pflua
pflua
is a high-performance network packet filtering library written
in Lua. It supports filters written in
pflang, the
filter language of the popular
tcpdump
tool. It's really fast: to our knowledge, it's the fastest pflang
implementation out there, by a wide margin. Read on for more details.
Getting started
$ git clone --recursive https://github.com/Igalia/pflua.git
$ cd pflua; make # Builds embedded LuaJIT
$ make check # Run builtin basic tests
Using pflua
Pflua is a library; you need an application to drive it.
The most simple way to use pflua is filtering packets from a file
captured by tcpdump
. For example:
$ cd tools
$ ../deps/luajit/usr/local/bin/luajit pflua-filter \
../tests/data/v4.pcap /tmp/foo.pcap "ip"
Filtered 43/43 packets from ../tests/data/v4.pcap to /tmp/foo.pcap.
See the source of pflua-filter for more information.
Pflua was made to be integrated into the Snabb Switch user-space networking toolkit, also written in Lua. A common deployment environment for Snabb is within the host virtual machine of a virtualized server, with Snabb having CPU affinity and complete control over a high-performance 10Gbit NIC, which it then routes to guest VMs. The administrator of such an environment might want to apply filters on the kinds of traffic passing into and out of the guests. To this end, we plan on integrating pflua into Snabb so as to provide a pleasant, expressive, high-performance filtering facility.
Given its high performance, it is also reasonable to deploy pflua on gateway routers and load-balancers, within virtualized networking appliances.
Implementation
Pflua can compile pflang filters in two ways.
The default compilation pipeline is pure Lua. First, a custom parser produces a high-level AST of a pflang filter expression. This AST is lowered to a primitive AST, with a limited set of operators and ways in which they can be combined. This representation is then exhaustively optimized, folding constants and tests, inferring ranges of expressions and packet offset values, hoisting assertions that post-dominate success continuations, etc. We then lower to A-normal form to give names to all intermediate values, perform common subexpression elimination, then inline named values that are only used once. We lower further to Static single assignment to give names to all blocks, which allows us to perform control-flow optimizations. Finally, we residualize Lua source code, using the control flow analysis from the SSA phase.
The resulting Lua function is a predicate of two parameters: the packet
as a uint8_t*
pointer, and its length. If the predicate is called
enough times, LuaJIT will kick in and optimize traces that run through
the function. Pleasantly, this results in machine code whose structure
reflects the actual packets that the filter sees, as branches that are
never taken are not residualized at all.
The other compilation pipeline starts with bytecode for the Berkeley
packet filter
VM.
Pflua can load up the libpcap
library and use it to compile a pflang
expression to BPF. In any case, whether you start from raw BPF or from
a pflang expression, the BPF is compiled directly to Lua source code,
which LuaJIT can gnaw on as it pleases.
We like the independence and optimization capabilities afforded by the
native pflang pipeline. However, though pflua does a good job in
implementing pflang, it is inevitable that there may be bugs or
differences of implementation relative to what libpcap
does. For that
reason, the libpcap
-to-bytecode pipeline can be a useful alternative
in some cases.
See the doc subdirectory for some examples of the Lua code generated for some simple pflang filters using these two pipelines.
Performance
To our knowledge, pflua is the fastest implementation of pflang out there. See https://github.com/Igalia/pflua-bench for our benchmarking experiments and results.
Pflua can beat other implementations because:
-
LuaJIT trace compilation results in machine code that reflects the actual traffic that your application sees
-
Pflua can hoist and eliminate bounds checks, whereas BPF is obligated to check that every packet access is valid
-
Pflua can work on data in network byte order, whereas BPF must convert to host byte order
-
Pflua takes advantage of LuaJIT's register allocator and excellent optimizing compiler, whereas e.g. the Linux kernel JIT has a limited optimizer
API documentation
None yet. See
pf.lua for the
high-level compile_filter
interface.
Bugs
Check our issue tracker for known bugs, and please file a bug if you find one. Cheers :)
Authors
Pflua was written by Katerina Barone-Adesi, Andy Wingo, Diego Pino, and Javier Muñoz at Igalia, S.L., as well as Peter Melnichenko. Development of pflua was supported by Luke Gorrie at Snabb Gmbh, purveyors of fine networking solutions. Thanks, Snabb!
Feedback is very welcome! If you are interested in pflua in a Snabb
context, probably the best thing is to post a message to the
snabb-devel
group. Or, if you like, you can contact Andy directly at
[email protected]
. If you have a problem that pflua can help solve,
let us know!