livox rust: bind host ports with SO_REUSEADDR, as SDK2 does by jeff-hykin · Pull Request #4240 · dimensionalOS/dimos

jeff-hykin · 2026-09-19T14:51:46Z

Don't forcefully reuse the Mid360 port (needed for R1 Pro stuff)

The vendor's livox_ros_driver2 (Livox SDK2) holds the same host ports and binds them with SO_REUSEADDR, so the C++ Point-LIO module (also SDK2) shares them and this driver, binding plainly, was the one process refused with EADDRINUSE. Bind the command and data sockets with reuse set. Also make the impostor-sender test run on macOS, where 127.0.0.2 is not on lo0: fall back to the routed interface address.

greptile-apps · 2026-09-19T15:02:48Z

Confidence Score: 4/5

Not safe to merge until the route-dependent test panic is addressed; the socket-sharing test gap is non-blocking but leaves the intended traffic behavior unverified.

Findings

Avoid route-dependent fallback ▶
Test actual socket traffic ▶

Summary

This change improves UDP socket binding compatibility and adds coverage around SDK2-style port sharing, but two test paths need attention. The SDK2 coexistence test confirms only that startup succeeds, not that handshake and sensor packets continue to flow while competing sockets remain bound. The unexpected-sender test also panics when its fallback cannot select a routed local address on an isolated runner.

_{Reviews (1) · Last reviewed commit: "livox rust: bind host ports with SO_REUS..."}

greptile-apps · 2026-09-19T15:02:51Z

+        );
+        assert!(
+            source.is_ok(),
+            "must coexist with an SDK2 binder: {:?}",
+            source.err()
+        );
+        stop.store(true, Ordering::Relaxed);


Test actual socket traffic

This test only verifies that LiveSource::start returns, then immediately stops the source. It can pass while the SDK2-style sockets receive the handshake ACKs or point and IMU datagrams instead. Keep those sockets bound through a completed fake-device handshake and assert receipt of both packet types. This is non-blocking, but without it the regression test can miss a broken data path.

Artifacts

Command output from the check

Ran the PR's existing coexistence test from the repository root; it passed but only ran the startup-and-stop path, confirming the limited scope.

Evidence from the check

Temporary executable command that adds a test retaining the three SDK2-style bindings, ACKing the full handshake, and requiring point and IMU receipt.

Command output from the check

Executed the focused validation from the repository root; it timed out waiting for a packet with competing SDK2-style sockets held, so full traffic coexistence was not demonstrated.

_{Ran code and verified through T-Rex}

greptile-apps · 2026-09-19T15:02:51Z

+        probe
+            .connect(SocketAddrV4::new(Ipv4Addr::new(10, 255, 255, 255), 9))
+            .expect("no 127.0.0.2 alias and no default route: nowhere to forge from");
+        match probe.local_addr().unwrap() {
+            std::net::SocketAddr::V4(addr) if !addr.ip().is_loopback() => *addr.ip(),
+            addr => panic!("routed address {addr} is not a usable impostor"),
+        }


Avoid route-dependent fallback

If 127.0.0.2 is unavailable and the runner has no default route, this fallback panics after UDP address selection fails, before the sender-filter assertion runs. The unexpected-sender test therefore fails on isolated or offline runners even though the production behavior is unaffected. Use a deterministic local setup or handle an unavailable route without panicking. This must be fixed before merging because it makes the test suite fail in supported isolated environments.

Artifacts

Evidence from the check

Evidence file captured while the check ran.

Command output from the check

The full command output behind this check.

Command output from the check

The full command output behind this check.

Command output from the check

The full command output behind this check.

Command output from the check

The full command output behind this check.

_{Ran code and verified through T-Rex}

greptile-apps · 2026-09-19T15:02:57Z

Comments Outside Diff

These findings sit on lines the diff does not cover, so they could not be posted inline. Each one leaves this list once its file changes.

