Lock-protected SSH now observes installed Baseline02 (compiled Linux 8584ccd,
Reborn b71b468, root v1.21, taint 0), superseding awaiting-installation wording
for identity alone. The isolated Wi-Fi pass fixes measured observer deadlines,
new-lease DNS scheduling, redundant RSSI/list/policy work and association-bound
Online proof, plus useful redacted Wi-Fi diagnostic enums/counters. Baseline
platform readiness took 27.99–32.24 s despite IP in 2.68–5.88 s; four 8 MiB
HTTP transfers passed hashes at RX 43.94–44.92 / TX 46.38–46.94 Mb/s. Physical
after measurements could not obtain the shared lock; source checks pass but
whole-platform acceptance remains PARTIAL. The checksum repair, firmware/register semantics, conservative
CAM automatic default and Bluetooth implementation remain unchanged. Exact
results and remaining physical gates belong to the
Wi-Fi optimization report.
Source tests do not close cold-boot, coexistence, electrical-power, endurance or
full-suspend acceptance. Earlier physical failures below retain their scope.
UART01 uses an independent Wi-Fi observer and bounded transport integrity alongside USB checks. Radios are off for guarded C3 entry windows. This does not qualify Bluetooth peer playback, optional Auto/ABR/codecs, radio endurance or full-system-suspend restoration. Baseline02 preserves matched firmware/provisioning and the current private experiment software; qualification and public distribution remain separate.
Baseline02 seal, UART01 hardware and current state carry exact identities, counters and limits.
The following text retains earlier dated decisions and results. Its “current”, “next” and candidate labels belong to those sessions; the Baseline02 summary above takes precedence. Historical failures and seal-time NOT_RUN receipts remain evidence for their exact images.
Latest bounded physical result, 2026-09-27: Fix01 independently verified four small Wi-Fi hash roundtrips totaling 1 MiB in each direction. An isolated CONSYS retry after a mandatory WMT timeout restored firmware, RF calibration, WLAN and HCI. Full suspend restoration and peer Bluetooth audio were NOT_TESTED; the final health check still reported supplicant_unavailable despite association, route and working TCP. Earlier radio discovery notes below retain their original scope. Fix01 result.
2026-09-15: fresh M5 entry audit supersedes the deferred research wording below. Own retained BTIF/DMA/IRQ and CONSYS_MT6582 metadata, exact stock firmware hashes/header order and private read-only NVRAM/PROTECT copies are now recorded. Actual controller stepping, calibration record mapping and radio addresses remain unresolved. Current POWER-02 charging fails; M4 acceptance and the integrated M5 deployment remain blocked. No native Wi-Fi/HCI or connectivity qualification is claimed.
2026-09-10: additional research supplies native cold-start/A2DP records with complete modem calibration followed by MD1 off. It also records failed EDR remedies and a Basic Rate workaround. Validate our own revision and calibration; never adopt another player's fixture. M5 remains deferred, and FM reception is excluded for the owner's older board.
Date: 2026-09-08. Related tasks: Y2E-101/105/115. No radio scan, enable, pair, connect or firmware-loading action was performed.
CONFIRMED current metadata: /system/etc/firmware lists several WIFI_RAM_CODE variants, WMT_SOC.cfg, MT6572/82 patch blobs and an mt6627 directory. /system/lib/modules/wlan.ko is a symlink naming wlan_consys_mt6582.ko; target availability/loading is not established by that listing. /proc/modules is empty. Firmware presence does not prove which blob is selected, which transport is used, or whether a radio is populated and operational.
CONFIRMED static ramdisk evidence: WMT/connection loader declarations and STP-related node setup; a ttyMT2 connectivity comment. These identify a vendor ecosystem, not a standard Linux HCI transport contract or safe diagnostic UART. Current ro.hardware=mt6582 is observed; reconcile it with older Bluetooth HAL-selection inferences instead of copying them.
The early BLUETOOTH_STACK_INVESTIGATION.md describes BlueAngel artifacts and originally lacked runtime evidence. Later archived route observations have different scope. Neither justifies promising BlueZ/PipeWire compatibility or specific negotiated codecs on native Linux.
UNKNOWN: exact controller revision, bus/transport, power/reset/clock sequence, required firmware/module hashes and order, calibration storage/ownership, MAC persistence and standard-kernel driver applicability. Keep addresses, bonds and network credentials private. Next proof is a bounded driver/firmware identity inventory with no radio state changes.