("%s failed with %d\n", 
+248                 "Sorry, registering the character device ", ret_val); 
+249        return ret_val; 
+250    } 
+251 
+252    Major = ret_val; 
+253 
+254    cls = class_create(THIS_MODULE, DEVICE_FILE_NAME); 
+255    device_create(cls, NULL, MKDEV(Major, MAJOR_NUM), NULL, DEVICE_FILE_NAME); 
+256 
+257    pr_info("Device created on /dev/%s\n", DEVICE_FILE_NAME); 
+258 
+259    return 0; 
+260} 
+261 
+262/* 
+263 * Cleanup - unregister the appropriate file from /proc 
+264 */ 
+265void cleanup_module() 
+266{ 
+267    device_destroy(cls, MKDEV(Major, 0)); 
+268    class_destroy(cls); 
+269 
+270    /* 
+271     * Unregister the device 
+272     */ 
+273    unregister_chrdev(Major, DEVICE_NAME); 
+274} 
+275 
+276MODULE_LICENSE("GPL");

    - 
    1/* 
-2 *  chardev.h - the header file with the ioctl definitions. 
-3 * 
-4 *  The declarations here have to be in a header file, because 
-5 *  they need to be known both to the kernel module 
-6 *  (in chardev.c) and the process calling ioctl (ioctl.c) 
-7 */ 
-8 
-9#ifndef CHARDEV_H 
-10#define CHARDEV_H 
-11 
-12#include <linux/ioctl.h> 
-13 
-14/* 
-15 * The major device number. We can't rely on dynamic 
-16 * registration any more, because ioctls need to know 
-17 * it. 
-18 */ 
-19#define MAJOR_NUM 100 
-20 
-21/* 
-22 * Set the message of the device driver 
-23 */ 
-24#define IOCTL_SET_MSG _IOW(MAJOR_NUM, 0, char *) 
-25/* 
-26 * _IOW means that we're creating an ioctl command 
-27 * number for passing information from a user process 
-28 * to the kernel module. 
-29 * 
-30 * The first arguments, MAJOR_NUM, is the major device 
-31 * number we're using. 
-32 * 
-33 * The second argument is the number of the command 
-34 * (there could be several with different meanings). 
-35 * 
-36 * The third argument is the type we want to get from 
-37 * the process to the kernel. 
-38 */ 
-39 
-40/* 
-41 * Get the message of the device driver 
-42 */ 
-43#define IOCTL_GET_MSG _IOR(MAJOR_NUM, 1, char *) 
-44/* 
-45 * This IOCTL is used for output, to get the message 
-46 * of the device driver. However, we still need the 
-47 * buffer to place the message in to be input, 
-48 * as it is allocated by the process. 
-49 */ 
-50 
-51/* 
-52 * Get the n'th byte of the message 
-53 */ 
-54#define IOCTL_GET_NTH_BYTE _IOWR(MAJOR_NUM, 2, int) 
-55/* 
-56 * The IOCTL is used for both input and output. It 
-57 * receives from the user a number, n, and returns 
-58 * Message[n]. 
-59 */ 
-60 
-61/* 
-62 * The name of the device file 
-63 */ 
-64#define DEVICE_FILE_NAME "char_dev" 
-65 
-66#endif
    + 
    1/* 
+2 *  chardev.h - the header file with the ioctl definitions. 
+3 * 
+4 *  The declarations here have to be in a header file, because 
+5 *  they need to be known both to the kernel module 
+6 *  (in chardev.c) and the process calling ioctl (ioctl.c) 
+7 */ 
+8 
+9#ifndef CHARDEV_H 
+10#define CHARDEV_H 
+11 
+12#include <linux/ioctl.h> 
+13 
+14/* 
+15 * The major device number. We can't rely on dynamic 
+16 * registration any more, because ioctls need to know 
+17 * it. 
+18 */ 
+19#define MAJOR_NUM 100 
+20 
+21/* 
+22 * Set the message of the device driver 
+23 */ 
+24#define IOCTL_SET_MSG _IOW(MAJOR_NUM, 0, char *) 
+25/* 
+26 * _IOW means that we're creating an ioctl command 
+27 * number for passing information from a user process 
+28 * to the kernel module. 
+29 * 
+30 * The first arguments, MAJOR_NUM, is the major device 
+31 * number we're using. 
+32 * 
+33 * The second argument is the number of the command 
+34 * (there could be several with different meanings). 
+35 * 
+36 * The third argument is the type we want to get from 
+37 * the process to the kernel. 
+38 */ 
+39 
+40/* 
+41 * Get the message of the device driver 
+42 */ 
+43#define IOCTL_GET_MSG _IOR(MAJOR_NUM, 1, char *) 
+44/* 
+45 * This IOCTL is used for output, to get the message 
+46 * of the device driver. However, we still need the 
+47 * buffer to place the message in to be input, 
+48 * as it is allocated by the process. 
+49 */ 
+50 
+51/* 
+52 * Get the n'th byte of the message 
+53 */ 
+54#define IOCTL_GET_NTH_BYTE _IOWR(MAJOR_NUM, 2, int) 
+55/* 
+56 * The IOCTL is used for both input and output. It 
+57 * receives from the user a number, n, and returns 
+58 * Message[n]. 
+59 */ 
+60 
+61/* 
+62 * The name of the device file 
+63 */ 
+64#define DEVICE_FILE_NAME "char_dev" 
+65 
+66#endif

    - 
    1/* 
-2 *  ioctl.c 
-3 */ 
-4#include <linux/cdev.h> 
-5#include <linux/fs.h> 
-6#include <linux/init.h> 
-7#include <linux/ioctl.h> 
-8#include <linux/module.h> 
-9#include <linux/slab.h> 
-10#include <linux/uaccess.h> 
-11 
-12struct ioctl_arg { 
-13    unsigned int reg; 
-14    unsigned int val; 
-15}; 
-16 
-17/* Documentation/ioctl/ioctl-number.txt */ 
-18#define IOC_MAGIC '\x66' 
-19 
-20#define IOCTL_VALSET _IOW(IOC_MAGIC, 0, struct ioctl_arg) 
-21#define IOCTL_VALGET _IOR(IOC_MAGIC, 1, struct ioctl_arg) 
-22#define IOCTL_VALGET_NUM _IOR(IOC_MAGIC, 2, int) 
-23#define IOCTL_VALSET_NUM _IOW(IOC_MAGIC, 3, int) 
-24 
-25#define IOCTL_VAL_MAXNR 3 
-26#define DRIVER_NAME "ioctltest" 
-27 
-28static unsigned int test_ioctl_major = 0; 
-29static unsigned int num_of_dev = 1; 
-30static struct cdev test_ioctl_cdev; 
-31static int ioctl_num = 0; 
-32 
-33struct test_ioctl_data { 
-34    unsigned char val; 
-35    rwlock_t lock; 
-36}; 
-37 
-38static long test_ioctl_ioctl(struct file *filp, 
-39                             unsigned int cmd, 
-40                             unsigned long arg) 
-41{ 
-42    struct test_ioctl_data *ioctl_data = filp->private_data; 
-43    int retval = 0; 
-44    unsigned char val; 
-45    struct ioctl_arg data; 
-46    memset(&data, 0, sizeof(data)); 
-47 
-48    switch (cmd) { 
-49    case IOCTL_VALSET: 
-50 
-51        /* 
-52        if (!capable(CAP_SYS_ADMIN)) { 
-53         retval = -EPERM; 
-54         goto done; 
-55        } 
-56        if (!access_ok(VERIFY_READ, (void __user *)arg, _IOC_SIZE(cmd))) { 
-57         retval = -EFAULT; 
-58         goto done; 
-59        } 
-60        */ 
-61        if (copy_from_user(&data, (int __user *) arg, sizeof(data))) { 
-62            retval = -EFAULT; 
-63            goto done; 
-64        } 
-65 
-66        pr_alert("IOCTL set val:%x .\n", data.val); 
-67        write_lock(&ioctl_data->lock); 
-68        ioctl_data->val = data.val; 
-69        write_unlock(&ioctl_data->lock); 
-70        break; 
-71 
-72    case IOCTL_VALGET: 
-73        /* 
-74        if (!access_ok(VERIFY_WRITE, (void __user *)arg, _IOC_SIZE(cmd))) { 
-75                                     retval = -EFAULT; 
-76                                     goto done; 
-77                             } 
-78        */ 
-79        read_lock(&ioctl_data->lock); 
-80        val = ioctl_data->val; 
-81        read_unlock(&ioctl_data->lock); 
-82        data.val = val; 
-83 
-84        if (copy_to_user((int __user *) arg, &data, sizeof(data))) { 
-85            retval = -EFAULT; 
-86            goto done; 
-87        } 
-88 
-89        break; 
-90 
-91    case IOCTL_VALGET_NUM: 
-92        retval = __put_user(ioctl_num, (int __user *) arg); 
-93        break; 
-94 
-95    case IOCTL_VALSET_NUM: 
-96        /* 
-97        if (!capable(CAP_SYS_ADMIN)) 
-98         return -EPERM; 
-99        */ 
-100        ioctl_num = arg; 
-101        break; 
-102 
-103    default: 
-104        retval = -ENOTTY; 
-105    } 
-106 
-107done: 
-108    return retval; 
-109} 
-110 
-111ssize_t test_ioctl_read(struct file *filp, 
-112                        char __user *buf, 
-113                        size_t count, 
-114                        loff_t *f_pos) 
-115{ 
-116    struct test_ioctl_data *ioctl_data = filp->private_data; 
-117    unsigned char val; 
-118    int retval; 
-119    int i = 0; 
-120    read_lock(&ioctl_data->lock); 
-121    val = ioctl_data->val; 
-122    read_unlock(&ioctl_data->lock); 
-123 
-124    for (; i < count; i++) { 
-125        if (copy_to_user(&buf[i], &val, 1)) { 
-126            retval = -EFAULT; 
-127            goto out; 
-128        } 
-129    } 
-130 
-131    retval = count; 
-132out: 
-133    return retval; 
-134} 
-135 
-136static int test_ioctl_close(struct inode *inode, struct file *filp) 
-137{ 
-138    pr_alert("%s call.\n", __func__); 
-139 
-140    if (filp->private_data) { 
-141        kfree(filp->private_data); 
-142        filp->private_data = NULL; 
-143    } 
-144 
-145    return 0; 
-146} 
-147 
-148static int test_ioctl_open(struct inode *inode, struct file *filp) 
-149{ 
-150    struct test_ioctl_data *ioctl_data; 
-151    pr_alert("%s call.\n", __func__); 
-152    ioctl_data = kmalloc(sizeof(struct test_ioctl_data), GFP_KERNEL); 
-153 
-154    if (ioctl_data == NULL) { 
-155        return -ENOMEM; 
-156    } 
-157 
-158    rwlock_init(&ioctl_data->lock); 
-159    ioctl_data->val = 0xFF; 
-160    filp->private_data = ioctl_data; 
-161    return 0; 
-162} 
-163 
-164struct file_operations fops = { 
-165    .owner = THIS_MODULE, 
-166    .open = test_ioctl_open, 
-167    .release = test_ioctl_close, 
-168    .read = test_ioctl_read, 
-169    .unlocked_ioctl = test_ioctl_ioctl, 
-170}; 
-171 
-172static int ioctl_init(void) 
-173{ 
-174    dev_t dev = MKDEV(test_ioctl_major, 0); 
-175    int alloc_ret = 0; 
-176    int cdev_ret = 0; 
-177    alloc_ret = alloc_chrdev_region(&dev, 0, num_of_dev, DRIVER_NAME); 
-178 
-179    if (alloc_ret) { 
-180        goto error; 
-181    } 
-182 
-183    test_ioctl_major = MAJOR(dev); 
-184    cdev_init(&test_ioctl_cdev, &fops); 
-185    cdev_ret = cdev_add(&test_ioctl_cdev, dev, num_of_dev); 
-186 
-187    if (cdev_ret) { 
-188        goto error; 
-189    } 
-190 
-191    pr_alert("%s driver(major: %d) installed.\n", DRIVER_NAME, 
-192             test_ioctl_major); 
-193    return 0; 
-194error: 
-195 
-196    if (cdev_ret == 0) { 
-197        cdev_del(&test_ioctl_cdev); 
-198    } 
-199 
-200    if (alloc_ret == 0) { 
-201        unregister_chrdev_region(dev, num_of_dev); 
-202    } 
-203 
-204    return -1; 
-205} 
-206 
-207static void ioctl_exit(void) 
-208{ 
-209    dev_t dev = MKDEV(test_ioctl_major, 0); 
-210    cdev_del(&test_ioctl_cdev); 
-211    unregister_chrdev_region(dev, num_of_dev); 
-212    pr_alert("%s driver removed.\n", DRIVER_NAME); 
-213} 
-214 
-215module_init(ioctl_init); 
-216module_exit(ioctl_exit); 
-217 
-218MODULE_LICENSE("GPL"); 
-219MODULE_DESCRIPTION("This is test_ioctl module");
    -

    + 

    1/* 
+2 *  ioctl.c 
+3 */ 
+4#include <linux/cdev.h> 
+5#include <linux/fs.h> 
+6#include <linux/init.h> 
+7#include <linux/ioctl.h> 
+8#include <linux/module.h> 
+9#include <linux/slab.h> 
+10#include <linux/uaccess.h> 
+11 
+12struct ioctl_arg { 
+13    unsigned int reg; 
+14    unsigned int val; 
+15}; 
+16 
+17/* Documentation/ioctl/ioctl-number.txt */ 
+18#define IOC_MAGIC '\x66' 
+19 
+20#define IOCTL_VALSET _IOW(IOC_MAGIC, 0, struct ioctl_arg) 
+21#define IOCTL_VALGET _IOR(IOC_MAGIC, 1, struct ioctl_arg) 
+22#define IOCTL_VALGET_NUM _IOR(IOC_MAGIC, 2, int) 
+23#define IOCTL_VALSET_NUM _IOW(IOC_MAGIC, 3, int) 
+24 
+25#define IOCTL_VAL_MAXNR 3 
+26#define DRIVER_NAME "ioctltest" 
+27 
+28static unsigned int test_ioctl_major = 0; 
+29static unsigned int num_of_dev = 1; 
+30static struct cdev test_ioctl_cdev; 
+31static int ioctl_num = 0; 
+32 
+33struct test_ioctl_data { 
+34    unsigned char val; 
+35    rwlock_t lock; 
+36}; 
+37 
+38static long test_ioctl_ioctl(struct file *filp, 
+39                             unsigned int cmd, 
+40                             unsigned long arg) 
+41{ 
+42    struct test_ioctl_data *ioctl_data = filp->private_data; 
+43    int retval = 0; 
+44    unsigned char val; 
+45    struct ioctl_arg data; 
+46    memset(&data, 0, sizeof(data)); 
+47 
+48    switch (cmd) { 
+49    case IOCTL_VALSET: 
+50 
+51        /* 
+52        if (!capable(CAP_SYS_ADMIN)) { 
+53         retval = -EPERM; 
+54         goto done; 
+55        } 
+56        if (!access_ok(VERIFY_READ, (void __user *)arg, _IOC_SIZE(cmd))) { 
+57         retval = -EFAULT; 
+58         goto done; 
+59        } 
+60        */ 
+61        if (copy_from_user(&data, (int __user *) arg, sizeof(data))) { 
+62            retval = -EFAULT; 
+63            goto done; 
+64        } 
+65 
+66        pr_alert("IOCTL set val:%x .\n", data.val); 
+67        write_lock(&ioctl_data->lock); 
+68        ioctl_data->val = data.val; 
+69        write_unlock(&ioctl_data->lock); 
+70        break; 
+71 
+72    case IOCTL_VALGET: 
+73        /* 
+74        if (!access_ok(VERIFY_WRITE, (void __user *)arg, _IOC_SIZE(cmd))) { 
+75                                     retval = -EFAULT; 
+76                                     goto done; 
+77                             } 
+78        */ 
+79        read_lock(&ioctl_data->lock); 
+80        val = ioctl_data->val; 
+81        read_unlock(&ioctl_data->lock); 
+82        data.val = val; 
+83 
+84        if (copy_to_user((int __user *) arg, &data, sizeof(data))) { 
+85            retval = -EFAULT; 
+86            goto done; 
+87        } 
+88 
+89        break; 
+90 
+91    case IOCTL_VALGET_NUM: 
+92        retval = __put_user(ioctl_num, (int __user *) arg); 
+93        break; 
+94 
+95    case IOCTL_VALSET_NUM: 
+96        /* 
+97        if (!capable(CAP_SYS_ADMIN)) 
+98         return -EPERM; 
+99        */ 
+100        ioctl_num = arg; 
+101        break; 
+102 
+103    default: 
+104        retval = -ENOTTY; 
+105    } 
+106 
+107done: 
+108    return retval; 
+109} 
+110 
+111ssize_t test_ioctl_read(struct file *filp, 
+112                        char __user *buf, 
+113                        size_t count, 
+114                        loff_t *f_pos) 
+115{ 
+116    struct test_ioctl_data *ioctl_data = filp->private_data; 
+117    unsigned char val; 
+118    int retval; 
+119    int i = 0; 
+120    read_lock(&ioctl_data->lock); 
+121    val = ioctl_data->val; 
+122    read_unlock(&ioctl_data->lock); 
+123 
+124    for (; i < count; i++) { 
+125        if (copy_to_user(&buf[i], &val, 1)) { 
+126            retval = -EFAULT; 
+127            goto out; 
+128        } 
+129    } 
+130 
+131    retval = count; 
+132out: 
+133    return retval; 
+134} 
+135 
+136static int test_ioctl_close(struct inode *inode, struct file *filp) 
+137{ 
+138    pr_alert("%s call.\n", __func__); 
+139 
+140    if (filp->private_data) { 
+141        kfree(filp->private_data); 
+142        filp->private_data = NULL; 
+143    } 
+144 
+145    return 0; 
+146} 
+147 
+148static int test_ioctl_open(struct inode *inode, struct file *filp) 
+149{ 
+150    struct test_ioctl_data *ioctl_data; 
+151    pr_alert("%s call.\n", __func__); 
+152    ioctl_data = kmalloc(sizeof(struct test_ioctl_data), GFP_KERNEL); 
+153 
+154    if (ioctl_data == NULL) { 
+155        return -ENOMEM; 
+156    } 
+157 
+158    rwlock_init(&ioctl_data->lock); 
+159    ioctl_data->val = 0xFF; 
+160    filp->private_data = ioctl_data; 
+161    return 0; 
+162} 
+163 
+164struct file_operations fops = { 
+165    .owner = THIS_MODULE, 
+166    .open = test_ioctl_open, 
+167    .release = test_ioctl_close, 
+168    .read = test_ioctl_read, 
+169    .unlocked_ioctl = test_ioctl_ioctl, 
+170}; 
+171 
+172static int ioctl_init(void) 
+173{ 
+174    dev_t dev = MKDEV(test_ioctl_major, 0); 
+175    int alloc_ret = 0; 
+176    int cdev_ret = 0; 
+177    alloc_ret = alloc_chrdev_region(&dev, 0, num_of_dev, DRIVER_NAME); 
+178 
+179    if (alloc_ret) { 
+180        goto error; 
+181    } 
+182 
+183    test_ioctl_major = MAJOR(dev); 
+184    cdev_init(&test_ioctl_cdev, &fops); 
+185    cdev_ret = cdev_add(&test_ioctl_cdev, dev, num_of_dev); 
+186 
+187    if (cdev_ret) { 
+188        goto error; 
+189    } 
+190 
+191    pr_alert("%s driver(major: %d) installed.\n", DRIVER_NAME, 
+192             test_ioctl_major); 
+193    return 0; 
+194error: 
+195 
+196    if (cdev_ret == 0) { 
+197        cdev_del(&test_ioctl_cdev); 
+198    } 
+199 
+200    if (alloc_ret == 0) { 
+201        unregister_chrdev_region(dev, num_of_dev); 
+202    } 
+203 
+204    return -1; 
+205} 
+206 
+207static void ioctl_exit(void) 
+208{ 
+209    dev_t dev = MKDEV(test_ioctl_major, 0); 
+210    cdev_del(&test_ioctl_cdev); 
+211    unregister_chrdev_region(dev, num_of_dev); 
+212    pr_alert("%s driver removed.\n", DRIVER_NAME); 
+213} 
+214 
+215module_init(ioctl_init); 
+216module_exit(ioctl_exit); 
+217 
+218MODULE_LICENSE("GPL"); 
+219MODULE_DESCRIPTION("This is test_ioctl module");
    +

    -

    0.10 System Calls

    -

    So far, the only thing we’ve done was to use well defined kernel mechanisms to +

    0.10 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 are mostly on your own. -

    If you are not being sensible and using a virtual machine then this is where kernel +

    If you are not being sensible and using a virtual machine then this is where kernel programming can become hazardous. While writing the example below, I killed the open() system call. This meant I could not open any files, I could not run any programs, and I could not shutdown the system. I had @@ -2863,7 +2888,7 @@ was doing this on some live mission critical system then that could have been a possible outcome. To ensure you do not lose any files, even within a test environment, please run sync right before you do the insmod and the rmmod. -

    Forget about /proc files, forget about device files. They are just minor details. +

    Forget about /proc files, forget about device files. They are just minor details. Minutiae in the vast expanse of the universe. 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 @@ -2871,11 +2896,11 @@ 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 not +

    In general, a process is not supposed to be able to access the kernel. It can not access kernel memory and it can’t call kernel functions. The hardware of the CPU enforces this (that is the reason why it is called “protected mode” or “page protection”). -

    System calls are an exception to this general rule. What happens is that the +

    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, @@ -2883,7 +2908,7 @@ this is done by means of interrupt 0x80. The hardware knows that once you jump t 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 +

    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 @@ -2891,13 +2916,13 @@ function, and after it returns, does a few system checks and then return back to process (or to a different process, if the process time ran out). If you want to read this code, it is 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 +

    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 +

    The source code here is an example of such a kernel module. We want to “spy” on a certain user, and to pr_info() 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 @@ -2908,7 +2933,7 @@ file. -

    The init_module function replaces the appropriate location in sys_call_table +

    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 @@ -2918,7 +2943,7 @@ A_open, which will call the original sys_open when it is 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 +

    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 @@ -2932,7 +2957,7 @@ pointing to A_open, so it will not restore it to 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 +

    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 @@ -2945,191 +2970,191 @@ Depending on your kernel version, you might even need to hand apply the patch.

