Conversation
- TrexTestRun: optional pcap argument restricting TRex traffic to a single profile pcap (set_pcap on BaseTrexClientManager) - STF profile is rebuilt and re-uploaded when the pcap changes - add generated MTU test pcaps (flow A at IP length 1500/1501/7000/ 9000/9001, 1000 packets each) as LFS objects
TRex infrastructure for functional tests that need exact packet counts per replayed pcap: * TrexTestRun.execute(pcap=, single_pass=) - select the pcap and optionally disable the duration fallback per execution, so the traffic volume is bounded by the pcap content, not by time * set_props accepts duration=None -> STL single-pass: every pcap is transmitted exactly once instead of looping for a duration; ASTF and STF still require a duration (explicit guards) * _selectable_pcaps snapshot taken at the end of __init__ so that set_pcap can re-select any profile pcap later, not just the ones remaining after the previous narrowing-down
suricata_conf_file now falls back to a suricata.yaml placed next to the test file when no --suricata-cfg is given, letting a test pin capture-specific settings (e.g. DPDK queue/RSS config) without a dedicated command-line path.
Verify jumbo frame handling on a DPDK interface with MTU 9000 (functional_tests/mtu/). The test is parametrized by stage so each stage is a separate pytest node and runs in isolation: * stage 1 (IP len 1500): baseline, standard frames received whole * stage 2 (IP len 1501): off-by-one above the standard MTU * stage 3 (IP len 7000): mid-range jumbo, exposes undersized buffers * stage 4 (IP len 9000): exactly the configured MTU boundary * stage 5 (IP len 9001): one byte above the MTU - frames must be rejected Delivery is asserted on decoder counters from eve-stats (pkts and ethernet frame counts; bytes as the full L2 frame length, no CRC - GET_PKT_LEN semantics). Uses a test-local suricata.yaml (mtu 9000 forcing ~9216-byte mbufs, 4 RX queues, interrupt-mode off) and the new MtuProfile (STL, all 5 pcaps at weight 1). The test forces dpdk.interfaces[0].mtu to 9000 regardless of the parametrized value.
Recreates all five MTU test pcaps (flow A at IP length 1500/1501/ 7000/9000/9001, 1000 packets each) from scapy - update STAGES and re-run when the test stages change.
There was a problem hiding this comment.
🟡 Changes recommended
The new single_pass path can propagate duration=None through runner hooks whose contract is still typed/structured as int, and the newly added test_settings.json is empty (invalid JSON).
Once you've addressed the issues Copilot identified, you can request another Copilot review.
Pull request overview
This PR adds an MTU/jumbo-frame focused functional test that replays MTU-sized pcaps against Suricata+DPDK via TRex, and extends the TRex traffic profile manager/test runner to support selecting a single pcap and doing STL “single-pass” replays.
Changes:
- Added a new parameterized MTU functional test (
functional_tests/mtu/) plus a test-local Suricata config and a pcap generator script. - Extended the TRex profile manager to (a) select a single pcap from a profile and (b) support STL replay bounded by pcap content (
duration=None). - Updated the runner and docs to allow choosing a pcap per run and documenting the MTU test workflow.
File summaries
| File | Description |
|---|---|
| util/test_runner.py | Adds pcap selection + single_pass support to the runner API (but needs contract/typing fixes). |
| README.md | Documents the new MTU functional test and how to run individual stages. |
| conftest.py | Allows using a per-test suricata.yaml when present in the test directory. |
| functional_tests/mtu/test_mtu.py | Implements the parameterized MTU test stages and assertions based on eve-stats.json. |
| functional_tests/mtu/test_settings.json | New settings file (currently empty/invalid JSON). |
| functional_tests/mtu/suricata.yaml | Test-local Suricata configuration tailored for MTU=9000 DPDK capture. |
| functional_tests/mtu/generate_pcaps.py | Script to generate the MTU-stage pcaps. |
| assets/trex/traffic_profiles/trex_client_manager.py | Adds set_pcap() and STL single-pass replay (duration=None) support. |
| assets/trex/traffic_profiles/functional_tests/mtu_profile/profile.py | New TRex profile enumerating the MTU-stage pcaps. |
| assets/trex/traffic_profiles/pcaps/mtu_flow_a_*_1000p.pcap | Adds MTU-stage pcap artifacts via Git LFS pointers. |
Review details
Suppressed comments (3)
util/test_runner.py:60
single_pass=Trueis documented as explicitly disabling the duration, but if a caller passesdurationas well the current logic keeps it. This makessingle_passineffective and also allowsduration=Noneto flow into the runner without an explicit decision.
