While troubleshooting or developing new features in OVS, a considerable
amount of time is spent analyzing flows (whether that's Openflow flows
or datapath flows). Currently, OVS has tools to dump flows with
different levels of verbosity as well as to filter flows prior to
dumping them, e.g: 'ovs-ofctl dump-flows', 'ovs-appctl
dpctl/dump-flows', etc.
The output of these commands is considered stable so it should be
possible to write more layered tools that enable advanced flow analysis.
However, the way flows are formatted into strings is not trivial to
parse.
This series introduces the a flow parsing library capable of parsing
both Openflow and DPIF flows.
The library is based on generic key-value and list parsers and a number
of common decoders. Based on that, an Openflow flow parser and a DPIF
flow parser are introduced by defining a way to decode each possible
field and action they might contain.
The library has the following features:
- Parsed key-value pairs keep some metadata that refer to the original
strings they were extracted from. That way the flows can be printed
and formatted in flexible ways.
- It includes a basic flow filtering mechanism. A filter can be defined
combining logical (||, &&, !), arithmetical (<, >, =) or mask (~=)
operations
- It supports IPAddress and Ethernet masking (based on netaddr)
- The decoder to use for each key (match or action) is set explicitly to
avoid expensive runtime type-guessing.
- The decoders to use for Openflow fields is automatically generated
based on meta-flow.h
- Additional dependencies:
- netaddr: For IP and Ethernet Address management.
- pyparsing: For filtering syntax.
- pytest: For unit tests.
One key goal challenge of including this library is avoiding diversion
between the C code that prints/parses the flows and the python parsing
code. To that effect, the series introduces the following mechanisms:
- Decoding information of openflow fields is automatically generated
based on meta-flow.h
- The calls to ovs-ofctl made from tests/ofp-actions.at are wrapped by a
python script that also runs the python parsers. If an exception is
raised by the python code (meaning it was not capable of parsing the
flow string), the test will fail
- The calls to the test-odp made from tests/odp.at are wrapped by a
python script that also runs the python parsers. If an exception is
raised by the python code (meaning it was not capable of parsing the
flow string), the test will fail.
- A python unit test testsuite ensures python code works and it's easy
to add more flow examples to it
- A dependency check is introduced. The python parsing code mainly
depends on lib/ofp-actions.c and lib/odp-util.c. This series stores
the md5sum of such files and adds a build target that ensures the
files have not been changed. That way, anyone who modifies those files
will get a warning the fist time they build the project. Dependency
digests are easily updated using a string so hopefully this warning
would not be too inconvenient.
Library usage
-------------
from ovs.flows.ofp import OFPFlow
flow = OFPFlow("cookie=0x2b32ab4d, table=41, n_packets=11, n_bytes=462,
priority=33,ip,reg15=0x2/0x2,nw_src=10.128.0.2/24
actions=move:NXM_OF_TCP_DST[]->NXM_NX_XXREG0[32..47],ct(table=16,zone=NXM_NX_REG13[0..15],nat)")
flow.info
{'cookie': 724740941, 'table': 41, 'n_packets': 11, 'n_bytes': 462}
flow.match
{'priority': 33, 'ip': True, 'reg15': Mask32('0x2/0x2'), 'nw_src':
IPMask('10.128.0.2/24')}
flow.actions
[{'move': {'src': {'field': 'NXM_OF_TCP_DST'}, 'dst': {'field':
'NXM_NX_XXREG0', 'start': 32, 'end': 47}}}, {'ct': {'table': 16, 'zone':
{'field': 'NXM_NX_REG13', 'start': 0, 'end': 15}, 'nat': True}}]
from ovs.flows.filter import OFFilter
filt = OFFilter("nw_src ~= 10.128.0.10 and (table = 42 or n_packets > 0)")
filt.evaluate(flow)
True
V4 -> V5:
- rename tests/test-ovs-{dp,of}parse.py to tests/test-{dp,of}parse.py
- add dependencies to netaddr and pyparsing.
In pip: add a new extras_requirement called "flow"
In packages: add them in the Suggests section
In CI: install them in linux-prepare.sh
V3 -> V4:
- renamed the library ovs/flows -> ovs/flow. Note I've kept Acked-by
tags for patches where only this change was done.
- unit tests: instead of wrapping ovs-ofctl/test-odp tools, just add
another test to the correspondent *.at files as per Ilya's suggestion.
- Dropped patch 13.
- Addressed Ilya's comments on unit test checks (old patch 14/18, new
patch 13/17).
V2 -> V3:
- Simplified KV and list decoding code (Mark Michelson's suggestion)
- Fixed typos
- Added missing files to FLAKE8_PYFILES
- Go back to a simplified ipv4/6 regexp for ip-port range extraction.
Also added specific unit test for ip-port range decoding.
- Adapt ofp encap() action decoder to support new header types: mpls and mpls_mc
(the need for change was detected by patch 13)
V1 -> V2:
- list/kv parsers: changed the API to accept the string to parse in the
constructor.
- list/kv parsers: allow re.split() to return less than 3 elements
(enables support for python 3.6).
- decoders: add a more accurate IPv6 regexp and remove confusing max()
in IPMask and EthMask.
- odp/ofp flows: remove the *Factory class and implement caching of
decoders using class variables.
- odp/ofp flows: homogenize names of functions and made them static.
- moved pytest unit tests from a build target to a testsuite and added
their requirements to python/test_requirements.txt.
- Formatting fixes and missing dots (lots of them!).
RFC -> V1:
- filters: created a class to represent the filtering result. That way
we can store more information such as what key actually triggered the
match. This enables functionality such as highlighting of keys based
on an expression
- Formatted python code according to flake8 requirements
- Split ofp actions in ofp_act.py
- drop ofparse utility (will send a RFC to the mail list soon)
- Moved the initialization of the decoders objects to a factory so they
are cached. This significantly decreases parsing time of large dumps.
Adrian Moreno (17):
python: add generic Key-Value parser
python: add mask, ip and eth decoders
python: add list parser
build-aux: split extract-ofp-fields
build-aux: generate ofp field decoders
python: add flow base class
python: introduce OpenFlow Flow parsing
python: add ovs datapath flow parsing
python: add flow filtering syntax
python: add a json encoder to flow fields
tests: verify flows in ofp-actions are parseable
tests: verify flows in odp.at are parseable
python: introduce unit tests
python: add unit tests for ListParser
python: add unit tests for openflow parsing
python: add unit tests to datapath parsing
python: add unit tests for filtering engine