    - 
    1/* 
-2 *  syscall.c 
-3 * 
-4 *  System call "stealing" sample. 
-5 * 
-6 *  Disables page protection at a processor level by 
-7 *  changing the 16th bit in the cr0 register (could be Intel specific) 
-8 * 
-9 *  Based on example by Peter Jay Salzman and 
-10 *  https://bbs.archlinux.org/viewtopic.php?id=139406 
-11 */ 
-12 
-13#include <linux/delay.h> 
-14#include <linux/kernel.h> 
-15#include <linux/module.h> 
-16#include <linux/moduleparam.h> /* which will have params */ 
-17#include <linux/syscalls.h> 
-18#include <linux/unistd.h> /* The list of system calls */ 
-19 
-20/* 
-21 * For the current (process) structure, we need 
-22 * this to know who the current user is. 
-23 */ 
-24#include <linux/sched.h> 
-25#include <linux/uaccess.h> 
-26 
-27unsigned long **sys_call_table; 
-28unsigned long original_cr0; 
-29 
-30/* 
-31 * UID we want to spy on - will be filled from the 
-32 * command line 
-33 */ 
-34static int uid; 
-35module_param(uid, int, 0644); 
-36 
-37/* 
-38 * A pointer to the original system call. The reason 
-39 * we keep this, rather than call the original function 
-40 * (sys_open), is because somebody else might have 
-41 * replaced the system call before us. Note that this 
-42 * is not 100% safe, because if another module 
-43 * replaced sys_open before us, then when we're inserted 
-44 * we'll call the function in that module - and it 
-45 * might be removed before we are. 
-46 * 
-47 * Another reason for this is that we can't get sys_open. 
-48 * It's a static variable, so it is not exported. 
-49 */ 
-50asmlinkage int (*original_call)(const char *, int, int); 
-51 
-52/* 
-53 * The function we'll replace sys_open (the function 
-54 * called when you call the open system call) with. To 
-55 * find the exact prototype, with the number and type 
-56 * of arguments, we find the original function first 
-57 * (it's at fs/open.c). 
-58 * 
-59 * In theory, this means that we're tied to the 
-60 * current version of the kernel. In practice, the 
-61 * system calls almost never change (it would wreck havoc 
-62 * and require programs to be recompiled, since the system 
-63 * calls are the interface between the kernel and the 
-64 * processes). 
-65 */ 
-66asmlinkage int our_sys_open(const char *filename, int flags, int mode) 
-67{ 
-68    int i = 0; 
-69    char ch; 
-70 
-71    /* 
-72     * Report the file, if relevant 
-73     */ 
-74    pr_info("Opened file by %d: ", uid); 
-75    do { 
-76        get_user(ch, filename + i); 
-77        i++; 
-78        pr_info("%c", ch); 
-79    } while (ch != 0); 
-80    pr_info("\n"); 
-81 
-82    /* 
-83     * Call the original sys_open - otherwise, we lose 
-84     * the ability to open files 
-85     */ 
-86    return original_call(filename, flags, mode); 
-87} 
-88 
-89static unsigned long **aquire_sys_call_table(void) 
-90{ 
-91    unsigned long int offset = PAGE_OFFSET; 
-92    unsigned long **sct; 
-93 
-94    while (offset < ULLONG_MAX) { 
-95        sct = (unsigned long **) offset; 
-96 
-97        if (sct[__NR_close] == (unsigned long *) ksys_close) 
-98            return sct; 
-99 
-100        offset += sizeof(void *); 
-101    } 
-102 
-103    return NULL; 
-104} 
-105 
-106static int __init syscall_start(void) 
-107{ 
-108    if (!(sys_call_table = aquire_sys_call_table())) 
-109        return -1; 
-110 
-111    original_cr0 = read_cr0(); 
-112 
-113    write_cr0(original_cr0 & ~0x00010000); 
-114 
-115    /* keep track of the original open function */ 
-116    original_call = (void *) sys_call_table[__NR_open]; 
-117 
-118    /* use our open function instead */ 
-119    sys_call_table[__NR_open] = (unsigned long *) our_sys_open; 
-120 
-121    write_cr0(original_cr0); 
-122 
-123    pr_info("Spying on UID:%d\n", uid); 
-124 
-125    return 0; 
-126} 
-127 
-128static void __exit syscall_end(void) 
-129{ 
-130    if (!sys_call_table) { 
-131        return; 
-132    } 
-133 
-134    /* 
-135     * Return the system call back to normal 
-136     */ 
-137    if (sys_call_table[__NR_open] != (unsigned long *) our_sys_open) { 
-138        pr_alert("Somebody else also played with the "); 
-139        pr_alert("open system call\n"); 
-140        pr_alert("The system may be left in "); 
-141        pr_alert("an unstable state.\n"); 
-142    } 
-143 
-144    write_cr0(original_cr0 & ~0x00010000); 
-145    sys_call_table[__NR_open] = (unsigned long *) original_call; 
-146    write_cr0(original_cr0); 
-147 
-148    msleep(2000); 
-149} 
-150 
-151module_init(syscall_start); 
-152module_exit(syscall_end); 
-153 
-154MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  syscall.c 
+3 * 
+4 *  System call "stealing" sample. 
+5 * 
+6 *  Disables page protection at a processor level by 
+7 *  changing the 16th bit in the cr0 register (could be Intel specific) 
+8 * 
+9 *  Based on example by Peter Jay Salzman and 
+10 *  https://bbs.archlinux.org/viewtopic.php?id=139406 
+11 */ 
+12 
+13#include <linux/delay.h> 
+14#include <linux/kernel.h> 
+15#include <linux/module.h> 
+16#include <linux/moduleparam.h> /* which will have params */ 
+17#include <linux/syscalls.h> 
+18#include <linux/unistd.h> /* The list of system calls */ 
+19 
+20/* 
+21 * For the current (process) structure, we need 
+22 * this to know who the current user is. 
+23 */ 
+24#include <linux/sched.h> 
+25#include <linux/uaccess.h> 
+26 
+27unsigned long **sys_call_table; 
+28unsigned long original_cr0; 
+29 
+30/* 
+31 * UID we want to spy on - will be filled from the 
+32 * command line 
+33 */ 
+34static int uid; 
+35module_param(uid, int, 0644); 
+36 
+37/* 
+38 * A pointer to the original system call. The reason 
+39 * we keep this, rather than call the original function 
+40 * (sys_open), is because somebody else might have 
+41 * replaced the system call before us. Note that this 
+42 * is not 100% safe, because if another module 
+43 * replaced sys_open before us, then when we're inserted 
+44 * we'll call the function in that module - and it 
+45 * might be removed before we are. 
+46 * 
+47 * Another reason for this is that we can't get sys_open. 
+48 * It's a static variable, so it is not exported. 
+49 */ 
+50asmlinkage int (*original_call)(const char *, int, int); 
+51 
+52/* 
+53 * The function we'll replace sys_open (the function 
+54 * called when you call the open system call) with. To 
+55 * find the exact prototype, with the number and type 
+56 * of arguments, we find the original function first 
+57 * (it's at fs/open.c). 
+58 * 
+59 * In theory, this means that we're tied to the 
+60 * current version of the kernel. In practice, the 
+61 * system calls almost never change (it would wreck havoc 
+62 * and require programs to be recompiled, since the system 
+63 * calls are the interface between the kernel and the 
+64 * processes). 
+65 */ 
+66asmlinkage int our_sys_open(const char *filename, int flags, int mode) 
+67{ 
+68    int i = 0; 
+69    char ch; 
+70 
+71    /* 
+72     * Report the file, if relevant 
+73     */ 
+74    pr_info("Opened file by %d: ", uid); 
+75    do { 
+76        get_user(ch, filename + i); 
+77        i++; 
+78        pr_info("%c", ch); 
+79    } while (ch != 0); 
+80    pr_info("\n"); 
+81 
+82    /* 
+83     * Call the original sys_open - otherwise, we lose 
+84     * the ability to open files 
+85     */ 
+86    return original_call(filename, flags, mode); 
+87} 
+88 
+89static unsigned long **aquire_sys_call_table(void) 
+90{ 
+91    unsigned long int offset = PAGE_OFFSET; 
+92    unsigned long **sct; 
+93 
+94    while (offset < ULLONG_MAX) { 
+95        sct = (unsigned long **) offset; 
+96 
+97        if (sct[__NR_close] == (unsigned long *) ksys_close) 
+98            return sct; 
+99 
+100        offset += sizeof(void *); 
+101    } 
+102 
+103    return NULL; 
+104} 
+105 
+106static int __init syscall_start(void) 
+107{ 
+108    if (!(sys_call_table = aquire_sys_call_table())) 
+109        return -1; 
+110 
+111    original_cr0 = read_cr0(); 
+112 
+113    write_cr0(original_cr0 & ~0x00010000); 
+114 
+115    /* keep track of the original open function */ 
+116    original_call = (void *) sys_call_table[__NR_open]; 
+117 
+118    /* use our open function instead */ 
+119    sys_call_table[__NR_open] = (unsigned long *) our_sys_open; 
+120 
+121    write_cr0(original_cr0); 
+122 
+123    pr_info("Spying on UID:%d\n", uid); 
+124 
+125    return 0; 
+126} 
+127 
+128static void __exit syscall_end(void) 
+129{ 
+130    if (!sys_call_table) { 
+131        return; 
+132    } 
+133 
+134    /* 
+135     * Return the system call back to normal 
+136     */ 
+137    if (sys_call_table[__NR_open] != (unsigned long *) our_sys_open) { 
+138        pr_alert("Somebody else also played with the "); 
+139        pr_alert("open system call\n"); 
+140        pr_alert("The system may be left in "); 
+141        pr_alert("an unstable state.\n"); 
+142    } 
+143 
+144    write_cr0(original_cr0 & ~0x00010000); 
+145    sys_call_table[__NR_open] = (unsigned long *) original_call; 
+146    write_cr0(original_cr0); 
+147 
+148    msleep(2000); 
+149} 
+150 
+151module_init(syscall_start); 
+152module_exit(syscall_end); 
+153 
+154MODULE_LICENSE("GPL");
    +

    -

    0.11 Blocking Processes and threads

    +

    0.11 Blocking Processes and threads

    -

    +

    -

    0.11.1 Sleep

    -

    What do you do when somebody asks you for something you can’t do right away? If +

    0.11.1 Sleep

    +

    What do you do when somebody asks you for something you can’t do right away? If you’re a human being and you’re bothered by a human being, the only thing you can say is: "Not right now, I’m busy. Go away!". But if you’re a kernel module and you’re bothered by a process, you have another possibility. You can put the process to sleep until you can service it. After all, processes are being put to sleep by the kernel and woken up all the time (that’s the way multiple processes appear to run on the same time on a single CPU). -

    This kernel module is an example of this. The file (called /proc/sleep) can only +

    This kernel module is an example of this. The file (called /proc/sleep) can only be opened by a single process at a time. If the file is already open, the kernel module calls wait_event_interruptible. The easiest way to keep a file open is to open it with:

    - 
    1tail -f
    -

    This function changes the status of the task (a task is the kernel data + 

    1tail -f
    +

    This function changes the status of the task (a task is the kernel data structure which holds information about a process and the system call it’s in, if any) to TASK_INTERRUPTIBLE, which means that the task will not run until it is woken up somehow, and adds it to WaitQ, the queue of tasks waiting to access the file. Then, the function calls the scheduler to context switch to a different process, one which has some use for the CPU. -

    When a process is done with the file, it closes it, and module_close is called. That +

    When a process is done with the file, it closes it, and module_close is called. That function wakes up all the processes in the queue (there’s no mechanism to only wake up one of them). It then returns and the process which just closed the file can continue to run. In time, the scheduler decides that that @@ -3137,14 +3162,14 @@ process has had enough and gives control of the CPU to another process. Eventually, one of the processes which was in the queue will be given control of the CPU by the scheduler. It starts at the point right after the call to module_interruptible_sleep_on. -

    This means that the process is still in kernel mode - as far as the process is +

    This means that the process is still in kernel mode - as far as the process is concerned, it issued the open system call and the system call hasn’t returned yet. The process doesn’t know somebody else used the CPU for most of the time between the moment it issued the call and the moment it returned. -

    It can then proceed to set a global variable to tell all the other processes that the +

    It can then proceed to set a global variable to tell all the other processes that the file is still open and go on with its life. When the other processes get a piece of the CPU, they’ll see that global variable and go back to sleep. -

    So we’ll use tail -f to keep the file open in the background, while trying to +

    So we’ll use tail -f to keep the file open in the background, while trying to access it with another process (again in the background, so that we need not switch to a different vt). As soon as the first background process is killed with kill %1 , the second is woken up, is able to access the file and finally @@ -3152,15 +3177,15 @@ terminates. -

    To make our life more interesting, module_close doesn’t have a monopoly +

    To make our life more interesting, module_close doesn’t have a monopoly on waking up the processes which wait to access the file. A signal, such as Ctrl +c (SIGINT) can also wake up a process. This is because we used module_interruptible_sleep_on. We could have used module_sleep_on instead, but that would have resulted in extremely angry users whose Ctrl+c’s are ignored. -

    In that case, we want to return with -EINTR immediately. This is important so +

    In that case, we want to return with -EINTR immediately. This is important so users can, for example, kill the process before it receives the file. -

    There is one more point to remember. Some times processes don’t want to sleep, +

    There is one more point to remember. Some times processes don’t want to sleep, they want either to get what they want immediately, or to be told it cannot be done. Such processes use the O_NONBLOCK flag when opening the file. The kernel is supposed to respond by returning with the error code -EAGAIN from operations @@ -3193,510 +3218,510 @@ $ cat_nonblock /proc/sleep Last input: $ -

