world leader in high performance signal processing
Trace: » procfs-guide

Linux Kernel Procfs Guide

Introduction

The /proc file system (procfs) is a special file system in the linux kernel. It's a virtual file system: it is not associated with a block device but exists only in memory. The files in the procfs are there to allow userland programs access to certain information from the kernel (like process information in /proc/[0-9]+/), but also for debug purposes (like /proc/ksyms).

This guide describes the use of the procfs file system from within the Linux kernel. It starts by introducing all relevant functions to manage the files within the file system. After that it shows how to communicate with userland, and some tips and tricks will be pointed out. Finally a complete example will be shown.

Note that the files in /proc/sys are sysctl files: they don't belong to procfs and are governed by a completely different API described in the Kernel API book.

Managing procfs entries

This chapter describes the functions that various kernel components use to populate the procfs with files, symlinks, device nodes, and directories.

A minor note before we start: if you want to use any of the procfs functions, be sure to include the correct header file! This should be one of the first lines in your code:

#include <linux/proc_fs.h>
    

Creating a regular file

This function creates a regular file with the name name, file mode mode in the directory parent. To create a file in the root of the procfs, use NULL as parent parameter. When successful, the function will return a pointer to the freshly created struct proc_dir_entry; otherwise it will return NULL. describes how to do something useful with regular files.

Note that it is specifically supported that you can pass a path that spans multiple directories. For example create_proc_entry(“drivers/via0/info”) will create the via0 directory if necessary, with standard 0755 permissions.

If you only want to be able to read the file, the function create_proc_read_entry described in may be used to create and initialise the procfs entry in one single call.

Creating a symlink

This creates a symlink in the procfs directory parent that points from name to dest. This translates in userland to ln -sdestname.

Creating a directory

Create a directory name in the procfs directory parent.

Removing an entry

Removes the entry name in the directory parent from the procfs. Entries are removed by their name, not by the struct proc_dir_entry returned by the various create functions. Note that this function doesn't recursively remove entries.

Be sure to free the data entry from the struct proc_dir_entry before remove_proc_entry is called (that is: if there was some data allocated, of course). See for more information on using the data entry.

Communicating with userland

Instead of reading (or writing) information directly from kernel memory, procfs works with call back functions for files: functions that are called when a specific file is being read or written. Such functions have to be initialised after the procfs file is created by setting the read_proc and/or write_proc fields in the struct proc_dir_entry* that the function create_proc_entry returned:

struct proc_dir_entry* entry;

entry->read_proc = read_proc_foo;
entry->write_proc = write_proc_foo;
    

If you only want to use a the read_proc, the function create_proc_read_entry described in may be used to create and initialise the procfs entry in one single call.

Reading data

The read function is a call back function that allows userland processes to read data from the kernel. The read function should have the following format:

The read function should write its information into the buffer, which will be exactly PAGE_SIZE bytes long.

The parameter peof should be used to signal that the end of the file has been reached by writing 1 to the memory location peof points to.

The data parameter can be used to create a single call back function for several files, see .

The rest of the parameters and the return value are described by a comment in fs/proc/generic.c as follows:

shows how to use a read call back function.

Writing data

The write call back function allows a userland process to write data to the kernel, so it has some kind of control over the kernel. The write function should have the following format:

The write function should read count bytes at maximum from the buffer. Note that the buffer doesn't live in the kernel's memory space, so it should first be copied to kernel space with copy_from_user. The file parameter is usually ignored. shows how to use the data parameter.

Again, shows how to use this call back function.

A single call back for many files

When a large number of almost identical files is used, it's quite inconvenient to use a separate call back function for each file. A better approach is to have a single call back function that distinguishes between the files by using the data field in struct proc_dir_entry. First of all, the data field has to be initialised:

struct proc_dir_entry* entry;
struct my_file_data *file_data;

file_data = kmalloc(sizeof(struct my_file_data), GFP_KERNEL);
entry->data = file_data;
      

The data field is a void *, so it can be initialised with anything.

Now that the data field is set, the read_proc and write_proc can use it to distinguish between files because they get it passed into their data parameter:

int foo_read_func(char *page, char **start, off_t off,
                  int count, int *eof, void *data)
{
        int len;

        if(data == file_data) {
                /* special case for this file */
        } else {
                /* normal processing */
        }

        return len;
}
      

Be sure to free the data data field when removing the procfs entry.

Tips and tricks

Convenience functions

This function creates a regular file in exactly the same way as create_proc_entry from does, but also allows to set the read function read_proc in one call. This function can set the data as well, like explained in .