SDK2 coexistence test does not validate usable data-plane coexistence ▶
- Bug
  - starts_while_an_sdk2_process_holds_the_host_ports passes after LiveSource::start and immediately stops. A focused run keeping identical SDK2-style competing bindings alive while ACKing the complete handshake and transmitting point/IMU packets timed out waiting for receipt.
- Cause
  - The added test neither provides a device to acknowledge the asynchronous control handshake nor calls recv; with SO_REUSEPORT, datagrams may be assigned to the competing sockets rather than this source.
- Fix
  - Extend the coexistence test with a fake device that ACKs every handshake step and sends point and IMU packets while the competing sockets remain alive; assert handshake completion and receipt of both packet types. If sharing is intentionally bind-only, narrow the test name and claim accordingly.

codecov · 2026-09-19T15:13:21Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

@@           Coverage Diff           @@
##             main    #4240   +/-   ##
=======================================
  Coverage   79.71%   79.71%           
=======================================
  Files        1508     1508           
  Lines      142607   142607           
  Branches    12209    12209           
=======================================
  Hits       113681   113681           
  Misses      25485    25485           
  Partials     3441     3441

Components	Coverage Δ
Tests	`95.41% <ø> (+<0.01%)`	⬆️

Flag	Coverage Δ
OS-ubuntu-24.04-arm	`75.35% <ø> (-0.01%)`	⬇️
OS-ubuntu-latest	`76.16% <ø> (+<0.01%)`	⬆️
Py-3.10	`75.94% <ø> (+<0.01%)`	⬆️
Py-3.11	`76.14% <ø> (-0.01%)`	⬇️
Py-3.12	`76.15% <ø> (+<0.01%)`	⬆️
SelfHosted-Large	`30.61% <ø> (ø)`
SelfHosted-Linux	`36.65% <ø> (+<0.01%)`	⬆️
SelfHosted-macOS	`36.98% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.
see 2 files with indirect coverage changes

🚀 New features to boost your workflow:

📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

aclauer · 2026-09-19T20:26:45Z

+/// over a specific one). A vendor SDK2 process -- `livox_ros_driver2` on the
+/// Galaxea R1 -- holds the same host ports, and the kernel only lets two
+/// sockets share a port when both asked for it. A plain bind is refused with
+/// `EADDRINUSE`; the C++ Point-LIO module, which binds through SDK2, never


Can we just remove this commentary about deleted code? Otherwise lgtm!

jeff-hykin deployed to cachix September 19, 2026 14:51 — with GitHub Actions Active

jeff-hykin marked this pull request as ready for review September 19, 2026 14:54

jeff-hykin requested review from Dreamsorcerer, aclauer, leshy, mustafab0, paul-nechifor and spomichter as code owners September 19, 2026 14:54

jeff-hykin added the backport:skip Skip creating a backport to any release branches label Sep 19, 2026

greptile-apps Bot reviewed Sep 19, 2026

View reviewed changes

github-actions Bot added the ready-to-merge Required CI checks have passed on this PR label Sep 19, 2026

aclauer previously approved these changes Sep 19, 2026

View reviewed changes

livox: drop the history from bind_shared's doc comment

be4947a

jeff-hykin dismissed aclauer’s stale review via be4947a September 19, 2026 20:29

github-actions Bot removed the ready-to-merge Required CI checks have passed on this PR label Sep 19, 2026

jeff-hykin deployed to cachix September 19, 2026 20:30 — with GitHub Actions Active

aclauer approved these changes Sep 19, 2026

View reviewed changes

github-actions Bot added the ready-to-merge Required CI checks have passed on this PR label Sep 19, 2026

aclauer approved these changes Sep 20, 2026

View reviewed changes

Navigation Menu

livox rust: bind host ports with SO_REUSEADDR, as SDK2 does - #4240

livox rust: bind host ports with SO_REUSEADDR, as SDK2 does#4240
jeff-hykin wants to merge 2 commits into
mainfrom
jeff/fix/livox_rust_reuse_addr

jeff-hykin commented Sep 19, 2026 •

edited

Loading

Uh oh!

greptile-apps Bot commented Sep 19, 2026 •

edited

Loading

Uh oh!

greptile-apps Bot Sep 19, 2026

Uh oh!

greptile-apps Bot Sep 19, 2026

Uh oh!

greptile-apps Bot commented Sep 19, 2026

Uh oh!

codecov Bot commented Sep 19, 2026 •

edited

Loading

Uh oh!

aclauer Sep 19, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

jeff-hykin commented Sep 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

greptile-apps Bot commented Sep 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Confidence Score: 4/5

Findings

Summary

Uh oh!

greptile-apps Bot Sep 19, 2026

Choose a reason for hiding this comment

Uh oh!

greptile-apps Bot Sep 19, 2026

Choose a reason for hiding this comment

Uh oh!

greptile-apps Bot commented Sep 19, 2026

Comments Outside Diff

Uh oh!

codecov Bot commented Sep 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

aclauer Sep 19, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

jeff-hykin commented Sep 19, 2026 •

edited

Loading

greptile-apps Bot commented Sep 19, 2026 •

edited

Loading

codecov Bot commented Sep 19, 2026 •

edited

Loading