14 releases

0.3.8 Feb 20, 2024
0.3.7 Feb 5, 2024
0.3.6 Jan 30, 2024
0.3.1 Jun 29, 2023
0.1.3 Apr 29, 2023

#15 in #wireless

MIT/Apache

115KB
3K SLoC

AWDL frame parser

AWDL(Apple Wireless Direct Link) is a protocol, for wireless P2P communication. AWDL is designed to be low latency and high bandwidth, while requiring only one chip for AWDL and "normal" WiFi. This parser can read and write the network frames used by this protocol. For further information see this stackoverflow post.

Support

Frame types

  • action
    • MIF
    • PSF
  • data(coming soon)

TLV

  • Arpa
  • Version
  • Synchronization parameters
  • Channel sequence
  • Election parameters
  • Election parameters v2
  • Synchronization tree
  • Data path state
  • HT capabilities
  • IEEE 802.11 Container
  • Service parameters
  • Service response

Features

The parser was designed to be able to run in low flash environments(i.e. ESP32, wasm). The listed time complexity refers to, if the parser runtime rises with input length. So O(1) means go ahead and O(n) means think before you parse. O(1) parsers are not benchmarked.

TLV feature Worst case time complexity
Arpa dns_sd_tlvs O(n)
Service Parameters dns_sd_tlvs O(n)
Service Respone dns_sd_tlvs min. O(n)
Channel Sequence sync_elect_tlvs O(2n)
ElectionParameters[V2] sync_elect_tlvs O(1)
Synchronization Parameters sync_elect_tlvs O(1)
Synchronization Tree sync_elect_tlvs O(n)
HTCapabilities data_tlvs O(1)
IEEE80211 Container data_tlvs O(n)
Data Path State data_tlvs O(n)
Version version_tlv O(1)

Do note please, that although the parsers are not yet present, the features are. Also some parsers were bundled into one feature. (For more information on this refer to Milan Stute's dissertation.)

no_std

The library doesn't require any allocations, due to the author sacrificing parts of his sanity, to use Iterators everywhere. Allocations are only used for testing, to verify that the reported sizes match reality.

Credits

Although the actual parser was written by me, the reverse engineering of the AWDL protocol was conducted by Milan Stute and SeeMoo-Lab. So kudos to them...

Dependencies

~1–1.7MB
~34K SLoC