    +

    - 
    1/* 
-2 *  sleep.c - create a /proc file, and if several processes try to open it at 
-3 *  the same time, put all but one to sleep 
-4 */ 
-5 
-6#include <linux/kernel.h>  /* We're doing kernel work */ 
-7#include <linux/module.h>  /* Specifically, a module */ 
-8#include <linux/proc_fs.h> /* Necessary because we use proc fs */ 
-9#include <linux/sched.h>   /* For putting processes to sleep and 
-10                                   waking them up */  
-11#include <linux/uaccess.h> /* for get_user and put_user */ 
-12#include <linux/version.h> 
-13 
-14#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
-15#define HAVE_PROC_OPS 
-16#endif 
-17 
-18/* 
-19 * The module's file functions 
-20 */ 
-21 
-22/* 
-23 * Here we keep the last message received, to prove that we can process our 
-24 * input 
-25 */ 
-26#define MESSAGE_LENGTH 80 
-27static char Message[MESSAGE_LENGTH]; 
-28 
-29static struct proc_dir_entry *Our_Proc_File; 
-30#define PROC_ENTRY_FILENAME "sleep" 
-31 
-32/* 
-33 * Since we use the file operations struct, we can't use the special proc 
-34 * output provisions - we have to use a standard read function, which is this 
-35 * function 
-36 */ 
-37static ssize_t module_output(struct file *file, /* see include/linux/fs.h   */ 
-38                             char *buf,         /* The buffer to put data to 
-39                                                   (in the user segment)    */  
-40                             size_t len,        /* The length of the buffer */ 
-41                             loff_t *offset) 
-42{ 
-43    static int finished = 0; 
-44    int i; 
-45    char message[MESSAGE_LENGTH + 30]; 
-46 
-47    /* 
-48     * Return 0 to signify end of file - that we have nothing 
-49     * more to say at this point. 
-50     */ 
-51    if (finished) { 
-52        finished = 0; 
-53        return 0; 
-54    } 
-55 
-56    /* 
-57     * If you don't understand this by now, you're hopeless as a kernel 
-58     * programmer. 
-59     */ 
-60    sprintf(message, "Last input:%s\n", Message); 
-61    for (i = 0; i < len && message[i]; i++) 
-62        put_user(message[i], buf + i); 
-63 
-64    finished = 1; 
-65    return i; /* Return the number of bytes "read" */ 
-66} 
-67 
-68/* 
-69 * This function receives input from the user when the user writes to the /proc 
-70 * file. 
-71 */ 
-72static ssize_t module_input(struct file *file, /* The file itself */ 
-73                            const char *buf,   /* The buffer with input */ 
-74                            size_t length,     /* The buffer's length */ 
-75                            loff_t *offset)    /* offset to file - ignore */ 
-76{ 
-77    int i; 
-78 
-79    /* 
-80     * Put the input into Message, where module_output will later be 
-81     * able to use it 
-82     */ 
-83    for (i = 0; i < MESSAGE_LENGTH - 1 && i < length; i++) 
-84        get_user(Message[i], buf + i); 
-85    /* 
-86     * we want a standard, zero terminated string 
-87     */ 
-88    Message[i] = '\0'; 
-89 
-90    /* 
-91     * We need to return the number of input characters used 
-92     */ 
-93    return i; 
-94} 
-95 
-96/* 
-97 * 1 if the file is currently open by somebody 
-98 */ 
-99int Already_Open = 0; 
-100 
-101/* 
-102 * Queue of processes who want our file 
-103 */ 
-104DECLARE_WAIT_QUEUE_HEAD(WaitQ); 
-105/* 
-106 * Called when the /proc file is opened 
-107 */ 
-108static int module_open(struct inode *inode, struct file *file) 
-109{ 
-110    /* 
-111     * If the file's flags include O_NONBLOCK, it means the process doesn't 
-112     * want to wait for the file.  In this case, if the file is already 
-113     * open, we should fail with -EAGAIN, meaning "you'll have to try 
-114     * again", instead of blocking a process which would rather stay awake. 
-115     */ 
-116    if ((file->f_flags & O_NONBLOCK) && Already_Open) 
-117        return -EAGAIN; 
-118 
-119    /* 
-120     * This is the correct place for try_module_get(THIS_MODULE) because 
-121     * if a process is in the loop, which is within the kernel module, 
-122     * the kernel module must not be removed. 
-123     */ 
-124    try_module_get(THIS_MODULE); 
-125 
-126    /* 
-127     * If the file is already open, wait until it isn't 
-128     */ 
-129 
-130    while (Already_Open) { 
-131        int i, is_sig = 0; 
-132 
-133        /* 
-134         * This function puts the current process, including any system 
-135         * calls, such as us, to sleep.  Execution will be resumed right 
-136         * after the function call, either because somebody called 
-137         * wake_up(&WaitQ) (only module_close does that, when the file 
-138         * is closed) or when a signal, such as Ctrl-C, is sent 
-139         * to the process 
-140         */ 
-141        wait_event_interruptible(WaitQ, !Already_Open); 
-142 
-143        /* 
-144         * If we woke up because we got a signal we're not blocking, 
-145         * return -EINTR (fail the system call).  This allows processes 
-146         * to be killed or stopped. 
-147         */ 
-148 
-149        /* 
-150         * Emmanuel Papirakis: 
-151         * 
-152         * This is a little update to work with 2.2.*.  Signals now are 
-153         * contained in two words (64 bits) and are stored in a structure that 
-154         * contains an array of two unsigned longs.  We now have to make 2 
-155         * checks in our if. 
-156         * 
-157         * Ori Pomerantz: 
-158         * 
-159         * Nobody promised me they'll never use more than 64 bits, or that this 
-160         * book won't be used for a version of Linux with a word size of 16 
-161         * bits.  This code would work in any case. 
-162         */ 
-163        for (i = 0; i < _NSIG_WORDS && !is_sig; i++) 
-164            is_sig = current->pending.signal.sig[i] & ~current->blocked.sig[i]; 
-165 
-166        if (is_sig) { 
-167            /* 
-168             * It's important to put module_put(THIS_MODULE) here, 
-169             * because for processes where the open is interrupted 
-170             * there will never be a corresponding close. If we 
-171             * don't decrement the usage count here, we will be 
-172             * left with a positive usage count which we'll have no 
-173             * way to bring down to zero, giving us an immortal 
-174             * module, which can only be killed by rebooting 
-175             * the machine. 
-176             */ 
-177            module_put(THIS_MODULE); 
-178            return -EINTR; 
-179        } 
-180    } 
-181 
-182    /* 
-183     * If we got here, Already_Open must be zero 
-184     */ 
-185 
-186    /* 
-187     * Open the file 
-188     */ 
-189    Already_Open = 1; 
-190    return 0; /* Allow the access */ 
-191} 
-192 
-193/* 
-194 * Called when the /proc file is closed 
-195 */ 
-196int module_close(struct inode *inode, struct file *file) 
-197{ 
-198    /* 
-199     * Set Already_Open to zero, so one of the processes in the WaitQ will 
-200     * be able to set Already_Open back to one and to open the file. All 
-201     * the other processes will be called when Already_Open is back to one, 
-202     * so they'll go back to sleep. 
-203     */ 
-204    Already_Open = 0; 
-205 
-206    /* 
-207     * Wake up all the processes in WaitQ, so if anybody is waiting for the 
-208     * file, they can have it. 
-209     */ 
-210    wake_up(&WaitQ); 
-211 
-212    module_put(THIS_MODULE); 
-213 
-214    return 0; /* success */ 
-215} 
-216 
-217/* 
-218 * Structures to register as the /proc file, with pointers to all the relevant 
-219 * functions. 
-220 */ 
-221 
-222/* 
-223 * File operations for our proc file. This is where we place pointers to all 
-224 * the functions called when somebody tries to do something to our file. NULL 
-225 * means we don't want to deal with something. 
-226 */ 
-227#ifdef HAVE_PROC_OPS 
-228static const struct proc_ops File_Ops_4_Our_Proc_File = { 
-229    .proc_read = module_output,   /* "read" from the file */ 
-230    .proc_write = module_input,   /* "write" to the file */ 
-231    .proc_open = module_open,     /* called when the /proc file is opened */ 
-232    .proc_release = module_close, /* called when it's closed */ 
-233}; 
-234#else 
-235static const struct file_operations File_Ops_4_Our_Proc_File = { 
-236    .read = module_output, 
-237    .write = module_input, 
-238    .open = module_open, 
-239    .release = module_close, 
-240}; 
-241#endif 
-242 
-243/* 
-244 * Module initialization and cleanup 
-245 */ 
-246 
-247/* 
-248 * Initialize the module - register the proc file 
-249 */ 
-250 
-251int init_module() 
-252{ 
-253    Our_Proc_File = 
-254        proc_create(PROC_ENTRY_FILENAME, 0644, NULL, &File_Ops_4_Our_Proc_File); 
-255    if (Our_Proc_File == NULL) { 
-256        remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
-257        pr_debug("Error: Could not initialize /proc/%s\n", PROC_ENTRY_FILENAME); 
-258        return -ENOMEM; 
-259    } 
-260    proc_set_size(Our_Proc_File, 80); 
-261    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
-262 
-263    pr_info("/proc/test created\n"); 
-264 
-265    return 0; 
-266} 
-267 
-268/* 
-269 * Cleanup - unregister our file from /proc.  This could get dangerous if 
-270 * there are still processes waiting in WaitQ, because they are inside our 
-271 * open function, which will get unloaded. I'll explain how to avoid removal 
-272 * of a kernel module in such a case in chapter 10. 
-273 */ 
-274void cleanup_module() 
-275{ 
-276    remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
-277    pr_debug("/proc/%s removed\n", PROC_ENTRY_FILENAME); 
-278} 
-279 
-280MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  sleep.c - create a /proc file, and if several processes try to open it at 
+3 *  the same time, put all but one to sleep 
+4 */ 
+5 
+6#include <linux/kernel.h>  /* We're doing kernel work */ 
+7#include <linux/module.h>  /* Specifically, a module */ 
+8#include <linux/proc_fs.h> /* Necessary because we use proc fs */ 
+9#include <linux/sched.h>   /* For putting processes to sleep and 
+10                                   waking them up */  
+11#include <linux/uaccess.h> /* for get_user and put_user */ 
+12#include <linux/version.h> 
+13 
+14#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 6, 0) 
+15#define HAVE_PROC_OPS 
+16#endif 
+17 
+18/* 
+19 * The module's file functions 
+20 */ 
+21 
+22/* 
+23 * Here we keep the last message received, to prove that we can process our 
+24 * input 
+25 */ 
+26#define MESSAGE_LENGTH 80 
+27static char Message[MESSAGE_LENGTH]; 
+28 
+29static struct proc_dir_entry *Our_Proc_File; 
+30#define PROC_ENTRY_FILENAME "sleep" 
+31 
+32/* 
+33 * Since we use the file operations struct, we can't use the special proc 
+34 * output provisions - we have to use a standard read function, which is this 
+35 * function 
+36 */ 
+37static ssize_t module_output(struct file *file, /* see include/linux/fs.h   */ 
+38                             char *buf,         /* The buffer to put data to 
+39                                                   (in the user segment)    */  
+40                             size_t len,        /* The length of the buffer */ 
+41                             loff_t *offset) 
+42{ 
+43    static int finished = 0; 
+44    int i; 
+45    char message[MESSAGE_LENGTH + 30]; 
+46 
+47    /* 
+48     * Return 0 to signify end of file - that we have nothing 
+49     * more to say at this point. 
+50     */ 
+51    if (finished) { 
+52        finished = 0; 
+53        return 0; 
+54    } 
+55 
+56    /* 
+57     * If you don't understand this by now, you're hopeless as a kernel 
+58     * programmer. 
+59     */ 
+60    sprintf(message, "Last input:%s\n", Message); 
+61    for (i = 0; i < len && message[i]; i++) 
+62        put_user(message[i], buf + i); 
+63 
+64    finished = 1; 
+65    return i; /* Return the number of bytes "read" */ 
+66} 
+67 
+68/* 
+69 * This function receives input from the user when the user writes to the /proc 
+70 * file. 
+71 */ 
+72static ssize_t module_input(struct file *file, /* The file itself */ 
+73                            const char *buf,   /* The buffer with input */ 
+74                            size_t length,     /* The buffer's length */ 
+75                            loff_t *offset)    /* offset to file - ignore */ 
+76{ 
+77    int i; 
+78 
+79    /* 
+80     * Put the input into Message, where module_output will later be 
+81     * able to use it 
+82     */ 
+83    for (i = 0; i < MESSAGE_LENGTH - 1 && i < length; i++) 
+84        get_user(Message[i], buf + i); 
+85    /* 
+86     * we want a standard, zero terminated string 
+87     */ 
+88    Message[i] = '\0'; 
+89 
+90    /* 
+91     * We need to return the number of input characters used 
+92     */ 
+93    return i; 
+94} 
+95 
+96/* 
+97 * 1 if the file is currently open by somebody 
+98 */ 
+99int Already_Open = 0; 
+100 
+101/* 
+102 * Queue of processes who want our file 
+103 */ 
+104DECLARE_WAIT_QUEUE_HEAD(WaitQ); 
+105/* 
+106 * Called when the /proc file is opened 
+107 */ 
+108static int module_open(struct inode *inode, struct file *file) 
+109{ 
+110    /* 
+111     * If the file's flags include O_NONBLOCK, it means the process doesn't 
+112     * want to wait for the file.  In this case, if the file is already 
+113     * open, we should fail with -EAGAIN, meaning "you'll have to try 
+114     * again", instead of blocking a process which would rather stay awake. 
+115     */ 
+116    if ((file->f_flags & O_NONBLOCK) && Already_Open) 
+117        return -EAGAIN; 
+118 
+119    /* 
+120     * This is the correct place for try_module_get(THIS_MODULE) because 
+121     * if a process is in the loop, which is within the kernel module, 
+122     * the kernel module must not be removed. 
+123     */ 
+124    try_module_get(THIS_MODULE); 
+125 
+126    /* 
+127     * If the file is already open, wait until it isn't 
+128     */ 
+129 
+130    while (Already_Open) { 
+131        int i, is_sig = 0; 
+132 
+133        /* 
+134         * This function puts the current process, including any system 
+135         * calls, such as us, to sleep.  Execution will be resumed right 
+136         * after the function call, either because somebody called 
+137         * wake_up(&WaitQ) (only module_close does that, when the file 
+138         * is closed) or when a signal, such as Ctrl-C, is sent 
+139         * to the process 
+140         */ 
+141        wait_event_interruptible(WaitQ, !Already_Open); 
+142 
+143        /* 
+144         * If we woke up because we got a signal we're not blocking, 
+145         * return -EINTR (fail the system call).  This allows processes 
+146         * to be killed or stopped. 
+147         */ 
+148 
+149        /* 
+150         * Emmanuel Papirakis: 
+151         * 
+152         * This is a little update to work with 2.2.*.  Signals now are 
+153         * contained in two words (64 bits) and are stored in a structure that 
+154         * contains an array of two unsigned longs.  We now have to make 2 
+155         * checks in our if. 
+156         * 
+157         * Ori Pomerantz: 
+158         * 
+159         * Nobody promised me they'll never use more than 64 bits, or that this 
+160         * book won't be used for a version of Linux with a word size of 16 
+161         * bits.  This code would work in any case. 
+162         */ 
+163        for (i = 0; i < _NSIG_WORDS && !is_sig; i++) 
+164            is_sig = current->pending.signal.sig[i] & ~current->blocked.sig[i]; 
+165 
+166        if (is_sig) { 
+167            /* 
+168             * It's important to put module_put(THIS_MODULE) here, 
+169             * because for processes where the open is interrupted 
+170             * there will never be a corresponding close. If we 
+171             * don't decrement the usage count here, we will be 
+172             * left with a positive usage count which we'll have no 
+173             * way to bring down to zero, giving us an immortal 
+174             * module, which can only be killed by rebooting 
+175             * the machine. 
+176             */ 
+177            module_put(THIS_MODULE); 
+178            return -EINTR; 
+179        } 
+180    } 
+181 
+182    /* 
+183     * If we got here, Already_Open must be zero 
+184     */ 
+185 
+186    /* 
+187     * Open the file 
+188     */ 
+189    Already_Open = 1; 
+190    return 0; /* Allow the access */ 
+191} 
+192 
+193/* 
+194 * Called when the /proc file is closed 
+195 */ 
+196int module_close(struct inode *inode, struct file *file) 
+197{ 
+198    /* 
+199     * Set Already_Open to zero, so one of the processes in the WaitQ will 
+200     * be able to set Already_Open back to one and to open the file. All 
+201     * the other processes will be called when Already_Open is back to one, 
+202     * so they'll go back to sleep. 
+203     */ 
+204    Already_Open = 0; 
+205 
+206    /* 
+207     * Wake up all the processes in WaitQ, so if anybody is waiting for the 
+208     * file, they can have it. 
+209     */ 
+210    wake_up(&WaitQ); 
+211 
+212    module_put(THIS_MODULE); 
+213 
+214    return 0; /* success */ 
+215} 
+216 
+217/* 
+218 * Structures to register as the /proc file, with pointers to all the relevant 
+219 * functions. 
+220 */ 
+221 
+222/* 
+223 * File operations for our proc file. This is where we place pointers to all 
+224 * the functions called when somebody tries to do something to our file. NULL 
+225 * means we don't want to deal with something. 
+226 */ 
+227#ifdef HAVE_PROC_OPS 
+228static const struct proc_ops File_Ops_4_Our_Proc_File = { 
+229    .proc_read = module_output,   /* "read" from the file */ 
+230    .proc_write = module_input,   /* "write" to the file */ 
+231    .proc_open = module_open,     /* called when the /proc file is opened */ 
+232    .proc_release = module_close, /* called when it's closed */ 
+233}; 
+234#else 
+235static const struct file_operations File_Ops_4_Our_Proc_File = { 
+236    .read = module_output, 
+237    .write = module_input, 
+238    .open = module_open, 
+239    .release = module_close, 
+240}; 
+241#endif 
+242 
+243/* 
+244 * Module initialization and cleanup 
+245 */ 
+246 
+247/* 
+248 * Initialize the module - register the proc file 
+249 */ 
+250 
+251int init_module() 
+252{ 
+253    Our_Proc_File = 
+254        proc_create(PROC_ENTRY_FILENAME, 0644, NULL, &File_Ops_4_Our_Proc_File); 
+255    if (Our_Proc_File == NULL) { 
+256        remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
+257        pr_debug("Error: Could not initialize /proc/%s\n", PROC_ENTRY_FILENAME); 
+258        return -ENOMEM; 
+259    } 
+260    proc_set_size(Our_Proc_File, 80); 
+261    proc_set_user(Our_Proc_File, GLOBAL_ROOT_UID, GLOBAL_ROOT_GID); 
+262 
+263    pr_info("/proc/test created\n"); 
+264 
+265    return 0; 
+266} 
+267 
+268/* 
+269 * Cleanup - unregister our file from /proc.  This could get dangerous if 
+270 * there are still processes waiting in WaitQ, because they are inside our 
+271 * open function, which will get unloaded. I'll explain how to avoid removal 
+272 * of a kernel module in such a case in chapter 10. 
+273 */ 
+274void cleanup_module() 
+275{ 
+276    remove_proc_entry(PROC_ENTRY_FILENAME, NULL); 
+277    pr_debug("/proc/%s removed\n", PROC_ENTRY_FILENAME); 
+278} 
+279 
+280MODULE_LICENSE("GPL");

    - 
    1/* 
-2 *  cat_nonblock.c - open a file and display its contents, but exit rather than 
-3 *  wait for input. 
-4 */ 
-5#include <errno.h>  /* for errno */ 
-6#include <fcntl.h>  /* for open */ 
-7#include <stdio.h>  /* standard I/O */ 
-8#include <stdlib.h> /* for exit */ 
-9#include <unistd.h> /* for read */ 
-10 
-11#define MAX_BYTES 1024 * 4 
-12 
-13int main(int argc, char *argv[]) 
-14{ 
-15    int fd;                 /* The file descriptor for the file to read */ 
-16    size_t bytes;           /* The number of bytes read */ 
-17    char buffer[MAX_BYTES]; /* The buffer for the bytes */ 
-18 
-19    /* Usage */ 
-20    if (argc != 2) { 
-21        printf("Usage: %s <filename>\n", argv[0]); 
-22        puts("Reads the content of a file, but doesn't wait for input"); 
-23        exit(-1); 
-24    } 
-25 
-26    /* Open the file for reading in non blocking mode */ 
-27    fd = open(argv[1], O_RDONLY | O_NONBLOCK); 
-28 
-29    /* If open failed */ 
-30    if (fd == -1) { 
-31        puts(errno == EAGAIN ? "Open would block" : "Open failed"); 
-32        exit(-1); 
-33    } 
-34 
-35    /* Read the file and output its contents */ 
-36    do { 
-37        /* Read characters from the file */ 
-38        bytes = read(fd, buffer, MAX_BYTES); 
-39 
-40        /* If there's an error, report it and die */ 
-41        if (bytes == -1) { 
-42            if (errno = EAGAIN) 
-43                puts("Normally I'd block, but you told me not to"); 
-44            else 
-45                puts("Another read error"); 
-46            exit(-1); 
-47        } 
-48 
-49        /* Print the characters */ 
-50        if (bytes > 0) { 
-51            for (int i = 0; i < bytes; i++) 
-52                putchar(buffer[i]); 
-53        } 
-54 
-55        /* While there are no errors and the file isn't over */ 
-56    } while (bytes > 0); 
-57 
-58    return 0; 
-59}
    -

    + 

    1/* 
+2 *  cat_nonblock.c - open a file and display its contents, but exit rather than 
+3 *  wait for input. 
+4 */ 
+5#include <errno.h>  /* for errno */ 
+6#include <fcntl.h>  /* for open */ 
+7#include <stdio.h>  /* standard I/O */ 
+8#include <stdlib.h> /* for exit */ 
+9#include <unistd.h> /* for read */ 
+10 
+11#define MAX_BYTES 1024 * 4 
+12 
+13int main(int argc, char *argv[]) 
+14{ 
+15    int fd;                 /* The file descriptor for the file to read */ 
+16    size_t bytes;           /* The number of bytes read */ 
+17    char buffer[MAX_BYTES]; /* The buffer for the bytes */ 
+18 
+19    /* Usage */ 
+20    if (argc != 2) { 
+21        printf("Usage: %s <filename>\n", argv[0]); 
+22        puts("Reads the content of a file, but doesn't wait for input"); 
+23        exit(-1); 
+24    } 
+25 
+26    /* Open the file for reading in non blocking mode */ 
+27    fd = open(argv[1], O_RDONLY | O_NONBLOCK); 
+28 
+29    /* If open failed */ 
+30    if (fd == -1) { 
+31        puts(errno == EAGAIN ? "Open would block" : "Open failed"); 
+32        exit(-1); 
+33    } 
+34 
+35    /* Read the file and output its contents */ 
+36    do { 
+37        /* Read characters from the file */ 
+38        bytes = read(fd, buffer, MAX_BYTES); 
+39 
+40        /* If there's an error, report it and die */ 
+41        if (bytes == -1) { 
+42            if (errno = EAGAIN) 
+43                puts("Normally I'd block, but you told me not to"); 
+44            else 
+45                puts("Another read error"); 
+46            exit(-1); 
+47        } 
+48 
+49        /* Print the characters */ 
+50        if (bytes > 0) { 
+51            for (int i = 0; i < bytes; i++) 
+52                putchar(buffer[i]); 
+53        } 
+54 
+55        /* While there are no errors and the file isn't over */ 
+56    } while (bytes > 0); 
+57 
+58    return 0; 
+59}
    +

    -

    0.11.2 Completions

    -

    Sometimes one thing should happen before another within a module having multiple +

    0.11.2 Completions

    +

    Sometimes one thing should happen before another within a module having multiple threads. Rather than using /proc/sleep commands the kernel has another way to do this which allows timeouts or interrupts to also happen. -

    In the following example two threads are started, but one needs to start before +

    In the following example two threads are started, but one needs to start before another.

