Open Wireless Link (OWL) is an open implementation of the Apple Wireless Direct Link (AWDL) ad hoc protocol for Linux and macOS written in C and part of the Open Wireless Link project.
OWL runs in user space and makes use of Linux’s Netlink API for Wi-Fi specific operations such as channel switching and to integrate itself in the Linux networking stack by providing a virtual network interface such that existing IPv6-capable programs can use AWDL without modification.
You can use OWL to in combination with OpenDrop to enable Apple AirDrop functionality on Linux-based systems such as a Raspberry Pi 3.
OWL is experimental software and is the result of reverse engineering efforts by the Open Wireless Link project. Therefore, it does not support all features of AWDL or might be incompatible with future AWDL versions. OWL is not affiliated with or endorsed by Apple Inc. Use this code at your own risk.
Wi-Fi card with active monitor mode. To use OWL, you will need a Wi-Fi card supporting active monitor mode with frame injection. We recommend the Atheros AR9280 chip (IEEE 802.11n) which we used to develop and test this code. Configurations that do not support active monitor mode, i.e., ACK received frames, might suffer from throughput degradation because the sender will re-transmit each frame up to 7 times as per the IEEE 802.11 standard. Have a look at issue #9 if you want to find out whether your card meets the requirements.
No support for virtual machines and WSL. OWL requires direct access to the Wi-Fi card which means that virtualized environments (except with USB or PCIe passthrough) and Windows Subsystem for Linux (WSL, see issue #8), are not supported.
If you're an Arch Linux user, then you can install owl
from the
owlink-git
AUR package, e.g.,
yay -S owlink-git
The project is build using CMake. To build the project, simply clone this repository in <OWLDIR>
and run
cd <OWLDIR>
git submodule update --init
mkdir build
cd build
cmake ..
make
sudo make install
OWL requires libpcap
for frame injection and reception, libev
for event handling, and libnl
(Linux only) for interactions with the system's networking stack which have to be installed on the target system.
On Debian Linux,
sudo apt install libpcap-dev libev-dev libnl-3-dev libnl-genl-3-dev libnl-route-3-dev
On Fedora Linux,
sudo dnf install libpcap-devel libev-devel libnl3-devel
On macOS, you need to add support for tun/tap devices, e.g., via tuntaposx. You can install everything via Homebrew:
brew install libpcap libev
brew cask install tuntap
Simply run
sudo owl -i <WLAN_IFACE>
You may increase the log level with -v
and -vv
and daemonize the program with -D
. For other options, have a look at daemon/owl.c
.
When started, OWL creates a virtual network interface awdl0
with a link-local IPv6 address. Discovered AWDL peers are automatically added (and removed) to (and from) the system's neighbor table. Run ip n
to see a list of all current neighbors.
The following figure shows the core structure of OWL.
We provide a coarse structure of the most important components and files to facilitate navigating the code base.
daemon/
Contains the active components that interact with the system.core.{c,h}
Schedules all relevant functions on the event loop.io.{c,h}
Platform-specific functions to send and receive frames.netutils.{c,h}
Platform-specific functions to interact with the system's networking stack.owl.c
Containsmain()
and sets up thecore
based on user arguments.
googletest/
The runtime for running the tests.radiotap/
Library for parsing radiotap headers.src/
Contains platform-independent AWDL code.channel.{c,h}
Utilities for managing the channel sequence.election.{c,h}
Code for running the election process.frame.{c,h}
The corresponding header file contains the definitions of all TLVs.peers.{c,h}
Manages the peer table.rx.{c,h}
Functions for handling a received data and action frames including parsing TLVs.schedule.{c,h}
Functions to determine when and which frames should be sent.state.{c,h}
Consolidates the AWDL state.sync.{c,h}
Synchronization: managing (extended) availability windows.tx.{c,h}
Crafting valid data and action frames ready for transmission.version.{c,h}
Parse and create AWDL version numbers.wire.{c,h}
Mini-library for safely reading and writing primitives types from and to a frame buffer.
tests/
The actual test cases for this repository.README.md
This file.
- OWL uses static election metric and counter values, so it either takes part as a slave (low values) or wins the election (high values). See
AWDL_ELECTION_METRIC_INIT
andAWDL_ELECTION_COUNTER_INIT
and ininclude/election.h
. - The channel sequence does not adjust itself automatically to current network load and/or other triggers. This would require a better understanding of Apple's implementation. Currently, the channel sequence is fixed when initializing. See
awdl_chanseq_init_static()
insrc/state.{c,h}
. - OWL does not allow a concurrent connection to an AP. This means, that when started, the Wi-Fi interface exclusively uses AWDL. To work around this, OWL could create a new monitor interface (instead of making the Wi-Fi interface one) and adjust its channel sequence to include the channel of the AP network.
-
Milan Stute, Sashank Narain, Alex Mariotto, Alexander Heinrich, David Kreitschmann, Guevara Noubir, and Matthias Hollick. A Billion Open Interfaces for Eve and Mallory: MitM, DoS, and Tracking Attacks on iOS and macOS Through Apple Wireless Direct Link. 28th USENIX Security Symposium (USENIX Security ’19), August 14–16, 2019, Santa Clara, CA, USA. Link
-
Milan Stute, David Kreitschmann, and Matthias Hollick. One Billion Apples’ Secret Sauce: Recipe for the Apple Wireless Direct Link Ad hoc Protocol. The 24th Annual International Conference on Mobile Computing and Networking (MobiCom '18), October 29–November 2, 2018, New Delhi, India. ACM. doi:10.1145/3241539.3241566
-
Milan Stute, David Kreitschmann, and Matthias Hollick. Demo: Linux Goes Apple Picking: Cross-Platform Ad hoc Communication with Apple Wireless Direct Link. The 24th Annual International Conference on Mobile Computing and Networking (MobiCom '18), October 29–November 2, 2018, New Delhi, India. ACM. doi:10.1145/3241539.3267716