plans: a device that accepts power has one inlet, not one per port #2129

Merged
jaylouisw merged 1 commit from plan/1119-0937 into main 2026-11-19 09:47:15 +00:00
Owner

Plan only - no code, no engine/test/workflow edits.

docs/plans/poe-pd-inference.md, measured on 0af1f392:

  • derivePortHardware() records poeOutDeclared (engine/media.js:1089) so a guessed PSE port can be told from a declared one. There is no poeInDeclared anywhere in the tree, so the same question cannot be asked of a powered device.
  • media.js:1104-1107 applies a device-level claim to every copper port. isPoePd() (:1147) returns true on the poe-pd capability, on category wireless|ptp|camera, or on a model regex containing unifi - none of which names a socket.
  • Measured through the engine over all of devices/: 1,661 PD ports on 649 devices; 251 declared, 1,410 inferred (85%); 99 devices hold three or more PD ports purely from the guess - USW-Pro-48 48 ports / 621.6 W, ECS-48S 49 / 634.6 W, UCG-Ultra 5 / 64.75 W, RB1100Dx4 13 / 168.3 W.
  • The guess is load-bearing: simulator.js:2589 charges the draw against the PSE, :2602 sets poeUnpowered, and :3054/:4106 then drop frames and refuse connections with E_POE.
  • The only way to say "no" is "poeIn": false, which six merged files already carry while docs/DEVICE_SCHEMA.md:204 says the field is "a number".
  • test/poe.test.js:21 uses UCG-Ultra as its "real powered device". Against #2124's branch (catalogue/1119-0852) that suite is red on and the PD declares what it draws (false W) - verified by running it.

Eight ordered steps, S/M each: the missing flag, sanctioning false, capping an inferred PD at one inlet (the one step that changes an answer), the 99 rows ten per PR, repairing the suite, a PD census, the two operator surfaces, then the gate. Closes no open issue directly; #2124 stops being red at step 5.

Plan only - no code, no engine/test/workflow edits. `docs/plans/poe-pd-inference.md`, measured on `0af1f392`: - `derivePortHardware()` records `poeOutDeclared` (`engine/media.js:1089`) so a guessed PSE port can be told from a declared one. There is **no `poeInDeclared`** anywhere in the tree, so the same question cannot be asked of a powered device. - `media.js:1104-1107` applies a *device-level* claim to every copper port. `isPoePd()` (`:1147`) returns true on the `poe-pd` capability, on category `wireless|ptp|camera`, or on a model regex containing `unifi` - none of which names a socket. - Measured through the engine over all of `devices/`: **1,661 PD ports on 649 devices; 251 declared, 1,410 inferred (85%); 99 devices hold three or more PD ports purely from the guess** - `USW-Pro-48` 48 ports / 621.6 W, `ECS-48S` 49 / 634.6 W, `UCG-Ultra` 5 / 64.75 W, `RB1100Dx4` 13 / 168.3 W. - The guess is load-bearing: `simulator.js:2589` charges the draw against the PSE, `:2602` sets `poeUnpowered`, and `:3054`/`:4106` then drop frames and refuse connections with `E_POE`. - The only way to say "no" is `"poeIn": false`, which six merged files already carry while `docs/DEVICE_SCHEMA.md:204` says the field is "a **number**". - `test/poe.test.js:21` uses `UCG-Ultra` as its "real powered device". Against #2124's branch (`catalogue/1119-0852`) that suite is red on `and the PD declares what it draws (false W)` - verified by running it. Eight ordered steps, S/M each: the missing flag, sanctioning `false`, capping an inferred PD at one inlet (the one step that changes an answer), the 99 rows ten per PR, repairing the suite, a PD census, the two operator surfaces, then the gate. Closes no open issue directly; #2124 stops being red at step 5.
plans: the PD side of PoE inference has no flag, so 1,410 ports are guessed
All checks were successful
CI / Classify the change (pull_request) Successful in 10s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 9s
CI / Test (Node 20) (pull_request) Successful in 52s
CI / Test (Node 18) (pull_request) Successful in 56s
CI / Test (Node 22) (pull_request) Successful in 54s
CI / Server smoke test (pull_request) Successful in 9s
CI / Docker build (pull_request) Successful in 21s
CI / Arm auto-merge (pull_request) Successful in 6s
c7ede70300
forgejo-actions scheduled this pull request to auto merge when all checks succeed 2026-11-19 09:45:06 +00:00
jaylouisw deleted branch plan/1119-0937 2026-11-19 09:47:16 +00:00
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge!2129
No description provided.