    - 
    1/* 
-2 *  completions.c 
-3 */ 
-4#include <linux/completion.h> 
-5#include <linux/init.h> 
-6#include <linux/kernel.h> 
-7#include <linux/kthread.h> 
-8#include <linux/module.h> 
-9 
-10static struct { 
-11    struct completion crank_comp; 
-12    struct completion flywheel_comp; 
-13} machine; 
-14 
-15static int machine_crank_thread(void *arg) 
-16{ 
-17    pr_info("Turn the crank\n"); 
-18 
-19    complete_all(&machine.crank_comp); 
-20    complete_and_exit(&machine.crank_comp, 0); 
-21} 
-22 
-23static int machine_flywheel_spinup_thread(void *arg) 
-24{ 
-25    wait_for_completion(&machine.crank_comp); 
-26 
-27    pr_info("Flywheel spins up\n"); 
-28 
-29    complete_all(&machine.flywheel_comp); 
-30    complete_and_exit(&machine.flywheel_comp, 0); 
-31} 
-32 
-33static int completions_init(void) 
-34{ 
-35    struct task_struct *crank_thread; 
-36    struct task_struct *flywheel_thread; 
-37 
-38    pr_info("completions example\n"); 
-39 
-40    init_completion(&machine.crank_comp); 
-41    init_completion(&machine.flywheel_comp); 
-42 
-43    crank_thread = kthread_create(machine_crank_thread, NULL, "KThread Crank"); 
-44    if (IS_ERR(crank_thread)) 
-45        goto ERROR_THREAD_1; 
-46 
-47    flywheel_thread = kthread_create(machine_flywheel_spinup_thread, NULL, 
-48                                     "KThread Flywheel"); 
-49    if (IS_ERR(flywheel_thread)) 
-50        goto ERROR_THREAD_2; 
-51 
-52    wake_up_process(flywheel_thread); 
-53    wake_up_process(crank_thread); 
-54 
-55    return 0; 
-56 
-57ERROR_THREAD_2: 
-58    kthread_stop(crank_thread); 
-59ERROR_THREAD_1: 
-60 
-61    return -1; 
-62} 
-63 
-64void completions_exit(void) 
-65{ 
-66    wait_for_completion(&machine.crank_comp); 
-67    wait_for_completion(&machine.flywheel_comp); 
-68 
-69    pr_info("completions exit\n"); 
-70} 
-71 
-72module_init(completions_init); 
-73module_exit(completions_exit); 
-74 
-75MODULE_DESCRIPTION("Completions example"); 
-76MODULE_LICENSE("GPL");
    -

    The machine structure stores the completion states for the two threads. At the + 

    1/* 
+2 *  completions.c 
+3 */ 
+4#include <linux/completion.h> 
+5#include <linux/init.h> 
+6#include <linux/kernel.h> 
+7#include <linux/kthread.h> 
+8#include <linux/module.h> 
+9 
+10static struct { 
+11    struct completion crank_comp; 
+12    struct completion flywheel_comp; 
+13} machine; 
+14 
+15static int machine_crank_thread(void *arg) 
+16{ 
+17    pr_info("Turn the crank\n"); 
+18 
+19    complete_all(&machine.crank_comp); 
+20    complete_and_exit(&machine.crank_comp, 0); 
+21} 
+22 
+23static int machine_flywheel_spinup_thread(void *arg) 
+24{ 
+25    wait_for_completion(&machine.crank_comp); 
+26 
+27    pr_info("Flywheel spins up\n"); 
+28 
+29    complete_all(&machine.flywheel_comp); 
+30    complete_and_exit(&machine.flywheel_comp, 0); 
+31} 
+32 
+33static int completions_init(void) 
+34{ 
+35    struct task_struct *crank_thread; 
+36    struct task_struct *flywheel_thread; 
+37 
+38    pr_info("completions example\n"); 
+39 
+40    init_completion(&machine.crank_comp); 
+41    init_completion(&machine.flywheel_comp); 
+42 
+43    crank_thread = kthread_create(machine_crank_thread, NULL, "KThread Crank"); 
+44    if (IS_ERR(crank_thread)) 
+45        goto ERROR_THREAD_1; 
+46 
+47    flywheel_thread = kthread_create(machine_flywheel_spinup_thread, NULL, 
+48                                     "KThread Flywheel"); 
+49    if (IS_ERR(flywheel_thread)) 
+50        goto ERROR_THREAD_2; 
+51 
+52    wake_up_process(flywheel_thread); 
+53    wake_up_process(crank_thread); 
+54 
+55    return 0; 
+56 
+57ERROR_THREAD_2: 
+58    kthread_stop(crank_thread); 
+59ERROR_THREAD_1: 
+60 
+61    return -1; 
+62} 
+63 
+64void completions_exit(void) 
+65{ 
+66    wait_for_completion(&machine.crank_comp); 
+67    wait_for_completion(&machine.flywheel_comp); 
+68 
+69    pr_info("completions exit\n"); 
+70} 
+71 
+72module_init(completions_init); 
+73module_exit(completions_exit); 
+74 
+75MODULE_DESCRIPTION("Completions example"); 
+76MODULE_LICENSE("GPL");
    +

    The machine structure stores the completion states for the two threads. At the exit point of each thread the respective completion state is updated, and wait_for_completion is used by the flywheel thread to ensure that it doesn’t begin prematurely. -

    So even though flywheel_thread is started first you should notice if you load this +

    So even though flywheel_thread is started first you should notice if you load this module and run dmesg that turning the crank always happens first because the flywheel thread waits for it to complete. -

    There are other variations upon the wait_for_completion function, which include +

    There are other variations upon the wait_for_completion function, which include timeouts or being interrupted, but this basic mechanism is enough for many common situations without adding a lot of complexity. -

    +

    -

    0.12 Avoiding Collisions and Deadlocks

    -

    If processes running on different CPUs or in different threads try to access the same +

    0.12 Avoiding Collisions and Deadlocks

    +

    If processes running on different CPUs or in different threads try to access the same memory then it’s possible that strange things can happen or your system can lock up. To avoid this various types of mutual exclusion kernel functions are available. These indicate if a section of code is "locked" or "unlocked" so that simultaneous attempts to run it can’t happen.

    -

    0.12.1 Mutex

    -

    You can use kernel mutexes (mutual exclusions) in much the same manner that you +

    0.12.1 Mutex

    +

    You can use kernel mutexes (mutual exclusions) in much the same manner that you might deploy them in userland. This may be all that’s needed to avoid collisions in most cases.

    - 
    1/* 
-2 *  example_mutex.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7#include <linux/mutex.h> 
-8 
-9DEFINE_MUTEX(mymutex); 
-10 
-11static int example_mutex_init(void) 
-12{ 
-13    int ret; 
-14 
-15    pr_info("example_mutex init\n"); 
-16 
-17    ret = mutex_trylock(&mymutex); 
-18    if (ret != 0) { 
-19        pr_info("mutex is locked\n"); 
-20 
-21        if (mutex_is_locked(&mymutex) == 0) 
-22            pr_info("The mutex failed to lock!\n"); 
-23 
-24        mutex_unlock(&mymutex); 
-25        pr_info("mutex is unlocked\n"); 
-26    } else 
-27        pr_info("Failed to lock\n"); 
-28 
-29    return 0; 
-30} 
-31 
-32static void example_mutex_exit(void) 
-33{ 
-34    pr_info("example_mutex exit\n"); 
-35} 
-36 
-37module_init(example_mutex_init); 
-38module_exit(example_mutex_exit); 
-39 
-40MODULE_DESCRIPTION("Mutex example"); 
-41MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  example_mutex.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7#include <linux/mutex.h> 
+8 
+9DEFINE_MUTEX(mymutex); 
+10 
+11static int example_mutex_init(void) 
+12{ 
+13    int ret; 
+14 
+15    pr_info("example_mutex init\n"); 
+16 
+17    ret = mutex_trylock(&mymutex); 
+18    if (ret != 0) { 
+19        pr_info("mutex is locked\n"); 
+20 
+21        if (mutex_is_locked(&mymutex) == 0) 
+22            pr_info("The mutex failed to lock!\n"); 
+23 
+24        mutex_unlock(&mymutex); 
+25        pr_info("mutex is unlocked\n"); 
+26    } else 
+27        pr_info("Failed to lock\n"); 
+28 
+29    return 0; 
+30} 
+31 
+32static void example_mutex_exit(void) 
+33{ 
+34    pr_info("example_mutex exit\n"); 
+35} 
+36 
+37module_init(example_mutex_init); 
+38module_exit(example_mutex_exit); 
+39 
+40MODULE_DESCRIPTION("Mutex example"); 
+41MODULE_LICENSE("GPL");
    +

    -

    0.12.2 Spinlocks

    -

    As the name suggests, spinlocks lock up the CPU that the code is running on, +

    0.12.2 Spinlocks

    +

    As the name suggests, spinlocks lock up the CPU that the code is running on, taking 100% of its resources. Because of this you should only use the spinlock mechanism around code which is likely to take no more than a few milliseconds to run and so won’t noticably slow anything down from the user’s point of @@ -3704,80 +3729,80 @@ view. -

    The example here is "irq safe" in that if interrupts happen during the lock then +

    The example here is "irq safe" in that if interrupts happen during the lock then they won’t be forgotten and will activate when the unlock happens, using the flags variable to retain their state.

    - 
    1/* 
-2 *  example_spinlock.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/interrupt.h> 
-6#include <linux/kernel.h> 
-7#include <linux/module.h> 
-8#include <linux/spinlock.h> 
-9 
-10DEFINE_SPINLOCK(sl_static); 
-11spinlock_t sl_dynamic; 
-12 
-13static void example_spinlock_static(void) 
-14{ 
-15    unsigned long flags; 
-16 
-17    spin_lock_irqsave(&sl_static, flags); 
-18    pr_info("Locked static spinlock\n"); 
-19 
-20    /* Do something or other safely. 
-21       Because this uses 100% CPU time this 
-22       code should take no more than a few 
-23       milliseconds to run */ 
-24 
-25    spin_unlock_irqrestore(&sl_static, flags); 
-26    pr_info("Unlocked static spinlock\n"); 
-27} 
-28 
-29static void example_spinlock_dynamic(void) 
-30{ 
-31    unsigned long flags; 
-32 
-33    spin_lock_init(&sl_dynamic); 
-34    spin_lock_irqsave(&sl_dynamic, flags); 
-35    pr_info("Locked dynamic spinlock\n"); 
-36 
-37    /* Do something or other safely. 
-38       Because this uses 100% CPU time this 
-39       code should take no more than a few 
-40       milliseconds to run */ 
-41 
-42    spin_unlock_irqrestore(&sl_dynamic, flags); 
-43    pr_info("Unlocked dynamic spinlock\n"); 
-44} 
-45 
-46static int example_spinlock_init(void) 
-47{ 
-48    pr_info("example spinlock started\n"); 
-49 
-50    example_spinlock_static(); 
-51    example_spinlock_dynamic(); 
-52 
-53    return 0; 
-54} 
-55 
-56static void example_spinlock_exit(void) 
-57{ 
-58    pr_info("example spinlock exit\n"); 
-59} 
-60 
-61module_init(example_spinlock_init); 
-62module_exit(example_spinlock_exit); 
-63 
-64MODULE_DESCRIPTION("Spinlock example"); 
-65MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  example_spinlock.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/interrupt.h> 
+6#include <linux/kernel.h> 
+7#include <linux/module.h> 
+8#include <linux/spinlock.h> 
+9 
+10DEFINE_SPINLOCK(sl_static); 
+11spinlock_t sl_dynamic; 
+12 
+13static void example_spinlock_static(void) 
+14{ 
+15    unsigned long flags; 
+16 
+17    spin_lock_irqsave(&sl_static, flags); 
+18    pr_info("Locked static spinlock\n"); 
+19 
+20    /* Do something or other safely. 
+21       Because this uses 100% CPU time this 
+22       code should take no more than a few 
+23       milliseconds to run */ 
+24 
+25    spin_unlock_irqrestore(&sl_static, flags); 
+26    pr_info("Unlocked static spinlock\n"); 
+27} 
+28 
+29static void example_spinlock_dynamic(void) 
+30{ 
+31    unsigned long flags; 
+32 
+33    spin_lock_init(&sl_dynamic); 
+34    spin_lock_irqsave(&sl_dynamic, flags); 
+35    pr_info("Locked dynamic spinlock\n"); 
+36 
+37    /* Do something or other safely. 
+38       Because this uses 100% CPU time this 
+39       code should take no more than a few 
+40       milliseconds to run */ 
+41 
+42    spin_unlock_irqrestore(&sl_dynamic, flags); 
+43    pr_info("Unlocked dynamic spinlock\n"); 
+44} 
+45 
+46static int example_spinlock_init(void) 
+47{ 
+48    pr_info("example spinlock started\n"); 
+49 
+50    example_spinlock_static(); 
+51    example_spinlock_dynamic(); 
+52 
+53    return 0; 
+54} 
+55 
+56static void example_spinlock_exit(void) 
+57{ 
+58    pr_info("example spinlock exit\n"); 
+59} 
+60 
+61module_init(example_spinlock_init); 
+62module_exit(example_spinlock_exit); 
+63 
+64MODULE_DESCRIPTION("Spinlock example"); 
+65MODULE_LICENSE("GPL");
    +

    -

    0.12.3 Read and write locks

    -

    Read and write locks are specialised kinds of spinlocks so that you can exclusively +

    0.12.3 Read and write locks

    +

    Read and write locks are specialised kinds of spinlocks so that you can exclusively read from something or write to something. Like the earlier spinlocks example the one below shows an "irq safe" situation in which if other functions were triggered from irqs which might also read and write to whatever you are concerned @@ -3787,68 +3812,68 @@ the system and cause users to start revolting against the tyranny of your module.

    - 
    1/* 
-2 *  example_rwlock.c 
-3 */ 
-4#include <linux/interrupt.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7 
-8DEFINE_RWLOCK(myrwlock); 
-9 
-10static void example_read_lock(void) 
-11{ 
-12    unsigned long flags; 
-13 
-14    read_lock_irqsave(&myrwlock, flags); 
-15    pr_info("Read Locked\n"); 
-16 
-17    /* Read from something */ 
-18 
-19    read_unlock_irqrestore(&myrwlock, flags); 
-20    pr_info("Read Unlocked\n"); 
-21} 
-22 
-23static void example_write_lock(void) 
-24{ 
-25    unsigned long flags; 
-26 
-27    write_lock_irqsave(&myrwlock, flags); 
-28    pr_info("Write Locked\n"); 
-29 
-30    /* Write to something */ 
-31 
-32    write_unlock_irqrestore(&myrwlock, flags); 
-33    pr_info("Write Unlocked\n"); 
-34} 
-35 
-36static int example_rwlock_init(void) 
-37{ 
-38    pr_info("example_rwlock started\n"); 
-39 
-40    example_read_lock(); 
-41    example_write_lock(); 
-42 
-43    return 0; 
-44} 
-45 
-46static void example_rwlock_exit(void) 
-47{ 
-48    pr_info("example_rwlock exit\n"); 
-49} 
-50 
-51module_init(example_rwlock_init); 
-52module_exit(example_rwlock_exit); 
-53 
-54MODULE_DESCRIPTION("Read/Write locks example"); 
-55MODULE_LICENSE("GPL");
    -

    Of course if you know for sure that there are no functions triggered by irqs + 

    1/* 
+2 *  example_rwlock.c 
+3 */ 
+4#include <linux/interrupt.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7 
+8DEFINE_RWLOCK(myrwlock); 
+9 
+10static void example_read_lock(void) 
+11{ 
+12    unsigned long flags; 
+13 
+14    read_lock_irqsave(&myrwlock, flags); 
+15    pr_info("Read Locked\n"); 
+16 
+17    /* Read from something */ 
+18 
+19    read_unlock_irqrestore(&myrwlock, flags); 
+20    pr_info("Read Unlocked\n"); 
+21} 
+22 
+23static void example_write_lock(void) 
+24{ 
+25    unsigned long flags; 
+26 
+27    write_lock_irqsave(&myrwlock, flags); 
+28    pr_info("Write Locked\n"); 
+29 
+30    /* Write to something */ 
+31 
+32    write_unlock_irqrestore(&myrwlock, flags); 
+33    pr_info("Write Unlocked\n"); 
+34} 
+35 
+36static int example_rwlock_init(void) 
+37{ 
+38    pr_info("example_rwlock started\n"); 
+39 
+40    example_read_lock(); 
+41    example_write_lock(); 
+42 
+43    return 0; 
+44} 
+45 
+46static void example_rwlock_exit(void) 
+47{ 
+48    pr_info("example_rwlock exit\n"); 
+49} 
+50 
+51module_init(example_rwlock_init); 
+52module_exit(example_rwlock_exit); 
+53 
+54MODULE_DESCRIPTION("Read/Write locks example"); 
+55MODULE_LICENSE("GPL");
    +

    Of course if you know for sure that there are no functions triggered by irqs which could possibly interfere with your logic then you can use the simpler read_lock(&myrwlock) and read_unlock(&myrwlock) or the corresponding write functions.

    -

    0.12.4 Atomic operations

    -

    If you’re doing simple arithmetic: adding, subtracting or bitwise operations then +

    0.12.4 Atomic operations

    +

    If you’re doing simple arithmetic: adding, subtracting or bitwise operations then there’s another way in the multi-CPU and multi-hyperthreaded world to stop other parts of the system from messing with your mojo. By using atomic operations you can be confident that your addition, subtraction or bit flip did actually happen @@ -3856,293 +3881,293 @@ and wasn’t overwritten by some other shenanigans. An example is shown below.

