fleet/ohm: pkgrel=6 — per-series converged (closes ka#29, includes besser#22 / #25 fixes) #37

2026-05-21T10:24:13Z

marfrit commented

2026-05-21 10:24:13 +00:00

Summary

fleet/ohm.yaml updated to point at the converged per-series patchset. Closes ka#29 (per-series reconstruction).

Two new includes:

driver/bes2600/tx-sdio-dma-oob-danctnix/ — already on disk from ka#17 but not previously included. The cumulative-c5x-danctnix shipped in pkgrel=3 did NOT have this bounce-buffer fix; KFENCE caught the OOB during pkgrel=4 soak; pkgrel=5 added it. Closes the regression surfaced in besser#22.
driver/bes2600/join-confirm-reset-danctnix/ (NEW scope) — cw1200 ancestor port (sta.c:1339-1344) with bes2600-specific PASSIVE-gate compensation. Closes besser#25. Verified on pkgrel=6 (srcversion 0E16463F): wsm_join_confirm cascade gone, plus a bonus finding — the periodic ~600 ms latency jitter on ohm was the same root cause and also went to zero.

Verification

Phase 7 on ohm passed user-acceptance (8h gate yanked early — fix is unambiguously working):

Trigger condition (PREV_AUTH_NOT_VALID deauth) observed within 1.5 min of boot
Zero wsm_join_confirm ret 1 text in subsequent logs
Zero bes2600_sdio_read_rx_batch sdio read error
Zero wifi_force_close / WARN_ON
Ping jitter: max RTT 612 ms → 13.9 ms, mdev 103.5 ms → 1.55 ms

Status

Build still hand-managed via boltzmann:~/src/besser/marfrit-besser/danctnix-besser-pkgbuild/kernel/PKGBUILD; ka-promote / ka-build template rendering still pending per the original TODOs in fleet/ohm.yaml.
cumulative-c5x-danctnix left in the manifest as historical fallback.

Upstream coordinates

bes2600-dkms (Mobian flavor): branch bes2600/wsm-join-confirm-reset, PR marfrit/bes2600-dkms#12 against cleanups
bes2600-dkms (danctnix flavor): branch bes2600/join-confirm-failure-reset top commit 3d833f8
marfrit/besser danctnix-besser-pkgbuild: noether/readme-pkgrel4-kernel-agent-flow commit df6792a5a (patches renumbered 0001-0023)

Closes #29.

## Summary fleet/ohm.yaml updated to point at the converged per-series patchset. Closes ka#29 (per-series reconstruction). Two new includes: 1. **`driver/bes2600/tx-sdio-dma-oob-danctnix/`** — already on disk from ka#17 but not previously included. The cumulative-c5x-danctnix shipped in pkgrel=3 did NOT have this bounce-buffer fix; KFENCE caught the OOB during pkgrel=4 soak; pkgrel=5 added it. Closes the regression surfaced in besser#22. 2. **`driver/bes2600/join-confirm-reset-danctnix/`** (NEW scope) — cw1200 ancestor port (sta.c:1339-1344) with bes2600-specific PASSIVE-gate compensation. Closes besser#25. Verified on pkgrel=6 (srcversion `0E16463F`): wsm_join_confirm cascade gone, plus a bonus finding — the periodic ~600 ms latency jitter on ohm was the same root cause and also went to zero. ## Verification Phase 7 on ohm passed user-acceptance (8h gate yanked early — fix is unambiguously working): - Trigger condition (`PREV_AUTH_NOT_VALID` deauth) observed within 1.5 min of boot - Zero `wsm_join_confirm ret 1` text in subsequent logs - Zero `bes2600_sdio_read_rx_batch sdio read error` - Zero `wifi_force_close` / `WARN_ON` - Ping jitter: max RTT 612 ms → 13.9 ms, mdev 103.5 ms → 1.55 ms ## Status - Build still hand-managed via `boltzmann:~/src/besser/marfrit-besser/danctnix-besser-pkgbuild/kernel/PKGBUILD`; ka-promote / ka-build template rendering still pending per the original TODOs in `fleet/ohm.yaml`. - `cumulative-c5x-danctnix` left in the manifest as historical fallback. ## Upstream coordinates - bes2600-dkms (Mobian flavor): branch `bes2600/wsm-join-confirm-reset`, PR marfrit/bes2600-dkms#12 against `cleanups` - bes2600-dkms (danctnix flavor): branch `bes2600/join-confirm-failure-reset` top commit `3d833f8` - marfrit/besser danctnix-besser-pkgbuild: `noether/readme-pkgrel4-kernel-agent-flow` commit `df6792a5a` (patches renumbered 0001-0023) Closes #29.

marfrit added 1 commit 2026-05-21 10:24:14 +00:00

fleet/ohm: pkgrel=6 — per-series converged with tx-sdio-dma-oob + join-confirm-reset 3d15c5367d

Two additions to fleet/ohm.yaml's includes for the bes2600 driver scope:

1. driver/bes2600/tx-sdio-dma-oob-danctnix/ — already on disk from
   ka#17 but not previously included. The cumulative-c5x-danctnix
   shipped in pkgrel=3 did NOT have this fix; pkgrel=4 per-series
   regressed because the staging-prep series was excluded. KFENCE
   caught the OOB during pkgrel=4 soak; pkgrel=5 included it.