"""Run one traffic+stats cycle.
`duration=None` falls back to `test_info.traffic_duration`; pass
`single_pass=True` to explicitly disable the duration so the traffic
is bounded by the pcap content instead of time (each pcap is
transmitted exactly once).
"""
if duration is None and not single_pass:
duration = self.test_info.traffic_duration
self._before_traffic(multiplier, duration, pcap)
util/test_runner.py:37
- With
single_pass=True,durationcan beNone, but_run_trafficcurrently types it asint. Any subclass that usesduration(sleep windows, rate calculations, etc.) could break onNone.
def _run_traffic(self, multiplier: float, duration: int, run_info: RunInfo):
"""Generate traffic. Suricata is already running."""
util/test_runner.py:104
TrexTestRunoverrides_before_traffic/_run_traffic; theirdurationannotations should match the base class now thatduration=Noneis valid for STL single-pass runs.
def _before_traffic(
self, multiplier: float, duration: int, pcap: str | None = None
):
if pcap is not None:
self.trex_client.set_pcap(pcap)
self.trex_client.set_props(multiplier, duration)
self.trex_client.prepare()
def _run_traffic(self, multiplier: float, duration: int, run_info: RunInfo):
# sample TRex's own counters at the measurement-window start
- Files reviewed: 13/14 changed files
- Comments generated: 2
- Review effort level: Lite
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Comment on lines
+31
to
+34
| def _before_traffic( | ||
| self, multiplier: float, duration: int, pcap: str | None = None | ||
| ): | ||
| """Prepare TRex before suricata starts (select pcap, reset, set_props).""" |
Comment on lines
+11
to
+14
| interface MTU). The test is parametrized by stage so a single stage can | ||
| be selected and run in isolation via its node ID, e.g. | ||
| `functional_tests/mtu/test_mtu.py::test_mtu[2]`: | ||
|
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces a new functional test for verifying jumbo frame (MTU) handling on DPDK interfaces, along with supporting infrastructure and documentation. The changes add a parameterized test that replays various packet capture files (pcaps) with increasing IP lengths to validate Suricata's ability to handle frames up to and beyond the configured MTU. The PR also includes the code to generate these pcaps, updates to the test framework to support single-pass pcap replay, and enhancements to the traffic profile manager. Below are the most important changes:
MTU Functional Test Implementation:
functional_tests/mtu/) that replays a set of pcaps with different IP lengths (1500, 1501, 7000, 9000, 9001 bytes) to verify correct jumbo frame handling and MTU enforcement in Suricata with DPDK. The test is fully parameterized and uses a test-local Suricata config.Traffic Profile and PCAP Management:
MtuProfileinassets/trex/traffic_profiles/functional_tests/mtu_profile/profile.pyto define the set of MTU test pcaps and manage their replay via TRex.mtu_flow_a_1500_1000p.pcap,mtu_flow_a_1501_1000p.pcap,mtu_flow_a_7000_1000p.pcap,mtu_flow_a_9000_1000p.pcap,mtu_flow_a_9001_1000p.pcap) to the repository.Enhancements to TRex Traffic Profile Manager:
BaseTrexClientManagerto support selecting and replaying individual pcaps from a profile, including logic for VLAN-tagged variants, and added theset_pcapmethod for dynamic pcap selection.Documentation:
README.mdwith detailed instructions and explanation of the new MTU functional test, including stage descriptions, usage examples, and notes on test behavior.Other Minor Updates:
These changes collectively enable robust testing of MTU and jumbo frame support in Suricata with DPDK, provide reusable infrastructure for pcap-driven functional tests, and improve test documentation and maintainability.