    - 
    1/* 
-2 *  example_atomic.c 
-3 */ 
-4#include <linux/interrupt.h> 
-5#include <linux/kernel.h> 
-6#include <linux/module.h> 
-7 
-8#define BYTE_TO_BINARY_PATTERN "%c%c%c%c%c%c%c%c" 
-9#define BYTE_TO_BINARY(byte)                                  \ 
-10    (byte & 0x80 ? '1' : '0'), (byte & 0x40 ? '1' : '0'),     \ 
-11        (byte & 0x20 ? '1' : '0'), (byte & 0x10 ? '1' : '0'), \ 
-12        (byte & 0x08 ? '1' : '0'), (byte & 0x04 ? '1' : '0'), \ 
-13        (byte & 0x02 ? '1' : '0'), (byte & 0x01 ? '1' : '0') 
-14 
-15static void atomic_add_subtract(void) 
-16{ 
-17    atomic_t debbie; 
-18    atomic_t chris = ATOMIC_INIT(50); 
-19 
-20    atomic_set(&debbie, 45); 
-21 
-22    /* subtract one */ 
-23    atomic_dec(&debbie); 
-24 
-25    atomic_add(7, &debbie); 
-26 
-27    /* add one */ 
-28    atomic_inc(&debbie); 
-29 
-30    pr_info("chris: %d, debbie: %d\n", atomic_read(&chris), 
-31            atomic_read(&debbie)); 
-32} 
-33 
-34static void atomic_bitwise(void) 
-35{ 
-36    unsigned long word = 0; 
-37 
-38    pr_info("Bits 0: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-39    set_bit(3, &word); 
-40    set_bit(5, &word); 
-41    pr_info("Bits 1: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-42    clear_bit(5, &word); 
-43    pr_info("Bits 2: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-44    change_bit(3, &word); 
-45 
-46    pr_info("Bits 3: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-47    if (test_and_set_bit(3, &word)) 
-48        pr_info("wrong\n"); 
-49    pr_info("Bits 4: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-50 
-51    word = 255; 
-52    pr_info("Bits 5: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
-53} 
-54 
-55static int example_atomic_init(void) 
-56{ 
-57    pr_info("example_atomic started\n"); 
-58 
-59    atomic_add_subtract(); 
-60    atomic_bitwise(); 
-61 
-62    return 0; 
-63} 
-64 
-65static void example_atomic_exit(void) 
-66{ 
-67    pr_info("example_atomic exit\n"); 
-68} 
-69 
-70module_init(example_atomic_init); 
-71module_exit(example_atomic_exit); 
-72 
-73MODULE_DESCRIPTION("Atomic operations example"); 
-74MODULE_LICENSE("GPL");
    + 
    1/* 
+2 *  example_atomic.c 
+3 */ 
+4#include <linux/interrupt.h> 
+5#include <linux/kernel.h> 
+6#include <linux/module.h> 
+7 
+8#define BYTE_TO_BINARY_PATTERN "%c%c%c%c%c%c%c%c" 
+9#define BYTE_TO_BINARY(byte)                                  \ 
+10    (byte & 0x80 ? '1' : '0'), (byte & 0x40 ? '1' : '0'),     \ 
+11        (byte & 0x20 ? '1' : '0'), (byte & 0x10 ? '1' : '0'), \ 
+12        (byte & 0x08 ? '1' : '0'), (byte & 0x04 ? '1' : '0'), \ 
+13        (byte & 0x02 ? '1' : '0'), (byte & 0x01 ? '1' : '0') 
+14 
+15static void atomic_add_subtract(void) 
+16{ 
+17    atomic_t debbie; 
+18    atomic_t chris = ATOMIC_INIT(50); 
+19 
+20    atomic_set(&debbie, 45); 
+21 
+22    /* subtract one */ 
+23    atomic_dec(&debbie); 
+24 
+25    atomic_add(7, &debbie); 
+26 
+27    /* add one */ 
+28    atomic_inc(&debbie); 
+29 
+30    pr_info("chris: %d, debbie: %d\n", atomic_read(&chris), 
+31            atomic_read(&debbie)); 
+32} 
+33 
+34static void atomic_bitwise(void) 
+35{ 
+36    unsigned long word = 0; 
+37 
+38    pr_info("Bits 0: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+39    set_bit(3, &word); 
+40    set_bit(5, &word); 
+41    pr_info("Bits 1: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+42    clear_bit(5, &word); 
+43    pr_info("Bits 2: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+44    change_bit(3, &word); 
+45 
+46    pr_info("Bits 3: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+47    if (test_and_set_bit(3, &word)) 
+48        pr_info("wrong\n"); 
+49    pr_info("Bits 4: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+50 
+51    word = 255; 
+52    pr_info("Bits 5: " BYTE_TO_BINARY_PATTERN, BYTE_TO_BINARY(word)); 
+53} 
+54 
+55static int example_atomic_init(void) 
+56{ 
+57    pr_info("example_atomic started\n"); 
+58 
+59    atomic_add_subtract(); 
+60    atomic_bitwise(); 
+61 
+62    return 0; 
+63} 
+64 
+65static void example_atomic_exit(void) 
+66{ 
+67    pr_info("example_atomic exit\n"); 
+68} 
+69 
+70module_init(example_atomic_init); 
+71module_exit(example_atomic_exit); 
+72 
+73MODULE_DESCRIPTION("Atomic operations example"); 
+74MODULE_LICENSE("GPL");
    -

    +

    -

    0.13 Replacing Print Macros

    -

    +

    0.13 Replacing Print Macros

    +

    -

    0.13.1 Replacement

    -

    In Section 1.2.1.2, I said that X and kernel module programming don’t mix. That’s +

    0.13.1 Replacement

    +

    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 the command to load the module came from. -

    "tty" is an abbreviation of teletype: originally a combination keyboard-printer +

    "tty" is an abbreviation of teletype: originally a combination keyboard-printer used to communicate with a Unix system, and today an abstraction for the text stream used for a Unix program, whether it’s a physical terminal, an xterm on an X display, a network connection used with ssh, etc. -

    The way this is done is by using current, a pointer to the currently running task, +

    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.

    - 
    1/* 
-2 *  print_string.c - Send output to the tty we're running on, regardless if it's 
-3 *  through X11, telnet, etc.  We do this by printing the string to the tty 
-4 *  associated with the current task. 
-5 */ 
-6#include <linux/init.h> 
-7#include <linux/kernel.h> 
-8#include <linux/module.h> 
-9#include <linux/sched.h> /* For current */ 
-10#include <linux/tty.h>   /* For the tty declarations */ 
-11 
-12MODULE_LICENSE("GPL"); 
-13 
-14static void print_string(char *str) 
-15{ 
-16    struct tty_struct *my_tty; 
-17    const struct tty_operations *ttyops; 
-18 
-19    /* 
-20     * The tty for the current task, for 2.6.6+ kernels 
-21     */ 
-22    my_tty = get_current_tty(); 
-23    ttyops = my_tty->driver->ops; 
-24 
-25    /* 
-26     * If my_tty is NULL, the current task has no tty you can print to 
-27     * (ie, if it's a daemon).  If so, there's nothing we can do. 
-28     */ 
-29    if (my_tty != NULL) { 
-30        /* 
-31         * my_tty->driver is a struct which holds the tty's functions, 
-32         * one of which (write) is used to write strings to the tty. 
-33         * It can be used to take a string either from the user's or 
-34         * kernel's memory segment. 
-35         * 
-36         * The function's 1st parameter is the tty to write to, 
-37         * because the same function would normally be used for all 
-38         * tty's of a certain type. 
-39         * The 2nd parameter is a pointer to a string. 
-40         * The 3rd parameter is the length of the string. 
-41         * 
-42         * As you will see below, sometimes it's necessary to use 
-43         * preprocessor stuff to create code that works for different 
-44         * kernel versions. The (naive) approach we've taken here 
-45         * does not scale well. The right way to deal with this 
-46         * is described in section 2 of 
-47         * linux/Documentation/SubmittingPatches 
-48         */ 
-49        (ttyops->write)(my_tty,       /* The tty itself */ 
-50                        str,          /* String                 */ 
-51                        strlen(str)); /* Length */ 
-52 
-53        /* 
-54         * ttys were originally hardware devices, which (usually) 
-55         * strictly followed the ASCII standard.  In ASCII, to move to 
-56         * a new line you need two characters, a carriage return and a 
-57         * line feed.  On Unix, the ASCII line feed is used for both 
-58         * purposes - so we can't just use \n, because it wouldn't have 
-59         * a carriage return and the next line will start at the 
-60         * column right after the line feed. 
-61         * 
-62         * This is why text files are different between Unix and 
-63         * MS Windows.  In CP/M and derivatives, like MS-DOS and 
-64         * MS Windows, the ASCII standard was strictly adhered to, 
-65         * and therefore a newline requirs both a LF and a CR. 
-66         */ 
-67        (ttyops->write)(my_tty, "\015\012", 2); 
-68    } 
-69} 
-70 
-71static int __init print_string_init(void) 
-72{ 
-73    print_string("The module has been inserted.  Hello world!"); 
-74    return 0; 
-75} 
-76 
-77static void __exit print_string_exit(void) 
-78{ 
-79    print_string("The module has been removed.  Farewell world!"); 
-80} 
-81 
-82module_init(print_string_init); 
-83module_exit(print_string_exit);
    -

    + 

    1/* 
+2 *  print_string.c - Send output to the tty we're running on, regardless if it's 
+3 *  through X11, telnet, etc.  We do this by printing the string to the tty 
+4 *  associated with the current task. 
+5 */ 
+6#include <linux/init.h> 
+7#include <linux/kernel.h> 
+8#include <linux/module.h> 
+9#include <linux/sched.h> /* For current */ 
+10#include <linux/tty.h>   /* For the tty declarations */ 
+11 
+12MODULE_LICENSE("GPL"); 
+13 
+14static void print_string(char *str) 
+15{ 
+16    struct tty_struct *my_tty; 
+17    const struct tty_operations *ttyops; 
+18 
+19    /* 
+20     * The tty for the current task, for 2.6.6+ kernels 
+21     */ 
+22    my_tty = get_current_tty(); 
+23    ttyops = my_tty->driver->ops; 
+24 
+25    /* 
+26     * If my_tty is NULL, the current task has no tty you can print to 
+27     * (ie, if it's a daemon).  If so, there's nothing we can do. 
+28     */ 
+29    if (my_tty != NULL) { 
+30        /* 
+31         * my_tty->driver is a struct which holds the tty's functions, 
+32         * one of which (write) is used to write strings to the tty. 
+33         * It can be used to take a string either from the user's or 
+34         * kernel's memory segment. 
+35         * 
+36         * The function's 1st parameter is the tty to write to, 
+37         * because the same function would normally be used for all 
+38         * tty's of a certain type. 
+39         * The 2nd parameter is a pointer to a string. 
+40         * The 3rd parameter is the length of the string. 
+41         * 
+42         * As you will see below, sometimes it's necessary to use 
+43         * preprocessor stuff to create code that works for different 
+44         * kernel versions. The (naive) approach we've taken here 
+45         * does not scale well. The right way to deal with this 
+46         * is described in section 2 of 
+47         * linux/Documentation/SubmittingPatches 
+48         */ 
+49        (ttyops->write)(my_tty,       /* The tty itself */ 
+50                        str,          /* String                 */ 
+51                        strlen(str)); /* Length */ 
+52 
+53        /* 
+54         * ttys were originally hardware devices, which (usually) 
+55         * strictly followed the ASCII standard.  In ASCII, to move to 
+56         * a new line you need two characters, a carriage return and a 
+57         * line feed.  On Unix, the ASCII line feed is used for both 
+58         * purposes - so we can't just use \n, because it wouldn't have 
+59         * a carriage return and the next line will start at the 
+60         * column right after the line feed. 
+61         * 
+62         * This is why text files are different between Unix and 
+63         * MS Windows.  In CP/M and derivatives, like MS-DOS and 
+64         * MS Windows, the ASCII standard was strictly adhered to, 
+65         * and therefore a newline requirs both a LF and a CR. 
+66         */ 
+67        (ttyops->write)(my_tty, "\015\012", 2); 
+68    } 
+69} 
+70 
+71static int __init print_string_init(void) 
+72{ 
+73    print_string("The module has been inserted.  Hello world!"); 
+74    return 0; 
+75} 
+76 
+77static void __exit print_string_exit(void) 
+78{ 
+79    print_string("The module has been removed.  Farewell world!"); 
+80} 
+81 
+82module_init(print_string_init); 
+83module_exit(print_string_exit);
    +

    -

    0.13.2 Flashing keyboard LEDs

    -

    In certain conditions, you may desire a simpler and more direct way to communicate +

    0.13.2 Flashing keyboard LEDs

    +

    In certain conditions, you may desire a simpler and more direct way to communicate to the external world. 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, compared to writing to a tty or a file. -

    The following source code illustrates a minimal kernel module which, when +

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

    - 
    1/* 
-2 *  kbleds.c - Blink keyboard leds until the module is unloaded. 
-3 */ 
-4 
-5#include <linux/init.h> 
-6#include <linux/kd.h> /* For KDSETLED */ 
-7#include <linux/module.h> 
-8#include <linux/tty.h> /* For fg_console, MAX_NR_CONSOLES */ 
-9#include <linux/vt.h> 
-10#include <linux/vt_kern.h> /* for fg_console */ 
-11 
-12#include <linux/console_struct.h> /* For vc_cons */ 
-13 
-14MODULE_DESCRIPTION("Example module illustrating the use of Keyboard LEDs."); 
-15MODULE_LICENSE("GPL"); 
-16 
-17struct timer_list my_timer; 
-18struct tty_driver *my_driver; 
-19char kbledstatus = 0; 
-20 
-21#define BLINK_DELAY HZ / 5 
-22#define ALL_LEDS_ON 0x07 
-23#define RESTORE_LEDS 0xFF 
-24 
-25/* 
-26 * Function my_timer_func blinks the keyboard LEDs periodically by invoking 
-27 * command KDSETLED of ioctl() on the keyboard driver. To learn more on virtual 
-28 * terminal ioctl operations, please see file: 
-29 *     /usr/src/linux/drivers/char/vt_ioctl.c, function vt_ioctl(). 
-30 * 
-31 * The argument to KDSETLED is alternatively set to 7 (thus causing the led 
-32 * mode to be set to LED_SHOW_IOCTL, and all the leds are lit) and to 0xFF 
-33 * (any value above 7 switches back the led mode to LED_SHOW_FLAGS, thus 
-34 * the LEDs reflect the actual keyboard status).  To learn more on this, 
-35 * please see file: 
-36 *     /usr/src/linux/drivers/char/keyboard.c, function setledstate(). 
-37 * 
-38 */ 
-39 
-40static void my_timer_func(unsigned long ptr) 
-41{ 
-42    unsigned long *pstatus = (unsigned long *) ptr; 
-43    struct tty_struct *t = vc_cons[fg_console].d->port.tty; 
-44 
-45    if (*pstatus == ALL_LEDS_ON) 
-46        *pstatus = RESTORE_LEDS; 
-47    else 
-48        *pstatus = ALL_LEDS_ON; 
-49 
-50    (my_driver->ops->ioctl)(t, KDSETLED, *pstatus); 
-51 
-52    my_timer.expires = jiffies + BLINK_DELAY; 
-53    add_timer(&my_timer); 
-54} 
-55 
-56static int __init kbleds_init(void) 
-57{ 
-58    int i; 
-59 
-60    pr_info("kbleds: loading\n"); 
-61    pr_info("kbleds: fgconsole is %x\n", fg_console); 
-62    for (i = 0; i < MAX_NR_CONSOLES; i++) { 
-63        if (!vc_cons[i].d) 
-64            break; 
-65        pr_info("poet_atkm: console[%i/%i] #%i, tty %lx\n", i, MAX_NR_CONSOLES, 
-66                vc_cons[i].d->vc_num, (unsigned long) vc_cons[i].d->port.tty); 
-67    } 
-68    pr_info("kbleds: finished scanning consoles\n"); 
-69 
-70    my_driver = vc_cons[fg_console].d->port.tty->driver; 
-71    pr_info("kbleds: tty driver magic %x\n", my_driver->magic); 
-72 
-73    /* 
-74     * Set up the LED blink timer the first time 
-75     */ 
-76    timer_setup(&my_timer, (void *) &my_timer_func, 
-77                (unsigned long) &kbledstatus); 
-78    my_timer.expires = jiffies + BLINK_DELAY; 
-79    add_timer(&my_timer); 
-80 
-81    return 0; 
-82} 
-83 
-84static void __exit kbleds_cleanup(void) 
-85{ 
-86    pr_info("kbleds: unloading...\n"); 
-87    del_timer(&my_timer); 
-88    (my_driver->ops->ioctl)(vc_cons[fg_console].d->port.tty, KDSETLED, 
-89                            RESTORE_LEDS); 
-90} 
-91 
-92module_init(kbleds_init); 
-93module_exit(kbleds_cleanup);
    -

    If none of the examples in this chapter fit your debugging needs there might yet + 

    1/* 
+2 *  kbleds.c - Blink keyboard leds until the module is unloaded. 
+3 */ 
+4 
+5#include <linux/init.h> 
+6#include <linux/kd.h> /* For KDSETLED */ 
+7#include <linux/module.h> 
+8#include <linux/tty.h> /* For fg_console, MAX_NR_CONSOLES */ 
+9#include <linux/vt.h> 
+10#include <linux/vt_kern.h> /* for fg_console */ 
+11 
+12#include <linux/console_struct.h> /* For vc_cons */ 
+13 
+14MODULE_DESCRIPTION("Example module illustrating the use of Keyboard LEDs."); 
+15MODULE_LICENSE("GPL"); 
+16 
+17struct timer_list my_timer; 
+18struct tty_driver *my_driver; 
+19char kbledstatus = 0; 
+20 
+21#define BLINK_DELAY HZ / 5 
+22#define ALL_LEDS_ON 0x07 
+23#define RESTORE_LEDS 0xFF 
+24 
+25/* 
+26 * Function my_timer_func blinks the keyboard LEDs periodically by invoking 
+27 * command KDSETLED of ioctl() on the keyboard driver. To learn more on virtual 
+28 * terminal ioctl operations, please see file: 
+29 *     /usr/src/linux/drivers/char/vt_ioctl.c, function vt_ioctl(). 
+30 * 
+31 * The argument to KDSETLED is alternatively set to 7 (thus causing the led 
+32 * mode to be set to LED_SHOW_IOCTL, and all the leds are lit) and to 0xFF 
+33 * (any value above 7 switches back the led mode to LED_SHOW_FLAGS, thus 
+34 * the LEDs reflect the actual keyboard status).  To learn more on this, 
+35 * please see file: 
+36 *     /usr/src/linux/drivers/char/keyboard.c, function setledstate(). 
+37 * 
+38 */ 
+39 
+40static void my_timer_func(unsigned long ptr) 
+41{ 
+42    unsigned long *pstatus = (unsigned long *) ptr; 
+43    struct tty_struct *t = vc_cons[fg_console].d->port.tty; 
+44 
+45    if (*pstatus == ALL_LEDS_ON) 
+46        *pstatus = RESTORE_LEDS; 
+47    else 
+48        *pstatus = ALL_LEDS_ON; 
+49 
+50    (my_driver->ops->ioctl)(t, KDSETLED, *pstatus); 
+51 
+52    my_timer.expires = jiffies + BLINK_DELAY; 
+53    add_timer(&my_timer); 
+54} 
+55 
+56static int __init kbleds_init(void) 
+57{ 
+58    int i; 
+59 
+60    pr_info("kbleds: loading\n"); 
+61    pr_info("kbleds: fgconsole is %x\n", fg_console); 
+62    for (i = 0; i < MAX_NR_CONSOLES; i++) { 
+63        if (!vc_cons[i].d) 
+64            break; 
+65        pr_info("poet_atkm: console[%i/%i] #%i, tty %lx\n", i, MAX_NR_CONSOLES, 
+66                vc_cons[i].d->vc_num, (unsigned long) vc_cons[i].d->port.tty); 
+67    } 
+68    pr_info("kbleds: finished scanning consoles\n"); 
+69 
+70    my_driver = vc_cons[fg_console].d->port.tty->driver; 
+71    pr_info("kbleds: tty driver magic %x\n", my_driver->magic); 
+72 
+73    /* 
+74     * Set up the LED blink timer the first time 
+75     */ 
+76    timer_setup(&my_timer, (void *) &my_timer_func, 
+77                (unsigned long) &kbledstatus); 
+78    my_timer.expires = jiffies + BLINK_DELAY; 
+79    add_timer(&my_timer); 
+80 
+81    return 0; 
+82} 
+83 
+84static void __exit kbleds_cleanup(void) 
+85{ 
+86    pr_info("kbleds: unloading...\n"); 
+87    del_timer(&my_timer); 
+88    (my_driver->ops->ioctl)(vc_cons[fg_console].d->port.tty, KDSETLED, 
+89                            RESTORE_LEDS); 
+90} 
+91 
+92module_init(kbleds_init); 
+93module_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 @@ -4155,64 +4180,64 @@ over a serial line. If you find yourself porting the kernel to some new and former unsupported architecture this is usually amongst 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 +

    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 sure it does not show up in production code. -

    +

    -

    0.14 Scheduling Tasks

    -

    There are two main ways of running tasks: tasklets and work queues. Tasklets are a +

    0.14 Scheduling Tasks

    +

    There are two main ways of running tasks: tasklets and work queues. Tasklets are a quick and easy way of scheduling a single function to be run, for example when triggered from an interrupt, whereas work queues are more complicated but also better suited to running multiple things in a sequence. -

    +

    -

    0.14.1 Tasklets

    -

    Here’s an example tasklet module. The tasklet_fn function runs for a few seconds +

    0.14.1 Tasklets

    +

    Here’s an example tasklet module. The tasklet_fn function runs for a few seconds and in the mean time execution of the example_tasklet_init function continues to the exit point.

