Documentation/networking/tuntap.txt

   1 Universal TUN/TAP device driver.
   2 Copyright (C) 1999-2000 Maxim Krasnyansky <max_mk@yahoo.com>
   3
   4   Linux, Solaris drivers
   5   Copyright (C) 1999-2000 Maxim Krasnyansky <max_mk@yahoo.com>
   6
   7   FreeBSD TAP driver
   8   Copyright (c) 1999-2000 Maksim Yevmenkin <m_evmenkin@yahoo.com>
   9
  10   Revision of this document 2002 by Florian Thiel <florian.thiel@gmx.net>
  11
  12 1. Description
  13   TUN/TAP provides packet reception and transmission for user space programs.
  14   It can be seen as a simple Point-to-Point or Ethernet device, which,
  15   instead of receiving packets from physical media, receives them from
  16   user space program and instead of sending packets via physical media
  17   writes them to the user space program.
  18
  19   In order to use the driver a program has to open /dev/net/tun and issue a
  20   corresponding ioctl() to register a network device with the kernel. A network
  21   device will appear as tunXX or tapXX, depending on the options chosen. When
  22   the program closes the file descriptor, the network device and all
  23   corresponding routes will disappear.
  24
  25   Depending on the type of device chosen the userspace program has to read/write
  26   IP packets (with tun) or ethernet frames (with tap). Which one is being used
  27   depends on the flags given with the ioctl().
  28
  29   The package from http://vtun.sourceforge.net/tun contains two simple examples
  30   for how to use tun and tap devices. Both programs work like a bridge between
  31   two network interfaces.
  32   br_select.c - bridge based on select system call.
  33   br_sigio.c  - bridge based on async io and SIGIO signal.
  34   However, the best example is VTun http://vtun.sourceforge.net :))
  35
  36 2. Configuration
  37   Create device node:
  38      mkdir /dev/net (if it doesn't exist already)
  39      mknod /dev/net/tun c 10 200
  40
  41   Set permissions:
  42      e.g. chmod 0666 /dev/net/tun
  43      There's no harm in allowing the device to be accessible by non-root users,
  44      since CAP_NET_ADMIN is required for creating network devices or for
  45      connecting to network devices which aren't owned by the user in question.
  46      If you want to create persistent devices and give ownership of them to
  47      unprivileged users, then you need the /dev/net/tun device to be usable by
  48      those users.
  49
  50   Driver module autoloading
  51
  52      Make sure that "Kernel module loader" - module auto-loading
  53      support is enabled in your kernel.  The kernel should load it on
  54      first access.
  55
  56   Manual loading
  57      insert the module by hand:
  58         modprobe tun
  59
  60   If you do it the latter way, you have to load the module every time you
  61   need it, if you do it the other way it will be automatically loaded when
  62   /dev/net/tun is being opened.
  63
  64 3. Program interface
  65   3.1 Network device allocation:
  66
  67   char *dev should be the name of the device with a format string (e.g.
  68   "tun%d"), but (as far as I can see) this can be any valid network device name.
  69   Note that the character pointer becomes overwritten with the real device name
  70   (e.g. "tun0")
  71
  72   #include <linux/if.h>
  73   #include <linux/if_tun.h>
  74
  75   int tun_alloc(char *dev)
  76   {
  77       struct ifreq ifr;
  78       int fd, err;
  79
  80       if( (fd = open("/dev/net/tun", O_RDWR)) < 0 )
  81          return tun_alloc_old(dev);
  82
  83       memset(&ifr, 0, sizeof(ifr));
  84
  85       /* Flags: IFF_TUN   - TUN device (no Ethernet headers)
  86        *        IFF_TAP   - TAP device
  87        *
  88        *        IFF_NO_PI - Do not provide packet information
  89        */
  90       ifr.ifr_flags = IFF_TUN;
  91       if( *dev )
  92          strncpy(ifr.ifr_name, dev, IFNAMSIZ);
  93
  94       if( (err = ioctl(fd, TUNSETIFF, (void *) &ifr)) < 0 ){
  95          close(fd);
  96          return err;
  97       }
  98       strcpy(dev, ifr.ifr_name);
  99       return fd;
 100   }
 101
 102   3.2 Frame format:
 103   If flag IFF_NO_PI is not set each frame format is:
 104      Flags [2 bytes]
 105      Proto [2 bytes]
 106      Raw protocol(IP, IPv6, etc) frame.
 107
 108   3.3 Multiqueue tuntap interface:
 109
 110   From version 3.8, Linux supports multiqueue tuntap which can uses multiple
 111   file descriptors (queues) to parallelize packets sending or receiving. The
 112   device allocation is the same as before, and if user wants to create multiple
 113   queues, TUNSETIFF with the same device name must be called many times with
 114   IFF_MULTI_QUEUE flag.
 115
 116   char *dev should be the name of the device, queues is the number of queues to
 117   be created, fds is used to store and return the file descriptors (queues)
 118   created to the caller. Each file descriptor were served as the interface of a
 119   queue which could be accessed by userspace.
