31 releases

0.16.0 Aug 27, 2024
0.15.0 Feb 5, 2024
0.14.1 Aug 4, 2023
0.14.0 May 13, 2022
0.4.0 Mar 19, 2019

#174 in Parser implementations

Download history 2707/week @ 2024-08-19 2829/week @ 2024-08-26 2770/week @ 2024-09-02 2372/week @ 2024-09-09 1853/week @ 2024-09-16 2460/week @ 2024-09-23 2033/week @ 2024-09-30 1444/week @ 2024-10-07 1817/week @ 2024-10-14 2970/week @ 2024-10-21 4092/week @ 2024-10-28 2481/week @ 2024-11-04 3413/week @ 2024-11-11 3660/week @ 2024-11-18 2272/week @ 2024-11-25 5695/week @ 2024-12-02

15,107 downloads per month
Used in 18 crates (17 directly)

MIT/Apache

180KB
3.5K SLoC

License: MIT Apache License 2.0 Crates.io Version docs.rs Github CI Minimum rustc version

PCAP and PCAPNG parsers

This crate contains several parsers for PCAP and PCAPNG files.

Compared to other similar projects, it is designed to offer a complete support of the many possible formats (legacy pcap, pcapng, little or big-endian, etc.) and features (pcapng files with multiple sections, interfaces, and endianness) while using only safe code and without copying data (zero-copy).

The code is available on Github and is part of the Rusticata project.

The pcap format(s)

The PCAP format (files usually ending with .pcap extension) is rather trivial. The PCAPNG format (usually .pcapng extension) is much more complex: it can be composed of multiple sections, each with multiple interfaces, having different capture lengths, time precision and even endianness!

These formats are more containers than data formats: packets contain data, formatted according to its interface linktype. There are many possible linktypes, defined in the linktypes registry. Support for parsing some of them is provided using the data feature (disabled by default).

This crate provides an abstraction over these different formats.

Parsing a file

pcap-parser provides several ways of parsing pcap data. Choosing the right one is mostly driven by resources: if the input file is small, the parse_pcap and parse_pcapng functions can be used directly.

Fine-grained functions are also available, to parse specifically some block types for example. They are listed in the pcap and pcapng modules.

If the input is larger and cannot fit into memory, then streaming parsers are available. They work by iterating on blocks, and so do not require to map the entire input. They cannot seek to a specific block, however.

Note: due to PCAPNG limitations, it is not possible to directly seek in a file to get a packet and handle it: the caller has to iterate though the file and store (at least) the interface descriptions for the current section, in order of appearance.

Example: streaming parsers

The following code shows how to parse a file in the pcap-ng format, using a PcapNGReader streaming parser. This reader provides a convenient abstraction over the file format, and takes care of the endianness.

use pcap_parser::*;
use pcap_parser::traits::PcapReaderIterator;
use std::fs::File;

let file = File::open(path).unwrap();
let mut num_blocks = 0;
let mut reader = PcapNGReader::new(65536, file).expect("PcapNGReader");
loop {
    match reader.next() {
        Ok((offset, _block)) => {
            println!("got new block");
            num_blocks += 1;
            reader.consume(offset);
        },
        Err(PcapError::Eof) => break,
        Err(PcapError::Incomplete(_)) => {
            reader.refill().unwrap();
        },
        Err(e) => panic!("error while reading: {:?}", e),
    }
}
println!("num_blocks: {}", num_blocks);

See PcapNGReader for a complete example, including handling of linktype and accessing packet data.

See also the pcapng module for more details about the new capture file format.

For legacy pcap files, use similar code with the LegacyPcapReader streaming parser.

See pcap-analyzer, in particular the libpcap-tools and pcap-info modules for more examples.

Example: generic streaming parsing

To create a pcap reader for input in either PCAP or PCAPNG format, use the create_reader function.

Serialization

Support for serialization (i.e. generating binary data) is available by enabling the serialize feature. Most structures gain the to_vec() method (provided by the ToVec trait).

Note: support is still experimental, though working. API may change in the future.

Changes

See CHANGELOG.md.

License

Licensed under either of

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

Dependencies

~1MB
~20K SLoC