    - 
    1/* 
-2 *  example_tasklet.c 
-3 */ 
-4#include <linux/delay.h> 
-5#include <linux/interrupt.h> 
-6#include <linux/kernel.h> 
-7#include <linux/module.h> 
-8 
-9static void tasklet_fn(unsigned long data) 
-10{ 
-11    pr_info("Example tasklet starts\n"); 
-12    mdelay(5000); 
-13    pr_info("Example tasklet ends\n"); 
-14} 
-15 
-16DECLARE_TASKLET(mytask, tasklet_fn, 0L); 
-17 
-18static int example_tasklet_init(void) 
-19{ 
-20    pr_info("tasklet example init\n"); 
-21    tasklet_schedule(&mytask); 
-22    mdelay(200); 
-23    pr_info("Example tasklet init continues...\n"); 
-24    return 0; 
-25} 
-26 
-27static void example_tasklet_exit(void) 
-28{ 
-29    pr_info("tasklet example exit\n"); 
-30    tasklet_kill(&mytask); 
-31} 
-32 
-33module_init(example_tasklet_init); 
-34module_exit(example_tasklet_exit); 
-35 
-36MODULE_DESCRIPTION("Tasklet example"); 
-37MODULE_LICENSE("GPL");
    -

    So with this example loaded dmesg should show: + 

    1/* 
+2 *  example_tasklet.c 
+3 */ 
+4#include <linux/delay.h> 
+5#include <linux/interrupt.h> 
+6#include <linux/kernel.h> 
+7#include <linux/module.h> 
+8 
+9static void tasklet_fn(unsigned long data) 
+10{ 
+11    pr_info("Example tasklet starts\n"); 
+12    mdelay(5000); 
+13    pr_info("Example tasklet ends\n"); 
+14} 
+15 
+16DECLARE_TASKLET(mytask, tasklet_fn, 0L); 
+17 
+18static int example_tasklet_init(void) 
+19{ 
+20    pr_info("tasklet example init\n"); 
+21    tasklet_schedule(&mytask); 
+22    mdelay(200); 
+23    pr_info("Example tasklet init continues...\n"); 
+24    return 0; 
+25} 
+26 
+27static void example_tasklet_exit(void) 
+28{ 
+29    pr_info("tasklet example exit\n"); 
+30    tasklet_kill(&mytask); 
+31} 
+32 
+33module_init(example_tasklet_init); 
+34module_exit(example_tasklet_exit); 
+35 
+36MODULE_DESCRIPTION("Tasklet example"); 
+37MODULE_LICENSE("GPL");
    +

    So with this example loaded dmesg should show: @@ -4223,57 +4248,57 @@ Example tasklet starts Example tasklet init continues... Example tasklet ends -

    -

    +

    +

    -

    0.14.2 Work queues

    -

    To add a task to the scheduler we can use a workqueue. The kernel then uses the +

    0.14.2 Work queues

    +

    To add a task to the scheduler we can use a workqueue. The kernel then uses the Completely Fair Scheduler (CFS) to execute work within the queue.

    - 
    1/* 
-2 *  sched.c 
-3 */ 
-4#include <linux/init.h> 
-5#include <linux/module.h> 
-6#include <linux/workqueue.h> 
-7 
-8static struct workqueue_struct *queue = NULL; 
-9static struct work_struct work; 
-10 
-11static void work_handler(struct work_struct *data) 
-12{ 
-13    pr_info("work handler function.\n"); 
-14} 
-15 
-16int init_module() 
-17{ 
-18    queue = alloc_workqueue("HELLOWORLD", WQ_UNBOUND, 1); 
-19    INIT_WORK(&work, work_handler); 
-20    schedule_work(&work); 
-21 
-22    return 0; 
-23} 
-24 
-25void cleanup_module() 
-26{ 
-27    destroy_workqueue(queue); 
-28} 
-29 
-30MODULE_LICENSE("GPL"); 
-31MODULE_DESCRIPTION("Workqueue example");
    -

    + 

    1/* 
+2 *  sched.c 
+3 */ 
+4#include <linux/init.h> 
+5#include <linux/module.h> 
+6#include <linux/workqueue.h> 
+7 
+8static struct workqueue_struct *queue = NULL; 
+9static struct work_struct work; 
+10 
+11static void work_handler(struct work_struct *data) 
+12{ 
+13    pr_info("work handler function.\n"); 
+14} 
+15 
+16int init_module() 
+17{ 
+18    queue = alloc_workqueue("HELLOWORLD", WQ_UNBOUND, 1); 
+19    INIT_WORK(&work, work_handler); 
+20    schedule_work(&work); 
+21 
+22    return 0; 
+23} 
+24 
+25void cleanup_module() 
+26{ 
+27    destroy_workqueue(queue); 
+28} 
+29 
+30MODULE_LICENSE("GPL"); 
+31MODULE_DESCRIPTION("Workqueue example");
    +

    -

    0.15 Interrupt Handlers

    -

    +

    0.15 Interrupt Handlers

    +

    -

    0.15.1 Interrupt Handlers

    -

    Except for the last chapter, everything we did in the kernel so far we’ve done as a +

    0.15.1 Interrupt Handlers

    +

    Except for the last chapter, everything we did in the kernel so far we’ve done as a response to a process asking for it, either by dealing with a special file, sending an ioctl(), or issuing a system call. But the job of the kernel isn’t just to respond to process requests. Another job, which is every bit as important, is to speak to the hardware connected to the machine. -

    There are two types of interaction between the CPU and the rest of the +

    There are two types of interaction between the CPU and the rest of the computer’s hardware. The first type is when the CPU gives orders to the hardware, the other is when the hardware needs to tell the CPU something. The second, called interrupts, is much harder to implement because it has to be dealt with when @@ -4283,14 +4308,14 @@ lost. -

    Under Linux, hardware interrupts are called IRQ’s (Interrupt ReQuests). There +

    Under Linux, hardware interrupts are called IRQ’s (Interrupt ReQuests). There are two types of IRQ’s, short and long. A short IRQ is one which is expected to take a very short period of time, during which the rest of the machine will be blocked and no other interrupts will be handled. A long IRQ is one which can take longer, and during which other interrupts may occur (but not interrupts from the same device). If at all possible, it’s better to declare an interrupt handler to be long. -

    When the CPU receives an interrupt, it stops whatever it’s doing (unless it’s +

    When the CPU receives an interrupt, it stops whatever it’s doing (unless it’s processing a more important interrupt, in which case it will deal with this one only when the more important one is done), saves certain parameters on the stack and calls the interrupt handler. This means that certain things @@ -4302,9 +4327,9 @@ the new information at a later time (this is called the "bottom half") and return. The kernel is then guaranteed to call the bottom half as soon as possible – and when it does, everything allowed in kernel modules will be allowed. -

    The way to implement this is to call request_irq() to get your interrupt handler +

    The way to implement this is to call request_irq() to get your interrupt handler called when the relevant IRQ is received. -

    In practice IRQ handling can be a bit more complex. Hardware is often +

    In practice IRQ handling can be a bit more complex. Hardware is often designed in a way that chains two interrupt controllers, so that all the IRQs from interrupt controller B are cascaded to a certain IRQ from interrupt controller A. Of course that requires that the kernel finds out which IRQ it @@ -4318,7 +4343,7 @@ another truckload of problems. It’s not enough to know if a certain IRQs has happend, it’s also important for what CPU(s) it was for. People still interested in more details, might want to do a web search for "APIC" now ;) -

    This function receives the IRQ number, the name of the function, flags, a name +

    This function receives the IRQ number, the name of the function, flags, a name for /proc/interrupts and a parameter to pass to the interrupt handler. Usually there is a certain number of IRQs available. How many IRQs there are is hardware-dependent. The flags can include SA_SHIRQ to indicate you’re willing to @@ -4329,333 +4354,333 @@ or if you’re both willing to share. -

    +

    -

    0.15.2 Detecting button presses

    -

    Many popular single board computers, such as Raspberry Pis or Beagleboards, have +

    0.15.2 Detecting button presses

    +

    Many popular single board computers, such as Raspberry Pis or Beagleboards, have a bunch of GPIO pins. Attaching buttons to those and then having a button press do something is a classic case in which you might need to use interrupts so that instead of having the CPU waste time and battery power polling for a change in input state it’s better for the input to trigger the CPU to then run a particular handling function. -

    Here’s an example where buttons are connected to GPIO numbers 17 and 18 and +

    Here’s an example where buttons are connected to GPIO numbers 17 and 18 and an LED is connected to GPIO 4. You can change those numbers to whatever is appropriate for your board.

    - 
    1/* 
-2 *  intrpt.c - Handling GPIO with interrupts 
-3 * 
-4 *  Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
-5 *  from: 
-6 *    https://github.com/wendlers/rpi-kmod-samples 
-7 * 
-8 *  Press one button to turn on a LED and another to turn it off 
-9 */ 
-10 
-11#include <linux/gpio.h> 
-12#include <linux/interrupt.h> 
-13#include <linux/kernel.h> 
-14#include <linux/module.h> 
-15 
-16static int button_irqs[] = {-1, -1}; 
-17 
-18/* Define GPIOs for LEDs. 
-19   Change the numbers for the GPIO on your board. */ 
-20static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
-21 
-22/* Define GPIOs for BUTTONS 
-23   Change the numbers for the GPIO on your board. */ 
-24static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
-25                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
-26 
-27/* 
-28 * interrupt function triggered when a button is pressed 
-29 */ 
-30static irqreturn_t button_isr(int irq, void *data) 
-31{ 
-32    /* first button */ 
-33    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
-34        gpio_set_value(leds[0].gpio, 1); 
-35    /* second button */ 
-36    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
-37        gpio_set_value(leds[0].gpio, 0); 
-38 
-39    return IRQ_HANDLED; 
-40} 
-41 
-42int init_module() 
-43{ 
-44    int ret = 0; 
-45 
-46    pr_info("%s\n", __func__); 
-47 
-48    /* register LED gpios */ 
-49    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
-50 
-51    if (ret) { 
-52        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
-53        return ret; 
-54    } 
-55 
-56    /* register BUTTON gpios */ 
-57    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
-58 
-59    if (ret) { 
-60        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
-61        goto fail1; 
-62    } 
-63 
-64    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
-65 
-66    ret = gpio_to_irq(buttons[0].gpio); 
-67 
-68    if (ret < 0) { 
-69        pr_err("Unable to request IRQ: %d\n", ret); 
-70        goto fail2; 
-71    } 
-72 
-73    button_irqs[0] = ret; 
-74 
-75    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
-76 
-77    ret = request_irq(button_irqs[0], button_isr, 
-78                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-79                      "gpiomod#button1", NULL); 
-80 
-81    if (ret) { 
-82        pr_err("Unable to request IRQ: %d\n", ret); 
-83        goto fail2; 
-84    } 
-85 
-86 
-87    ret = gpio_to_irq(buttons[1].gpio); 
-88 
-89    if (ret < 0) { 
-90        pr_err("Unable to request IRQ: %d\n", ret); 
-91        goto fail2; 
-92    } 
-93 
-94    button_irqs[1] = ret; 
-95 
-96    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
-97 
-98    ret = request_irq(button_irqs[1], button_isr, 
-99                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-100                      "gpiomod#button2", NULL); 
-101 
-102    if (ret) { 
-103        pr_err("Unable to request IRQ: %d\n", ret); 
-104        goto fail3; 
-105    } 
-106 
-107    return 0; 
-108 
-109/* cleanup what has been setup so far */ 
-110fail3: 
-111    free_irq(button_irqs[0], NULL); 
-112 
-113fail2: 
-114    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
-115 
-116fail1: 
-117    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-118 
-119    return ret; 
-120} 
-121 
-122void cleanup_module() 
-123{ 
-124    int i; 
-125 
-126    pr_info("%s\n", __func__); 
-127 
-128    /* free irqs */ 
-129    free_irq(button_irqs[0], NULL); 
-130    free_irq(button_irqs[1], NULL); 
-131 
-132    /* turn all LEDs off */ 
-133    for (i = 0; i < ARRAY_SIZE(leds); i++) 
-134        gpio_set_value(leds[i].gpio, 0); 
-135 
-136    /* unregister */ 
-137    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-138    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
-139} 
-140 
-141MODULE_LICENSE("GPL"); 
-142MODULE_DESCRIPTION("Handle some GPIO interrupts");
    -

    + 

    1/* 
+2 *  intrpt.c - Handling GPIO with interrupts 
+3 * 
+4 *  Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
+5 *  from: 
+6 *    https://github.com/wendlers/rpi-kmod-samples 
+7 * 
+8 *  Press one button to turn on a LED and another to turn it off 
+9 */ 
+10 
+11#include <linux/gpio.h> 
+12#include <linux/interrupt.h> 
+13#include <linux/kernel.h> 
+14#include <linux/module.h> 
+15 
+16static int button_irqs[] = {-1, -1}; 
+17 
+18/* Define GPIOs for LEDs. 
+19   Change the numbers for the GPIO on your board. */ 
+20static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
+21 
+22/* Define GPIOs for BUTTONS 
+23   Change the numbers for the GPIO on your board. */ 
+24static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
+25                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
+26 
+27/* 
+28 * interrupt function triggered when a button is pressed 
+29 */ 
+30static irqreturn_t button_isr(int irq, void *data) 
+31{ 
+32    /* first button */ 
+33    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
+34        gpio_set_value(leds[0].gpio, 1); 
+35    /* second button */ 
+36    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
+37        gpio_set_value(leds[0].gpio, 0); 
+38 
+39    return IRQ_HANDLED; 
+40} 
+41 
+42int init_module() 
+43{ 
+44    int ret = 0; 
+45 
+46    pr_info("%s\n", __func__); 
+47 
+48    /* register LED gpios */ 
+49    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
+50 
+51    if (ret) { 
+52        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
+53        return ret; 
+54    } 
+55 
+56    /* register BUTTON gpios */ 
+57    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
+58 
+59    if (ret) { 
+60        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
+61        goto fail1; 
+62    } 
+63 
+64    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
+65 
+66    ret = gpio_to_irq(buttons[0].gpio); 
+67 
+68    if (ret < 0) { 
+69        pr_err("Unable to request IRQ: %d\n", ret); 
+70        goto fail2; 
+71    } 
+72 
+73    button_irqs[0] = ret; 
+74 
+75    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
+76 
+77    ret = request_irq(button_irqs[0], button_isr, 
+78                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+79                      "gpiomod#button1", NULL); 
+80 
+81    if (ret) { 
+82        pr_err("Unable to request IRQ: %d\n", ret); 
+83        goto fail2; 
+84    } 
+85 
+86 
+87    ret = gpio_to_irq(buttons[1].gpio); 
+88 
+89    if (ret < 0) { 
+90        pr_err("Unable to request IRQ: %d\n", ret); 
+91        goto fail2; 
+92    } 
+93 
+94    button_irqs[1] = ret; 
+95 
+96    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
+97 
+98    ret = request_irq(button_irqs[1], button_isr, 
+99                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+100                      "gpiomod#button2", NULL); 
+101 
+102    if (ret) { 
+103        pr_err("Unable to request IRQ: %d\n", ret); 
+104        goto fail3; 
+105    } 
+106 
+107    return 0; 
+108 
+109/* cleanup what has been setup so far */ 
+110fail3: 
+111    free_irq(button_irqs[0], NULL); 
+112 
+113fail2: 
+114    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
+115 
+116fail1: 
+117    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+118 
+119    return ret; 
+120} 
+121 
+122void cleanup_module() 
+123{ 
+124    int i; 
+125 
+126    pr_info("%s\n", __func__); 
+127 
+128    /* free irqs */ 
+129    free_irq(button_irqs[0], NULL); 
+130    free_irq(button_irqs[1], NULL); 
+131 
+132    /* turn all LEDs off */ 
+133    for (i = 0; i < ARRAY_SIZE(leds); i++) 
+134        gpio_set_value(leds[i].gpio, 0); 
+135 
+136    /* unregister */ 
+137    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+138    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
+139} 
+140 
+141MODULE_LICENSE("GPL"); 
+142MODULE_DESCRIPTION("Handle some GPIO interrupts");
    +

    -

    0.15.3 Bottom Half

    -

    Suppose you want to do a bunch of stuff inside of an interrupt routine. A common +

    0.15.3 Bottom Half

    +

    Suppose you want to do a bunch of stuff inside of an interrupt routine. A common way to do that without rendering the interrupt unavailable for a significant duration is to combine it with a tasklet. This pushes the bulk of the work off into the scheduler. -

    The example below modifies the previous example to also run an additional task +

    The example below modifies the previous example to also run an additional task when an interrupt is triggered.