 120
 121   #include <linux/if.h>
 122   #include <linux/if_tun.h>
 123
 124   int tun_alloc_mq(char *dev, int queues, int *fds)
 125   {
 126       struct ifreq ifr;
 127       int fd, err, i;
 128
 129       if (!dev)
 130           return -1;
 131
 132       memset(&ifr, 0, sizeof(ifr));
 133       /* Flags: IFF_TUN   - TUN device (no Ethernet headers)
 134        *        IFF_TAP   - TAP device
 135        *
 136        *        IFF_NO_PI - Do not provide packet information
 137        *        IFF_MULTI_QUEUE - Create a queue of multiqueue device
 138        */
 139       ifr.ifr_flags = IFF_TAP | IFF_NO_PI | IFF_MULTI_QUEUE;
 140       strcpy(ifr.ifr_name, dev);
 141
 142       for (i = 0; i < queues; i++) {
 143           if ((fd = open("/dev/net/tun", O_RDWR)) < 0)
 144              goto err;
 145           err = ioctl(fd, TUNSETIFF, (void *)&ifr);
 146           if (err) {
 147              close(fd);
 148              goto err;
 149           }
 150           fds[i] = fd;
 151       }
 152
 153       return 0;
 154   err:
 155       for (--i; i >= 0; i--)
 156           close(fds[i]);
 157       return err;
 158   }
 159
 160   A new ioctl(TUNSETQUEUE) were introduced to enable or disable a queue. When
 161   calling it with IFF_DETACH_QUEUE flag, the queue were disabled. And when
 162   calling it with IFF_ATTACH_QUEUE flag, the queue were enabled. The queue were
 163   enabled by default after it was created through TUNSETIFF.
 164
 165   fd is the file descriptor (queue) that we want to enable or disable, when
 166   enable is true we enable it, otherwise we disable it
 167
 168   #include <linux/if.h>
 169   #include <linux/if_tun.h>
 170
 171   int tun_set_queue(int fd, int enable)
 172   {
 173       struct ifreq ifr;
 174
 175       memset(&ifr, 0, sizeof(ifr));
 176
 177       if (enable)
 178          ifr.ifr_flags = IFF_ATTACH_QUEUE;
 179       else
 180          ifr.ifr_flags = IFF_DETACH_QUEUE;
 181
 182       return ioctl(fd, TUNSETQUEUE, (void *)&ifr);
 183   }
 184
 185 Universal TUN/TAP device driver Frequently Asked Question.
 186
 187 1. What platforms are supported by TUN/TAP driver ?
 188 Currently driver has been written for 3 Unices:
 189    Linux kernels 2.2.x, 2.4.x
 190    FreeBSD 3.x, 4.x, 5.x
 191    Solaris 2.6, 7.0, 8.0
 192
 193 2. What is TUN/TAP driver used for?
 194 As mentioned above, main purpose of TUN/TAP driver is tunneling.
 195 It is used by VTun (http://vtun.sourceforge.net).
 196
 197 Another interesting application using TUN/TAP is pipsecd
 198 (http://perso.enst.fr/~beyssac/pipsec/), a userspace IPSec
 199 implementation that can use complete kernel routing (unlike FreeS/WAN).
 200
 201 3. How does Virtual network device actually work ?
 202 Virtual network device can be viewed as a simple Point-to-Point or
 203 Ethernet device, which instead of receiving packets from a physical
 204 media, receives them from user space program and instead of sending
 205 packets via physical media sends them to the user space program.
 206
 207 Let's say that you configured IPX on the tap0, then whenever
 208 the kernel sends an IPX packet to tap0, it is passed to the application
 209 (VTun for example). The application encrypts, compresses and sends it to
 210 the other side over TCP or UDP. The application on the other side decompresses
 211 and decrypts the data received and writes the packet to the TAP device,
 212 the kernel handles the packet like it came from real physical device.
 213
 214 4. What is the difference between TUN driver and TAP driver?
 215 TUN works with IP frames. TAP works with Ethernet frames.
 216
 217 This means that you have to read/write IP packets when you are using tun and
 218 ethernet frames when using tap.
 219
 220 5. What is the difference between BPF and TUN/TAP driver?
 221 BPF is an advanced packet filter. It can be attached to existing
 222 network interface. It does not provide a virtual network interface.
 223 A TUN/TAP driver does provide a virtual network interface and it is possible
 224 to attach BPF to this interface.
 225
 226 6. Does TAP driver support kernel Ethernet bridging?
 227 Yes. Linux and FreeBSD drivers support Ethernet bridging.