Linux Debugfs implementation as a userspace interface

-
debugfs is a Linux file system. It is a user space interface to and from the Linux kernel. It is mounted at /sys/kernel/debug. The kernel must be compiled with the CONFIG_DEBUG_FS option to use this file system. It exists in the Linux kernel since version 2.6.10-rc3 [1]
In the below example, you will see an example to a debugfs. A 32-bit flag is created under the /sys/kernel/debug/debugfs_ex directory called enable for read and write.


debugfs_ex.c


#include <linux/version.h>
#include <linux/module.h>
#include <linux/debugfs.h>

static struct dentry *dn; /* Dentry object for debugfs directory */

static int enable;  /* Enable flag for debugfs_ex module */

/*
 * debugfs_ex_init()
 */
static int __init debugfs_ex_init(void)
{
 printk(KERN_INFO "debugfs_ex init\n");

 dn = debugfs_create_dir("debugfs_ex", NULL);
 if (!dn) {
  printk(KERN_ERR "Failed to create debugfs_ex directory in debugfs\n");
  return -1;
 }

 /*
  * Create a 32-bit flag under debugfs_ex directory.
  * This flag is readable and writable.
  */
 if (!debugfs_create_u32("enable", 0666, dn, (u32 *)&enable)) {
  printk(KERN_ALERT "Failed to create enable file in debugfs\n");
  goto fail;
 }

 return 0;
fail:
 debugfs_remove_recursive(dn);

 printk(KERN_INFO "debugfs_ex init failed\n");

 return -1;
}

/*
 * debugfs_ex_exit()
 */
static void __exit debugfs_ex_exit(void)
{
 printk(KERN_INFO "debugfs_ex exit\n");

 if (dn) {
  printk(KERN_INFO "remove debugfs_ex node\n");
  debugfs_remove_recursive(dn);
 }

 printk(KERN_INFO "debugfs_ex exit complete\n");
}

module_init(debugfs_ex_init)
module_exit(debugfs_ex_exit)

MODULE_DESCRIPTION("Debugfs Example");
#ifdef MODULE_LICENSE
MODULE_LICENSE("Dual BSD/GPL");
#endif

Makefile:


obj-m += debugfs_ex.o

all:
 make -C /lib/modules/$(shell uname -r)/build M=$(shell pwd) modules

clean:
 make -C /lib/modules/$(shell uname -r)/build M=$(shell pwd) clean

Compilation:

You need to put the Makefile and debugfs_ex.c file into the same directory and simply run the make command to create the kernel loadable module debugfs_ex.ko

Usage:

Make sure you have the right permission to use your debugfs directory.

Install the module:
insmod ./debugfs_ex.ko

Write 5 to the flag:
echo 5 > /sys/kernel/debug/debugfs_ex/enable

Read the flag:
cat  /sys/kernel/debug/debugfs_ex/enable

Remove module:
rmmod debugfs_ex



[1] https://en.wikipedia.org/wiki/Debugfs















































Comments








Post a Comment
























Popular posts from this blog









How to Use container_of in Linux Kernel Development






-















In this post, I will explain the usage of linux's container_of  macro with a sample code.   If you know address of a data structure's field, you can get the address of the instance of that data structure by using the container_of  macro. This is the simplest explanation of this macro.   Let's say you have a function called with a pointer parameter and you know that the pointer is pointing an address of a data structure field.   struct  my_object {         int  a;         int  b;  } obj;  function(&obj->b);   If you want to access the other field "a" in the function, you need to use container_of  macro.   Let's see the below sample code and it's output to understand it easily.    /*   * my_object structure.   */  struct  my_object {         int  a;         int  b;         int  c; };  /*   * container_test()   *      Test function for usage of container_of macro.   *   * This function is passed a pointer which points a field of   * an allocated my_obj...








Read more










Notification Chains in Linux Kernel






-















  Notification chains in Linux kernel is a mechanism for the subsystems being informed by the events generated by the other subsystems.   In other words, if you are developing a new driver in Linux and you want to listen some events from other driver and handle these events in your own driver, you need to use the Linux's notification chains mechanism.   As an example, if you develop a driver and this driver needs to be informed about the net device events such as net device UP, DOWN events, the driver needs to register itself to the net device notifications with the  register_netdevice_notifier  function. Similarly, it unregisters itself with the unregister_netdevice_notifier  function. These functions are provided by the net device subsystem.   The APIs and the data structures of notification chains are defined in the include/linux/notifier.h  and  kernel/notifier.c  files under linux source code.   In this post, I am going to explain how this notification chains mechanism wo...








Read more










Build and Flash OpenWrt for Raspberry Pi






-














Image







   OpenWrt is an open source project for embedded operating system based on Linux, primarily used on embedded devices to route network traffic. [ 1 ]    In this post, I will write about how to get the OpenWrt source, compile it for your Raspberry Pi board type and flash the SD card with the generated image.    First of all, you need a Linux build environment installed with the requirements needed to build the OpenWrt. I use Ubuntu 14.04.    In the market, there are 4 types of Raspberry Pi boards. Type 1, 2, 3 and Zero. OpenWrt's latest trunk branch supports all of these types. I will go through trunk.    Check out the source:    git clone git://github.com/openwrt/openwrt.git    This will create a directory called "openwrt" in your current directory. Go inside this directory and run make package/symlinks command.    msezgin@ubuntu:~/openwrt/make package/symlinks    Now, we need to do the configuration before starting the build.    msezgin@ubuntu:~/openwrt/make menuconfig  ...








Read more