    - 
    1/* 
-2 * bottomhalf.c - Top and bottom half interrupt handling 
-3 * 
-4 * Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
-5 * from: 
-6 *    https://github.com/wendlers/rpi-kmod-samples 
-7 * 
-8 *  Press one button to turn on a LED and another to turn it off 
-9 */ 
-10 
-11#include <linux/delay.h> 
-12#include <linux/gpio.h> 
-13#include <linux/interrupt.h> 
-14#include <linux/kernel.h> 
-15#include <linux/module.h> 
-16 
-17static int button_irqs[] = {-1, -1}; 
-18 
-19/* Define GPIOs for LEDs. 
-20   Change the numbers for the GPIO on your board. */ 
-21static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
-22 
-23/* Define GPIOs for BUTTONS 
-24   Change the numbers for the GPIO on your board. */ 
-25static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
-26                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
-27 
-28/* Tasklet containing some non-trivial amount of processing */ 
-29static void bottomhalf_tasklet_fn(unsigned long data) 
-30{ 
-31    pr_info("Bottom half tasklet starts\n"); 
-32    /* do something which takes a while */ 
-33    mdelay(500); 
-34    pr_info("Bottom half tasklet ends\n"); 
-35} 
-36 
-37DECLARE_TASKLET(buttontask, bottomhalf_tasklet_fn, 0L); 
-38 
-39/* 
-40 * interrupt function triggered when a button is pressed 
-41 */ 
-42static irqreturn_t button_isr(int irq, void *data) 
-43{ 
-44    /* Do something quickly right now */ 
-45    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
-46        gpio_set_value(leds[0].gpio, 1); 
-47    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
-48        gpio_set_value(leds[0].gpio, 0); 
-49 
-50    /* Do the rest at leisure via the scheduler */ 
-51    tasklet_schedule(&buttontask); 
-52 
-53    return IRQ_HANDLED; 
-54} 
-55 
-56int init_module() 
-57{ 
-58    int ret = 0; 
-59 
-60    pr_info("%s\n", __func__); 
-61 
-62    /* register LED gpios */ 
-63    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
-64 
-65    if (ret) { 
-66        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
-67        return ret; 
-68    } 
-69 
-70    /* register BUTTON gpios */ 
-71    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
-72 
-73    if (ret) { 
-74        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
-75        goto fail1; 
-76    } 
-77 
-78    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
-79 
-80    ret = gpio_to_irq(buttons[0].gpio); 
-81 
-82    if (ret < 0) { 
-83        pr_err("Unable to request IRQ: %d\n", ret); 
-84        goto fail2; 
-85    } 
-86 
-87    button_irqs[0] = ret; 
-88 
-89    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
-90 
-91    ret = request_irq(button_irqs[0], button_isr, 
-92                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-93                      "gpiomod#button1", NULL); 
-94 
-95    if (ret) { 
-96        pr_err("Unable to request IRQ: %d\n", ret); 
-97        goto fail2; 
-98    } 
-99 
-100 
-101    ret = gpio_to_irq(buttons[1].gpio); 
-102 
-103    if (ret < 0) { 
-104        pr_err("Unable to request IRQ: %d\n", ret); 
-105        goto fail2; 
-106    } 
-107 
-108    button_irqs[1] = ret; 
-109 
-110    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
-111 
-112    ret = request_irq(button_irqs[1], button_isr, 
-113                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
-114                      "gpiomod#button2", NULL); 
-115 
-116    if (ret) { 
-117        pr_err("Unable to request IRQ: %d\n", ret); 
-118        goto fail3; 
-119    } 
-120 
-121    return 0; 
-122 
-123/* cleanup what has been setup so far */ 
-124fail3: 
-125    free_irq(button_irqs[0], NULL); 
-126 
-127fail2: 
-128    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
-129 
-130fail1: 
-131    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-132 
-133    return ret; 
-134} 
-135 
-136void cleanup_module() 
-137{ 
-138    int i; 
-139 
-140    pr_info("%s\n", __func__); 
-141 
-142    /* free irqs */ 
-143    free_irq(button_irqs[0], NULL); 
-144    free_irq(button_irqs[1], NULL); 
-145 
-146    /* turn all LEDs off */ 
-147    for (i = 0; i < ARRAY_SIZE(leds); i++) 
-148        gpio_set_value(leds[i].gpio, 0); 
-149 
-150    /* unregister */ 
-151    gpio_free_array(leds, ARRAY_SIZE(leds)); 
-152    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
-153} 
-154 
-155MODULE_LICENSE("GPL"); 
-156MODULE_DESCRIPTION("Interrupt with top and bottom half");
    -

    + 

    1/* 
+2 * bottomhalf.c - Top and bottom half interrupt handling 
+3 * 
+4 * Based upon the RPi example by Stefan Wendler (devnull@kaltpost.de) 
+5 * from: 
+6 *    https://github.com/wendlers/rpi-kmod-samples 
+7 * 
+8 *  Press one button to turn on a LED and another to turn it off 
+9 */ 
+10 
+11#include <linux/delay.h> 
+12#include <linux/gpio.h> 
+13#include <linux/interrupt.h> 
+14#include <linux/kernel.h> 
+15#include <linux/module.h> 
+16 
+17static int button_irqs[] = {-1, -1}; 
+18 
+19/* Define GPIOs for LEDs. 
+20   Change the numbers for the GPIO on your board. */ 
+21static struct gpio leds[] = {{4, GPIOF_OUT_INIT_LOW, "LED 1"}}; 
+22 
+23/* Define GPIOs for BUTTONS 
+24   Change the numbers for the GPIO on your board. */ 
+25static struct gpio buttons[] = {{17, GPIOF_IN, "LED 1 ON BUTTON"}, 
+26                                {18, GPIOF_IN, "LED 1 OFF BUTTON"}}; 
+27 
+28/* Tasklet containing some non-trivial amount of processing */ 
+29static void bottomhalf_tasklet_fn(unsigned long data) 
+30{ 
+31    pr_info("Bottom half tasklet starts\n"); 
+32    /* do something which takes a while */ 
+33    mdelay(500); 
+34    pr_info("Bottom half tasklet ends\n"); 
+35} 
+36 
+37DECLARE_TASKLET(buttontask, bottomhalf_tasklet_fn, 0L); 
+38 
+39/* 
+40 * interrupt function triggered when a button is pressed 
+41 */ 
+42static irqreturn_t button_isr(int irq, void *data) 
+43{ 
+44    /* Do something quickly right now */ 
+45    if (irq == button_irqs[0] && !gpio_get_value(leds[0].gpio)) 
+46        gpio_set_value(leds[0].gpio, 1); 
+47    else if (irq == button_irqs[1] && gpio_get_value(leds[0].gpio)) 
+48        gpio_set_value(leds[0].gpio, 0); 
+49 
+50    /* Do the rest at leisure via the scheduler */ 
+51    tasklet_schedule(&buttontask); 
+52 
+53    return IRQ_HANDLED; 
+54} 
+55 
+56int init_module() 
+57{ 
+58    int ret = 0; 
+59 
+60    pr_info("%s\n", __func__); 
+61 
+62    /* register LED gpios */ 
+63    ret = gpio_request_array(leds, ARRAY_SIZE(leds)); 
+64 
+65    if (ret) { 
+66        pr_err("Unable to request GPIOs for LEDs: %d\n", ret); 
+67        return ret; 
+68    } 
+69 
+70    /* register BUTTON gpios */ 
+71    ret = gpio_request_array(buttons, ARRAY_SIZE(buttons)); 
+72 
+73    if (ret) { 
+74        pr_err("Unable to request GPIOs for BUTTONs: %d\n", ret); 
+75        goto fail1; 
+76    } 
+77 
+78    pr_info("Current button1 value: %d\n", gpio_get_value(buttons[0].gpio)); 
+79 
+80    ret = gpio_to_irq(buttons[0].gpio); 
+81 
+82    if (ret < 0) { 
+83        pr_err("Unable to request IRQ: %d\n", ret); 
+84        goto fail2; 
+85    } 
+86 
+87    button_irqs[0] = ret; 
+88 
+89    pr_info("Successfully requested BUTTON1 IRQ # %d\n", button_irqs[0]); 
+90 
+91    ret = request_irq(button_irqs[0], button_isr, 
+92                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+93                      "gpiomod#button1", NULL); 
+94 
+95    if (ret) { 
+96        pr_err("Unable to request IRQ: %d\n", ret); 
+97        goto fail2; 
+98    } 
+99 
+100 
+101    ret = gpio_to_irq(buttons[1].gpio); 
+102 
+103    if (ret < 0) { 
+104        pr_err("Unable to request IRQ: %d\n", ret); 
+105        goto fail2; 
+106    } 
+107 
+108    button_irqs[1] = ret; 
+109 
+110    pr_info("Successfully requested BUTTON2 IRQ # %d\n", button_irqs[1]); 
+111 
+112    ret = request_irq(button_irqs[1], button_isr, 
+113                      IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING, 
+114                      "gpiomod#button2", NULL); 
+115 
+116    if (ret) { 
+117        pr_err("Unable to request IRQ: %d\n", ret); 
+118        goto fail3; 
+119    } 
+120 
+121    return 0; 
+122 
+123/* cleanup what has been setup so far */ 
+124fail3: 
+125    free_irq(button_irqs[0], NULL); 
+126 
+127fail2: 
+128    gpio_free_array(buttons, ARRAY_SIZE(leds)); 
+129 
+130fail1: 
+131    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+132 
+133    return ret; 
+134} 
+135 
+136void cleanup_module() 
+137{ 
+138    int i; 
+139 
+140    pr_info("%s\n", __func__); 
+141 
+142    /* free irqs */ 
+143    free_irq(button_irqs[0], NULL); 
+144    free_irq(button_irqs[1], NULL); 
+145 
+146    /* turn all LEDs off */ 
+147    for (i = 0; i < ARRAY_SIZE(leds); i++) 
+148        gpio_set_value(leds[i].gpio, 0); 
+149 
+150    /* unregister */ 
+151    gpio_free_array(leds, ARRAY_SIZE(leds)); 
+152    gpio_free_array(buttons, ARRAY_SIZE(buttons)); 
+153} 
+154 
+155MODULE_LICENSE("GPL"); 
+156MODULE_DESCRIPTION("Interrupt with top and bottom half");
    +

    -

    0.16 Crypto

    -

    At the dawn of the internet everybody trusted everybody completely…but that didn’t +

    0.16 Crypto

    +

    At the dawn of the internet everybody trusted everybody completely…but that didn’t work out so well. When this guide was originally written it was a more innocent era in which almost nobody actually gave a damn about crypto - least of all kernel developers. That’s certainly no longer the case now. To handle crypto stuff the kernel @@ -4664,295 +4689,295 @@ favourite hash functions. -

    +

    -

    0.16.1 Hash functions

    -

    Calculating and checking the hashes of things is a common operation. Here is a +

    0.16.1 Hash functions

    +

    Calculating and checking the hashes of things is a common operation. Here is a demonstration of how to calculate a sha256 hash within a kernel module.

    - 
    1/* 
-2 *  cryptosha256.c 
-3 */ 
-4#include <crypto/internal/hash.h> 
-5#include <linux/module.h> 
-6 
-7#define SHA256_LENGTH 32 
-8 
-9static void show_hash_result(char *plaintext, char *hash_sha256) 
-10{ 
-11    int i; 
-12    char str[SHA256_LENGTH * 2 + 1]; 
-13 
-14    pr_info("sha256 test for string: \"%s\"\n", plaintext); 
-15    for (i = 0; i < SHA256_LENGTH; i++) 
-16        sprintf(&str[i * 2], "%02x", (unsigned char) hash_sha256[i]); 
-17    str[i * 2] = 0; 
-18    pr_info("%s\n", str); 
-19} 
-20 
-21int cryptosha256_init(void) 
-22{ 
-23    char *plaintext = "This is a test"; 
-24    char hash_sha256[SHA256_LENGTH]; 
-25    struct crypto_shash *sha256; 
-26    struct shash_desc *shash; 
-27 
-28    sha256 = crypto_alloc_shash("sha256", 0, 0); 
-29    if (IS_ERR(sha256)) 
-30        return -1; 
-31 
-32    shash = kmalloc(sizeof(struct shash_desc) + crypto_shash_descsize(sha256), 
-33                    GFP_KERNEL); 
-34    if (!shash) 
-35        return -ENOMEM; 
-36 
-37    shash->tfm = sha256; 
-38 
-39    if (crypto_shash_init(shash)) 
-40        return -1; 
-41 
-42    if (crypto_shash_update(shash, plaintext, strlen(plaintext))) 
-43        return -1; 
-44 
-45    if (crypto_shash_final(shash, hash_sha256)) 
-46        return -1; 
-47 
-48    kfree(shash); 
-49    crypto_free_shash(sha256); 
-50 
-51    show_hash_result(plaintext, hash_sha256); 
-52 
-53    return 0; 
-54} 
-55 
-56void cryptosha256_exit(void) {} 
-57 
-58module_init(cryptosha256_init); 
-59module_exit(cryptosha256_exit); 
-60 
-61MODULE_DESCRIPTION("sha256 hash test"); 
-62MODULE_LICENSE("GPL");
    -

    Make and install the module: + 

    1/* 
+2 *  cryptosha256.c 
+3 */ 
+4#include <crypto/internal/hash.h> 
+5#include <linux/module.h> 
+6 
+7#define SHA256_LENGTH 32 
+8 
+9static void show_hash_result(char *plaintext, char *hash_sha256) 
+10{ 
+11    int i; 
+12    char str[SHA256_LENGTH * 2 + 1]; 
+13 
+14    pr_info("sha256 test for string: \"%s\"\n", plaintext); 
+15    for (i = 0; i < SHA256_LENGTH; i++) 
+16        sprintf(&str[i * 2], "%02x", (unsigned char) hash_sha256[i]); 
+17    str[i * 2] = 0; 
+18    pr_info("%s\n", str); 
+19} 
+20 
+21int cryptosha256_init(void) 
+22{ 
+23    char *plaintext = "This is a test"; 
+24    char hash_sha256[SHA256_LENGTH]; 
+25    struct crypto_shash *sha256; 
+26    struct shash_desc *shash; 
+27 
+28    sha256 = crypto_alloc_shash("sha256", 0, 0); 
+29    if (IS_ERR(sha256)) 
+30        return -1; 
+31 
+32    shash = kmalloc(sizeof(struct shash_desc) + crypto_shash_descsize(sha256), 
+33                    GFP_KERNEL); 
+34    if (!shash) 
+35        return -ENOMEM; 
+36 
+37    shash->tfm = sha256; 
+38 
+39    if (crypto_shash_init(shash)) 
+40        return -1; 
+41 
+42    if (crypto_shash_update(shash, plaintext, strlen(plaintext))) 
+43        return -1; 
+44 
+45    if (crypto_shash_final(shash, hash_sha256)) 
+46        return -1; 
+47 
+48    kfree(shash); 
+49    crypto_free_shash(sha256); 
+50 
+51    show_hash_result(plaintext, hash_sha256); 
+52 
+53    return 0; 
+54} 
+55 
+56void cryptosha256_exit(void) {} 
+57 
+58module_init(cryptosha256_init); 
+59module_exit(cryptosha256_exit); 
+60 
+61MODULE_DESCRIPTION("sha256 hash test"); 
+62MODULE_LICENSE("GPL");
    +

    Make and install the module:

    - 
    1make 
-2sudo insmod cryptosha256.ko 
-3dmesg
    -

    And you should see that the hash was calculated for the test string. -

    Finally, remove the test module: + 

    1make 
+2sudo insmod cryptosha256.ko 
+3dmesg
    +

    And you should see that the hash was calculated for the test string. +

    Finally, remove the test module:

    - 
    1sudo rmmod cryptosha256
    -

    + 

    1sudo rmmod cryptosha256
    +

    -

    0.16.2 Symmetric key encryption

    -

    Here is an example of symmetrically encrypting a string using the AES algorithm +

    0.16.2 Symmetric key encryption

    +

    Here is an example of symmetrically encrypting a string using the AES algorithm and a password.

    - 
    1/* 
-2 *  cryptosk.c 
-3 */ 
-4#include <crypto/internal/skcipher.h> 
-5#include <linux/crypto.h> 
-6#include <linux/module.h> 
-7 
-8#define SYMMETRIC_KEY_LENGTH 32 
-9#define CIPHER_BLOCK_SIZE 16 
-10 
-11struct tcrypt_result { 
-12    struct completion completion; 
-13    int err; 
-14}; 
-15 
-16struct skcipher_def { 
-17    struct scatterlist sg; 
-18    struct crypto_skcipher *tfm; 
-19    struct skcipher_request *req; 
-20    struct tcrypt_result result; 
-21    char *scratchpad; 
-22    char *ciphertext; 
-23    char *ivdata; 
-24}; 
-25 
-26static struct skcipher_def sk; 
-27 
-28static void test_skcipher_finish(struct skcipher_def *sk) 
-29{ 
-30    if (sk->tfm) 
-31        crypto_free_skcipher(sk->tfm); 
-32    if (sk->req) 
-33        skcipher_request_free(sk->req); 
-34    if (sk->ivdata) 
-35        kfree(sk->ivdata); 
-36    if (sk->scratchpad) 
-37        kfree(sk->scratchpad); 
-38    if (sk->ciphertext) 
-39        kfree(sk->ciphertext); 
-40} 
-41 
-42static int test_skcipher_result(struct skcipher_def *sk, int rc) 
-43{ 
-44    switch (rc) { 
-45    case 0: 
-46        break; 
-47    case -EINPROGRESS || -EBUSY: 
-48        rc = wait_for_completion_interruptible(&sk->result.completion); 
-49        if (!rc && !sk->result.err) { 
-50            reinit_completion(&sk->result.completion); 
-51            break; 
-52        } 
-53        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
-54                sk->result.err); 
-55        break; 
-56    default: 
-57        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
-58                sk->result.err); 
-59        break; 
-60    } 
-61 
-62    init_completion(&sk->result.completion); 
-63 
-64    return rc; 
-65} 
-66 
-67static void test_skcipher_callback(struct crypto_async_request *req, int error) 
-68{ 
-69    struct tcrypt_result *result = req->data; 
-70    /* int ret; */ 
-71 
-72    if (error == -EINPROGRESS) 
-73        return; 
-74 
-75    result->err = error; 
-76    complete(&result->completion); 
-77    pr_info("Encryption finished successfully\n"); 
-78 
-79    /* decrypt data */ 
-80    /* 
-81    memset((void*)sk.scratchpad, '-', CIPHER_BLOCK_SIZE); 
-82    ret = crypto_skcipher_decrypt(sk.req); 
-83    ret = test_skcipher_result(&sk, ret); 
-84    if (ret) 
-85        return; 
-86 
-87    sg_copy_from_buffer(&sk.sg, 1, sk.scratchpad, CIPHER_BLOCK_SIZE); 
-88    sk.scratchpad[CIPHER_BLOCK_SIZE-1] = 0; 
-89 
-90    pr_info("Decryption request successful\n"); 
-91    pr_info("Decrypted: %s\n", sk.scratchpad); 
-92    */ 
-93} 
-94 
-95static int test_skcipher_encrypt(char *plaintext, 
-96                                 char *password, 
-97                                 struct skcipher_def *sk) 
-98{ 
-99    int ret = -EFAULT; 
-100    unsigned char key[SYMMETRIC_KEY_LENGTH]; 
-101 
-102    if (!sk->tfm) { 
-103        sk->tfm = crypto_alloc_skcipher("cbc-aes-aesni", 0, 0); 
-104        if (IS_ERR(sk->tfm)) { 
-105            pr_info("could not allocate skcipher handle\n"); 
-106            return PTR_ERR(sk->tfm); 
-107        } 
-108    } 
-109 
-110    if (!sk->req) { 
-111        sk->req = skcipher_request_alloc(sk->tfm, GFP_KERNEL); 
-112        if (!sk->req) { 
-113            pr_info("could not allocate skcipher request\n"); 
-114            ret = -ENOMEM; 
-115            goto out; 
-116        } 
-117    } 
-118 
-119    skcipher_request_set_callback(sk->req, CRYPTO_TFM_REQ_MAY_BACKLOG, 
-120                                  test_skcipher_callback, &sk->result); 
-121 
-122    /* clear the key */ 
-123    memset((void *) key, '\0', SYMMETRIC_KEY_LENGTH); 
-124 
-125    /* Use the world's favourite password */ 
-126    sprintf((char *) key, "%s", password); 
-127 
-128    /* AES 256 with given symmetric key */ 
-129    if (crypto_skcipher_setkey(sk->tfm, key, SYMMETRIC_KEY_LENGTH)) { 
-130        pr_info("key could not be set\n"); 
-131        ret = -EAGAIN; 
-132        goto out; 
-133    } 
-134    pr_info("Symmetric key: %s\n", key); 
-135    pr_info("Plaintext: %s\n", plaintext); 
-136 
-137    if (!sk->ivdata) { 
-138        /* see https://en.wikipedia.org/wiki/Initialization_vector */ 
-139        sk->ivdata = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
-140        if (!sk->ivdata) { 
-141            pr_info("could not allocate ivdata\n"); 
-142            goto out; 
-143        } 
-144        get_random_bytes(sk->ivdata, CIPHER_BLOCK_SIZE); 
-145    } 
-146 
-147    if (!sk->scratchpad) { 
-148        /* The text to be encrypted */ 
-149        sk->scratchpad = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
-150        if (!sk->scratchpad) { 
-151            pr_info("could not allocate scratchpad\n"); 
-152            goto out; 
-153        } 
-154    } 
-155    sprintf((char *) sk->scratchpad, "%s", plaintext); 
-156 
-157    sg_init_one(&sk->sg, sk->scratchpad, CIPHER_BLOCK_SIZE); 
-158    skcipher_request_set_crypt(sk->req, &sk->sg, &sk->sg, CIPHER_BLOCK_SIZE, 
-159                               sk->ivdata); 
-160    init_completion(&sk->result.completion); 
-161 
-162    /* encrypt data */ 
-163    ret = crypto_skcipher_encrypt(sk->req); 
-164    ret = test_skcipher_result(sk, ret); 
-165    if (ret) 
-166        goto out; 
-167 
-168    pr_info("Encryption request successful\n"); 
-169 
-170out: 
-171    return ret; 
-172} 
-173 
-174int cryptoapi_init(void) 
-175{ 
-176    /* The world's favorite password */ 
-177    char *password = "password123"; 
-178 
-179    sk.tfm = NULL; 
-180    sk.req = NULL; 
-181    sk.scratchpad = NULL; 
-182    sk.ciphertext = NULL; 
-183    sk.ivdata = NULL; 
-184 
-185    test_skcipher_encrypt("Testing", password, &sk); 
-186    return 0; 
-187} 
-188 
-189void cryptoapi_exit(void) 
-190{ 
-191    test_skcipher_finish(&sk); 
-192} 
-193 
-194module_init(cryptoapi_init); 
-195module_exit(cryptoapi_exit); 
-196 
-197MODULE_DESCRIPTION("Symmetric key encryption example"); 
-198MODULE_LICENSE("GPL");
    -

