[josephnef]

Second item of #566. The M6 issue asked for tests/ran_handover.sh. The M6 rig tests are testrig's scripts/m6_*.sh, because they need the rig. testrig consumes devourer (its tests/ and build/), never the reverse, so a devourer entry point that calls testrig would invert that. Instead docs/cell-m6.md records where the suite lives, the verdict convention, and what each script asserts. devourer keeps the headless part (ctest cell_m6).

Docs only.

🤖 Generated with Claude Code

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can start a comment with 'qodo' or '@qodo' to chat about any finding

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Document where the M6 rig suite lives

📝 Documentation 🕐 Less than 10 minutes

AI Description

• Identify testrig as the home of hardware-dependent M6 tests without reversing the repository
 dependency.
• List each rig script’s assertions, verdict convention, and results location.
• Clarify that devourer retains the headless cell_m6 test.

Diagram

graph TD
  Scripts["M6 rig scripts"] --> Coordinator["M5 coordinator"] --> Rig["Physical rig"]
  Scripts --> Build["Devourer build"]
  Scripts --> Tests["Devourer tests"] --> Ctest["Headless cell_m6"]

Loading

High-Level Assessment

Documenting the existing testrig suite is appropriate because its tests require hardware. A devourer wrapper for the requested script was considered but would reverse the established dependency; the existing headless ctest remains in devourer.

Files changed (1) +25 / -0

Documentation (1) +25 / -0

cell-m6.mdLocate and catalog the M6 rig suite +25/-0

Locate and catalog the M6 rig suite

• Adds a section explaining why hardware-dependent M6 tests live in testrig rather than under devourer’s tests. It catalogs script assertions and records their verdict convention, coordinator, headless counterpart, and results location.

docs/cell-m6.md

[josephnef]

Added the second #566 item to this docs PR: the 802.11 roam baseline, measured on the rig (testrig m6_roam_baseline.sh). It's an ordinary roam on the same radios as the M6 handover: rtw89, two hostapd APs (WPA2-PSK), a wpa_supplicant client, 10 forced roams per run, three channel layouts. The KPI is M6's, the DL gap across the switch.

channels	DL gap per roam	returning to the first AP
149/157	379–598 ms	5600–5838 ms
157/149	435–663 ms	5729–5804 ms
149/149	415–542 ms	5671–5865 ms

The M6 switch costs 84–123 ms on the same radios. The 5.6–5.9 s return is reproducible: the first AP's authentication response goes unacknowledged three times, then the supplicant retries. The cause is not established, and the docs report it as this stack's behaviour, not a general Wi-Fi figure.

[josephnef]

Qodo: no issues found. Adding skip-qodo-gate.