2. driver/bes2600/join-confirm-reset-danctnix/ — NEW scope.
   cw1200 ancestor port (sta.c:1339-1344) with bes2600-specific
   PASSIVE-gate compensation in bes2600_unjoin_work. Closes
   besser#25. Verified pkgrel=6 srcversion 0E16463F: cascade gone,
   periodic ~600ms latency jitter also gone (same root cause).

Status note: per-series reconstruction is now converged. The
cumulative-c5x-danctnix entry is left as historical fallback;
ka#29's blocker (per-series mirrors not applying cleanly) was
resolved by manually reconstructing the per-series in
marfrit/bes2600-dkms bes2600/join-confirm-failure-reset (top
commit 3d833f8).

Build still hand-managed via boltzmann:~/src/besser/marfrit-besser/
danctnix-besser-pkgbuild/kernel/PKGBUILD; ka-promote / ka-build
template rendering still pending per the original TODOs.

Signed-off-by: Claude (noether) <claude@reauktion.de>

marfrit merged commit 84cdb5b4ee into main

2026-05-21 10:35:32 +00:00

marfrit deleted branch noether/ohm-pkgrel6-perseries-converged

2026-05-21 10:35:32 +00:00

marfrit referenced this issue from a commit

2026-05-21 10:35:34 +00:00

Merge pull request 'fleet/ohm: pkgrel=6 — per-series converged (closes ka#29, includes besser#22 / #25 fixes)' (#37) from noether/ohm-pkgrel6-perseries-converged into main

marfrit commented

2026-05-21 11:06:25 +00:00

Update: pkgrel=6 reproducible from manifest alone

Added commit 98ca36e which restructures the includes to make ka-promote ohm produce a deterministic, reproducible cumulative — without reaching back into the besser repo.

Change

New scope driver/bes2600/cumulative-pkgrel6-danctnix/ — single squashed diff equivalent to the 22-commit pkgrel=6 source state
Dropped from manifest: cumulative-c5x-danctnix (pkgrel=3 era), scs-arm-neon-build-fix (removed in pkgrel=4), queue-pending-record-lock-bh-danctnix (now in cumulative), tx-sdio-dma-oob-danctnix (now in cumulative), join-confirm-reset-danctnix (now in cumulative)

Resulting fleet/ohm.yaml includes just two entries:

  - driver/bes2600/cumulative-pkgrel6-danctnix/
  - driver/bes2600/scan-filter-5ghz-danctnix/

Verification

cd /tmp/verify-pkgrel6
tar -xf linux-7.0.tar.xz
zstd -dc v7.0-danctnix1.patch.zst | patch -sNp1
patch -sNp1 < ~/src/kernel-agent/build/ohm/v7.0-danctnix1/cumulative.patch
diff -rq drivers/staging/bes2600 ~/src/besser/.../pkgrel=6/src/.../drivers/staging/bes2600

Result: only build artifacts (.o, .cmd, .ko, .mod) differ. All source files bit-identical to the kernel running on ohm.

Effect

kernel-agent's ka-promote ohm is now the authoritative entry point for reproducing the ohm-live source state. The PKGBUILD template rendering (the template_at: TBD-... TODO in the manifest) is the next blocker for fully automating the pipeline.

Closes #29.

## Update: pkgrel=6 reproducible from manifest alone Added commit `98ca36e` which restructures the includes to make `ka-promote ohm` produce a deterministic, reproducible cumulative — without reaching back into the besser repo. ### Change - **New scope** `driver/bes2600/cumulative-pkgrel6-danctnix/` — single squashed diff equivalent to the 22-commit pkgrel=6 source state - Dropped from manifest: `cumulative-c5x-danctnix` (pkgrel=3 era), `scs-arm-neon-build-fix` (removed in pkgrel=4), `queue-pending-record-lock-bh-danctnix` (now in cumulative), `tx-sdio-dma-oob-danctnix` (now in cumulative), `join-confirm-reset-danctnix` (now in cumulative) Resulting `fleet/ohm.yaml` includes just two entries: ``` - driver/bes2600/cumulative-pkgrel6-danctnix/ - driver/bes2600/scan-filter-5ghz-danctnix/ ``` ### Verification ``` cd /tmp/verify-pkgrel6 tar -xf linux-7.0.tar.xz zstd -dc v7.0-danctnix1.patch.zst | patch -sNp1 patch -sNp1 < ~/src/kernel-agent/build/ohm/v7.0-danctnix1/cumulative.patch diff -rq drivers/staging/bes2600 ~/src/besser/.../pkgrel=6/src/.../drivers/staging/bes2600 ``` Result: only build artifacts (`.o`, `.cmd`, `.ko`, `.mod`) differ. All source files bit-identical to the kernel running on ohm. ### Effect kernel-agent's `ka-promote ohm` is now the authoritative entry point for reproducing the ohm-live source state. The PKGBUILD template rendering (the `template_at: TBD-...` TODO in the manifest) is the next blocker for fully automating the pipeline. Closes #29.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: marfrit/kernel-agent#37