    + 

    1/* 
+2 *  cryptosk.c 
+3 */ 
+4#include <crypto/internal/skcipher.h> 
+5#include <linux/crypto.h> 
+6#include <linux/module.h> 
+7 
+8#define SYMMETRIC_KEY_LENGTH 32 
+9#define CIPHER_BLOCK_SIZE 16 
+10 
+11struct tcrypt_result { 
+12    struct completion completion; 
+13    int err; 
+14}; 
+15 
+16struct skcipher_def { 
+17    struct scatterlist sg; 
+18    struct crypto_skcipher *tfm; 
+19    struct skcipher_request *req; 
+20    struct tcrypt_result result; 
+21    char *scratchpad; 
+22    char *ciphertext; 
+23    char *ivdata; 
+24}; 
+25 
+26static struct skcipher_def sk; 
+27 
+28static void test_skcipher_finish(struct skcipher_def *sk) 
+29{ 
+30    if (sk->tfm) 
+31        crypto_free_skcipher(sk->tfm); 
+32    if (sk->req) 
+33        skcipher_request_free(sk->req); 
+34    if (sk->ivdata) 
+35        kfree(sk->ivdata); 
+36    if (sk->scratchpad) 
+37        kfree(sk->scratchpad); 
+38    if (sk->ciphertext) 
+39        kfree(sk->ciphertext); 
+40} 
+41 
+42static int test_skcipher_result(struct skcipher_def *sk, int rc) 
+43{ 
+44    switch (rc) { 
+45    case 0: 
+46        break; 
+47    case -EINPROGRESS || -EBUSY: 
+48        rc = wait_for_completion_interruptible(&sk->result.completion); 
+49        if (!rc && !sk->result.err) { 
+50            reinit_completion(&sk->result.completion); 
+51            break; 
+52        } 
+53        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
+54                sk->result.err); 
+55        break; 
+56    default: 
+57        pr_info("skcipher encrypt returned with %d result %d\n", rc, 
+58                sk->result.err); 
+59        break; 
+60    } 
+61 
+62    init_completion(&sk->result.completion); 
+63 
+64    return rc; 
+65} 
+66 
+67static void test_skcipher_callback(struct crypto_async_request *req, int error) 
+68{ 
+69    struct tcrypt_result *result = req->data; 
+70    /* int ret; */ 
+71 
+72    if (error == -EINPROGRESS) 
+73        return; 
+74 
+75    result->err = error; 
+76    complete(&result->completion); 
+77    pr_info("Encryption finished successfully\n"); 
+78 
+79    /* decrypt data */ 
+80    /* 
+81    memset((void*)sk.scratchpad, '-', CIPHER_BLOCK_SIZE); 
+82    ret = crypto_skcipher_decrypt(sk.req); 
+83    ret = test_skcipher_result(&sk, ret); 
+84    if (ret) 
+85        return; 
+86 
+87    sg_copy_from_buffer(&sk.sg, 1, sk.scratchpad, CIPHER_BLOCK_SIZE); 
+88    sk.scratchpad[CIPHER_BLOCK_SIZE-1] = 0; 
+89 
+90    pr_info("Decryption request successful\n"); 
+91    pr_info("Decrypted: %s\n", sk.scratchpad); 
+92    */ 
+93} 
+94 
+95static int test_skcipher_encrypt(char *plaintext, 
+96                                 char *password, 
+97                                 struct skcipher_def *sk) 
+98{ 
+99    int ret = -EFAULT; 
+100    unsigned char key[SYMMETRIC_KEY_LENGTH]; 
+101 
+102    if (!sk->tfm) { 
+103        sk->tfm = crypto_alloc_skcipher("cbc-aes-aesni", 0, 0); 
+104        if (IS_ERR(sk->tfm)) { 
+105            pr_info("could not allocate skcipher handle\n"); 
+106            return PTR_ERR(sk->tfm); 
+107        } 
+108    } 
+109 
+110    if (!sk->req) { 
+111        sk->req = skcipher_request_alloc(sk->tfm, GFP_KERNEL); 
+112        if (!sk->req) { 
+113            pr_info("could not allocate skcipher request\n"); 
+114            ret = -ENOMEM; 
+115            goto out; 
+116        } 
+117    } 
+118 
+119    skcipher_request_set_callback(sk->req, CRYPTO_TFM_REQ_MAY_BACKLOG, 
+120                                  test_skcipher_callback, &sk->result); 
+121 
+122    /* clear the key */ 
+123    memset((void *) key, '\0', SYMMETRIC_KEY_LENGTH); 
+124 
+125    /* Use the world's favourite password */ 
+126    sprintf((char *) key, "%s", password); 
+127 
+128    /* AES 256 with given symmetric key */ 
+129    if (crypto_skcipher_setkey(sk->tfm, key, SYMMETRIC_KEY_LENGTH)) { 
+130        pr_info("key could not be set\n"); 
+131        ret = -EAGAIN; 
+132        goto out; 
+133    } 
+134    pr_info("Symmetric key: %s\n", key); 
+135    pr_info("Plaintext: %s\n", plaintext); 
+136 
+137    if (!sk->ivdata) { 
+138        /* see https://en.wikipedia.org/wiki/Initialization_vector */ 
+139        sk->ivdata = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
+140        if (!sk->ivdata) { 
+141            pr_info("could not allocate ivdata\n"); 
+142            goto out; 
+143        } 
+144        get_random_bytes(sk->ivdata, CIPHER_BLOCK_SIZE); 
+145    } 
+146 
+147    if (!sk->scratchpad) { 
+148        /* The text to be encrypted */ 
+149        sk->scratchpad = kmalloc(CIPHER_BLOCK_SIZE, GFP_KERNEL); 
+150        if (!sk->scratchpad) { 
+151            pr_info("could not allocate scratchpad\n"); 
+152            goto out; 
+153        } 
+154    } 
+155    sprintf((char *) sk->scratchpad, "%s", plaintext); 
+156 
+157    sg_init_one(&sk->sg, sk->scratchpad, CIPHER_BLOCK_SIZE); 
+158    skcipher_request_set_crypt(sk->req, &sk->sg, &sk->sg, CIPHER_BLOCK_SIZE, 
+159                               sk->ivdata); 
+160    init_completion(&sk->result.completion); 
+161 
+162    /* encrypt data */ 
+163    ret = crypto_skcipher_encrypt(sk->req); 
+164    ret = test_skcipher_result(sk, ret); 
+165    if (ret) 
+166        goto out; 
+167 
+168    pr_info("Encryption request successful\n"); 
+169 
+170out: 
+171    return ret; 
+172} 
+173 
+174int cryptoapi_init(void) 
+175{ 
+176    /* The world's favorite password */ 
+177    char *password = "password123"; 
+178 
+179    sk.tfm = NULL; 
+180    sk.req = NULL; 
+181    sk.scratchpad = NULL; 
+182    sk.ciphertext = NULL; 
+183    sk.ivdata = NULL; 
+184 
+185    test_skcipher_encrypt("Testing", password, &sk); 
+186    return 0; 
+187} 
+188 
+189void cryptoapi_exit(void) 
+190{ 
+191    test_skcipher_finish(&sk); 
+192} 
+193 
+194module_init(cryptoapi_init); 
+195module_exit(cryptoapi_exit); 
+196 
+197MODULE_DESCRIPTION("Symmetric key encryption example"); 
+198MODULE_LICENSE("GPL");
    +

    -

    0.17 Standardising the interfaces: The Device Model

    -

    Up to this point we’ve seen all kinds of modules doing all kinds of things, but there +

    0.17 Standardising the interfaces: The Device Model

    +

    Up to this point we’ve seen all kinds of modules doing all kinds of things, but there was no consistency in their interfaces with the rest of the kernel. To impose some consistency such that there is at minimum a standardised way to start, suspend and resume a device a device model was added. An example is show below, and you can @@ -4960,188 +4985,188 @@ use this as a template to add your own suspend, resume or other interface functions.

    - 
    1/* 
-2 *  devicemodel.c 
-3 */ 
-4#include <linux/kernel.h> 
-5#include <linux/module.h> 
-6#include <linux/platform_device.h> 
-7 
-8struct devicemodel_data { 
-9    char *greeting; 
-10    int number; 
-11}; 
-12 
-13static int devicemodel_probe(struct platform_device *dev) 
-14{ 
-15    struct devicemodel_data *pd = 
-16        (struct devicemodel_data *) (dev->dev.platform_data); 
-17 
-18    pr_info("devicemodel probe\n"); 
-19    pr_info("devicemodel greeting: %s; %d\n", pd->greeting, pd->number); 
-20 
-21    /* Your device initialisation code */ 
-22 
-23    return 0; 
-24} 
-25 
-26static int devicemodel_remove(struct platform_device *dev) 
-27{ 
-28    pr_info("devicemodel example removed\n"); 
-29 
-30    /* Your device removal code */ 
-31 
-32    return 0; 
-33} 
-34 
-35static int devicemodel_suspend(struct device *dev) 
-36{ 
-37    pr_info("devicemodel example suspend\n"); 
-38 
-39    /* Your device suspend code */ 
-40 
-41    return 0; 
-42} 
-43 
-44static int devicemodel_resume(struct device *dev) 
-45{ 
-46    pr_info("devicemodel example resume\n"); 
-47 
-48    /* Your device resume code */ 
-49 
-50    return 0; 
-51} 
-52 
-53static const struct dev_pm_ops devicemodel_pm_ops = { 
-54    .suspend = devicemodel_suspend, 
-55    .resume = devicemodel_resume, 
-56    .poweroff = devicemodel_suspend, 
-57    .freeze = devicemodel_suspend, 
-58    .thaw = devicemodel_resume, 
-59    .restore = devicemodel_resume, 
-60}; 
-61 
-62static struct platform_driver devicemodel_driver = { 
-63    .driver = 
-64        { 
-65            .name = "devicemodel_example", 
-66            .owner = THIS_MODULE, 
-67            .pm = &devicemodel_pm_ops, 
-68        }, 
-69    .probe = devicemodel_probe, 
-70    .remove = devicemodel_remove, 
-71}; 
-72 
-73static int devicemodel_init(void) 
-74{ 
-75    int ret; 
-76 
-77    pr_info("devicemodel init\n"); 
-78 
-79    ret = platform_driver_register(&devicemodel_driver); 
-80 
-81    if (ret) { 
-82        pr_err("Unable to register driver\n"); 
-83        return ret; 
-84    } 
-85 
-86    return 0; 
-87} 
-88 
-89static void devicemodel_exit(void) 
-90{ 
-91    pr_info("devicemodel exit\n"); 
-92    platform_driver_unregister(&devicemodel_driver); 
-93} 
-94 
-95MODULE_LICENSE("GPL"); 
-96MODULE_DESCRIPTION("Linux Device Model example"); 
-97 
-98module_init(devicemodel_init); 
-99module_exit(devicemodel_exit);
    + 
    1/* 
+2 *  devicemodel.c 
+3 */ 
+4#include <linux/kernel.h> 
+5#include <linux/module.h> 
+6#include <linux/platform_device.h> 
+7 
+8struct devicemodel_data { 
+9    char *greeting; 
+10    int number; 
+11}; 
+12 
+13static int devicemodel_probe(struct platform_device *dev) 
+14{ 
+15    struct devicemodel_data *pd = 
+16        (struct devicemodel_data *) (dev->dev.platform_data); 
+17 
+18    pr_info("devicemodel probe\n"); 
+19    pr_info("devicemodel greeting: %s; %d\n", pd->greeting, pd->number); 
+20 
+21    /* Your device initialisation code */ 
+22 
+23    return 0; 
+24} 
+25 
+26static int devicemodel_remove(struct platform_device *dev) 
+27{ 
+28    pr_info("devicemodel example removed\n"); 
+29 
+30    /* Your device removal code */ 
+31 
+32    return 0; 
+33} 
+34 
+35static int devicemodel_suspend(struct device *dev) 
+36{ 
+37    pr_info("devicemodel example suspend\n"); 
+38 
+39    /* Your device suspend code */ 
+40 
+41    return 0; 
+42} 
+43 
+44static int devicemodel_resume(struct device *dev) 
+45{ 
+46    pr_info("devicemodel example resume\n"); 
+47 
+48    /* Your device resume code */ 
+49 
+50    return 0; 
+51} 
+52 
+53static const struct dev_pm_ops devicemodel_pm_ops = { 
+54    .suspend = devicemodel_suspend, 
+55    .resume = devicemodel_resume, 
+56    .poweroff = devicemodel_suspend, 
+57    .freeze = devicemodel_suspend, 
+58    .thaw = devicemodel_resume, 
+59    .restore = devicemodel_resume, 
+60}; 
+61 
+62static struct platform_driver devicemodel_driver = { 
+63    .driver = 
+64        { 
+65            .name = "devicemodel_example", 
+66            .owner = THIS_MODULE, 
+67            .pm = &devicemodel_pm_ops, 
+68        }, 
+69    .probe = devicemodel_probe, 
+70    .remove = devicemodel_remove, 
+71}; 
+72 
+73static int devicemodel_init(void) 
+74{ 
+75    int ret; 
+76 
+77    pr_info("devicemodel init\n"); 
+78 
+79    ret = platform_driver_register(&devicemodel_driver); 
+80 
+81    if (ret) { 
+82        pr_err("Unable to register driver\n"); 
+83        return ret; 
+84    } 
+85 
+86    return 0; 
+87} 
+88 
+89static void devicemodel_exit(void) 
+90{ 
+91    pr_info("devicemodel exit\n"); 
+92    platform_driver_unregister(&devicemodel_driver); 
+93} 
+94 
+95MODULE_LICENSE("GPL"); 
+96MODULE_DESCRIPTION("Linux Device Model example"); 
+97 
+98module_init(devicemodel_init); 
+99module_exit(devicemodel_exit);
    -

    +

    -

    0.18 Optimizations

    -

    +

    0.18 Optimizations

    +

    -

    0.18.1 Likely and Unlikely conditions

    -

    Sometimes you might want your code to run as quickly as possible, especially if +

    0.18.1 Likely and Unlikely conditions

    +

    Sometimes you might want your code to run as quickly as possible, especially if it’s handling an interrupt or doing something which might cause noticible latency. If your code contains boolean conditions and if you know that the conditions are almost always likely to evaluate as either true or false, then you can allow the compiler to optimise for this using the likely and unlikely macros. -

    For example, when allocating memory you’re almost always expecting this to +

    For example, when allocating memory you’re almost always expecting this to succeed.

    - 
    1bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx); 
-2if (unlikely(!bvl)) { 
-3  mempool_free(bio, bio_pool); 
-4  bio = NULL; 
-5  goto out; 
-6}
    -

    When the unlikely macro is used the compiler alters its machine instruction + 

    1bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx); 
+2if (unlikely(!bvl)) { 
+3  mempool_free(bio, bio_pool); 
+4  bio = NULL; 
+5  goto out; 
+6}
    +

    When the unlikely macro is used the compiler alters its machine instruction output so that it continues along the false branch and only jumps if the condition is true. That avoids flushing the processor pipeline. The opposite happens if you use the likely macro.

    -

    0.19 Common Pitfalls

    -

    Before I send you on your way to go out into the world and write kernel modules, +

    0.19 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. -

    +

    -

    0.19.1 Using standard libraries

    -

    You can’t do that. In a kernel module you can only use kernel functions, which are +

    0.19.1 Using standard libraries

    +

    You can’t do that. In a kernel module you can only use kernel functions, which are the functions you can see in /proc/kallsyms. -

    +

    -

    0.19.2 Disabling interrupts

    -

    You might need to do this for a short time and that is OK, but if you don’t +

    0.19.2 Disabling interrupts

    +

    You might need to do this for a short time and that is OK, but if you don’t enable them afterwards, your system will be stuck and you’ll have to power it off. -

    +

    -

    0.19.3 Sticking your head inside a large carnivore

    -

    I probably don’t have to warn you about this, but I figured I will anyway, just in +

    0.19.3 Sticking your head inside a large carnivore

    +

    I probably don’t have to warn you about this, but I figured I will anyway, just in case. -

    +

    -

    0.20 Where To Go From Here?

    -

    I could easily have squeezed a few more chapters into this book. I could have added a +

    0.20 Where To Go From Here?

    +

    I could easily have squeezed a few more chapters into this book. I could have added a chapter about creating new file systems, or about adding new protocol stacks (as if there’s a need for that – you’d have to dig underground to find a protocol stack not supported by Linux). I could have added explanations of the kernel mechanisms we haven’t touched upon, such as bootstrapping or the disk interface. -

    However, I chose not to. My purpose in writing this book was to provide initiation +

    However, I chose not to. My purpose in writing this book was to provide initiation into the mysteries of kernel module programming and to teach the common techniques for that purpose. For people seriously interested in kernel programming, I recommend kernelnewbies.org and the Documentation subdirectory within the kernel source code which isn’t always easy to understand but can be a starting point for further investigation. Also, as Linus said, the best way to learn the kernel is to read the source code yourself. -

    If you’re interested in more examples of short kernel modules then searching on +

    If you’re interested in more examples of short kernel modules then searching on sites such as Github and Gitlab is a good way to start, although there is a lot of duplication of older LKMPG examples which may not compile with newer kernel versions. You will also be able to find examples of the use of kernel modules to attack or compromise systems or exfiltrate data and those can be useful for thinking about how to defend systems and learning about existing security mechanisms within the kernel. -

    I hope I have helped you in your quest to become a better programmer, or at +

    I hope I have helped you in your quest to become a better programmer, or at least to have fun through technology. And, if you do write useful kernel modules, I hope you publish them under the GPL, so I can use them too. -

    If you’d like to contribute to this guide or notice anything glaringly wrong, please +

    If you’d like to contribute to this guide or notice anything glaringly wrong, please create an issue at https://github.com/sysprog21/lkmpg. -

    Happy hacking. +

    Happy hacking.

    diff --git a/lkmpg0x.svg b/lkmpg0x.svg index e72111c..c2037bf 100644 --- a/lkmpg0x.svg +++ b/lkmpg0x.svg @@ -1,6 +1,6 @@ - + @@ -12,15 +12,15 @@ - - - - - - - - - - + + + + + + + + + + \ No newline at end of file