Modules

If procfs is being used from within a module, be sure to set the owner field in the struct proc_dir_entry to THIS_MODULE.

struct proc_dir_entry* entry;

entry->owner = THIS_MODULE;
      

Mode and ownership

Sometimes it is useful to change the mode and/or ownership of a procfs entry. Here is an example that shows how to achieve that:

struct proc_dir_entry* entry;

entry->mode =  S_IWUSR |S_IRUSR | S_IRGRP | S_IROTH;
entry->uid = 0;
entry->gid = 100;
      

Example

/*
 * procfs_example.c: an example proc interface
 *
 * Copyright (C) 2001, Erik Mouw (mouw@nl.linux.org)
 *
 * This file accompanies the procfs-guide in the Linux kernel
 * source. Its main use is to demonstrate the concepts and
 * functions described in the guide.
 *
 * This software has been developed while working on the LART
 * computing board (http://www.lartmaker.nl), which was sponsored
 * by the Delt University of Technology projects Mobile Multi-media
 * Communications and Ubiquitous Communications.
 *
 * This program is free software; you can redistribute
 * it and/or modify it under the terms of the GNU General
 * Public License as published by the Free Software
 * Foundation; either version 2 of the License, or (at your
 * option) any later version.
 *
 * This program is distributed in the hope that it will be
 * useful, but WITHOUT ANY WARRANTY; without even the implied
 * warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
 * PURPOSE.  See the GNU General Public License for more
 * details.
 * 
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place,
 * Suite 330, Boston, MA  02111-1307  USA
 *
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/jiffies.h>
#include <asm/uaccess.h>


#define MODULE_VERS "1.0"
#define MODULE_NAME "procfs_example"

#define FOOBAR_LEN 8

struct fb_data_t {
        char name[FOOBAR_LEN + 1];
        char value[FOOBAR_LEN + 1];
};


static struct proc_dir_entry *example_dir, *foo_file,
        *bar_file, *jiffies_file, *symlink;


struct fb_data_t foo_data, bar_data;


static int proc_read_jiffies(char *page, char **start,
                             off_t off, int count,
                             int *eof, void *data)
{
        int len;

        len = sprintf(page, "jiffies = %ld\n",
                      jiffies);

        return len;
}


static int proc_read_foobar(char *page, char **start,
                            off_t off, int count, 
                            int *eof, void *data)
{
        int len;
        struct fb_data_t *fb_data = (struct fb_data_t *)data;

        /* DON'T DO THAT - buffer overruns are bad */
        len = sprintf(page, "%s = '%s'\n", 
                      fb_data->name, fb_data->value);

        return len;
}


static int proc_write_foobar(struct file *file,
                             const char *buffer,
                             unsigned long count, 
                             void *data)
{
        int len;
        struct fb_data_t *fb_data = (struct fb_data_t *)data;

        if(count > FOOBAR_LEN)
                len = FOOBAR_LEN;
        else
                len = count;

        if(copy_from_user(fb_data->value, buffer, len))
                return -EFAULT;

        fb_data->value[len] = '\0';

        return len;
}


static int __init init_procfs_example(void)
{
        int rv = 0;

        /* create directory */
        example_dir = proc_mkdir(MODULE_NAME, NULL);
        if(example_dir == NULL) {
                rv = -ENOMEM;
                goto out;
        }
        /* create jiffies using convenience function */
        jiffies_file = create_proc_read_entry("jiffies", 
                                              0444, example_dir, 
                                              proc_read_jiffies,
                                              NULL);
        if(jiffies_file == NULL) {
                rv  = -ENOMEM;
                goto no_jiffies;
        }

        /* create foo and bar files using same callback
         * functions 
         */
        foo_file = create_proc_entry("foo", 0644, example_dir);
        if(foo_file == NULL) {
                rv = -ENOMEM;
                goto no_foo;
        }

        strcpy(foo_data.name, "foo");
        strcpy(foo_data.value, "foo");
        foo_file->data = &foo_data;
        foo_file->read_proc = proc_read_foobar;
        foo_file->write_proc = proc_write_foobar;
                
        bar_file = create_proc_entry("bar", 0644, example_dir);
        if(bar_file == NULL) {
                rv = -ENOMEM;
                goto no_bar;
        }

        strcpy(bar_data.name, "bar");
        strcpy(bar_data.value, "bar");
        bar_file->data = &bar_data;
        bar_file->read_proc = proc_read_foobar;
        bar_file->write_proc = proc_write_foobar;
                
        /* create symlink */
        symlink = proc_symlink("jiffies_too", example_dir, 
                               "jiffies");
        if(symlink == NULL) {
                rv = -ENOMEM;
                goto no_symlink;
        }

        /* everything OK */
        printk(KERN_INFO "%s %s initialised\n",
               MODULE_NAME, MODULE_VERS);
        return 0;

no_symlink:
        remove_proc_entry("bar", example_dir);
no_bar:
        remove_proc_entry("foo", example_dir);
no_foo:
        remove_proc_entry("jiffies", example_dir);
no_jiffies:                           
        remove_proc_entry(MODULE_NAME, NULL);
out:
        return rv;
}


static void __exit cleanup_procfs_example(void)
{
        remove_proc_entry("jiffies_too", example_dir);
        remove_proc_entry("bar", example_dir);
        remove_proc_entry("foo", example_dir);
        remove_proc_entry("jiffies", example_dir);
        remove_proc_entry(MODULE_NAME, NULL);

        printk(KERN_INFO "%s %s removed\n",
               MODULE_NAME, MODULE_VERS);
}


module_init(init_procfs_example);
module_exit(cleanup_procfs_example);

MODULE_AUTHOR("Erik Mouw");
MODULE_DESCRIPTION("procfs examples");
MODULE_LICENSE("GPL");

About This Book

Authors

  • Erik Mouw

Copyright 2001

  • Erik Mouw

Legal Notice

This documentation is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This documentation is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA

For more details see the file COPYING in the source distribution of Linux.