58dcb3d121
- LBC supervisor v1..v4 (linux/patches/b10b3-lbc-supervisor-v3.patch + b10/b10b3/lbc-v3-to-v4-fault-clamp.diff): CHARGING 4.20 V/450 mA -> HOLD 4.05 V after 60 min CV -> battery -> new cycle on USB insert; FAULT = VDD_MAX 4.00 V clamp - RAM tests C/B/D/D2/D3/D4 (safety timer only ends fast charge), B10B-4 actuator audit (VDD_MAX works, USB_SUSP does not), B10B-5 FAULT clamp regression, production run, warm reboot via PON reboot reason - cache = B10B-5 c3732515 (HW EDL write, readback + full partition verify), class-A persistent validation - test-only hooks/images are not part of the production kernel; images with the AP PSK are not published
72 lines
5.8 KiB
Markdown
72 lines
5.8 KiB
Markdown
# B10B2-RESULT — upstream pm8916_lbc charger, RAM only
|
||
|
||
Date: 2026-10-03. **Verdict: PASS with a caveat** (see "Not verified").
|
||
aboot/cache untouched, eMMC writes = 0, no manual PMIC writes before or after probe. Boot class A (RTC 5 s) via RESET-held lk1st fastboot.
|
||
|
||
## Image / DT
|
||
- kernel `7.2.7-aurora-b10b2` = b10b1 + `CONFIG_CHARGER_PM8916_LBC` (`linux/aurora-b10b2.config`).
|
||
- DTB `linux/dts/msm8916-jz08-aurora-b10b2.dts` = B10B-1 +
|
||
`&pm8916_usbin { disabled }`, `&pm8916_charger { okay; monitored-battery; qcom,fast-charge-safe-voltage = 4200000;
|
||
qcom,fast-charge-safe-current = 500000 }`, `&pm8916_bms { power-supplies = <&pm8916_charger> }`,
|
||
`&usb`/`&usb_hs_phy { extcon = <&pm8916_charger> }` → **USB extcon has exactly one owner** (`charger@1000`, USB=1; verified in the DTB and live).
|
||
- simple-battery unchanged from B10B-1 (3000 mAh, 3.4 V, **4.20 V B10 cap; the cell allows 4.35 V — not used**).
|
||
- image `b10/b10b2/out/aurora-b10b2.img` 9f8a9096 (`build-b10b2.sh`), driver `b10b2-boot.sh K`, observer `b10b2-mon.sh N P` (reads only, ALERT on VBAT > 4.22 V / PMIC > 60 °C).
|
||
out/ contains the B8F initramfs (PSK) → never publish.
|
||
|
||
## Baseline before probe (`baseline-b10b1.txt`, `baseline-summary.txt`; running B10B-1, same lk1st path)
|
||
CHG_STATUS 05, CHGR RT 20, CHG_CTRL 90, VDD_MAX/SAFE 08/08 (4.20 V), IBAT_MAX/SAFE 00/0a (90/990 mA), VIN_MIN 04 (4308 mV),
|
||
TCHG_MAX_EN/TCHG_MAX 80/1d (timer on, 120 min), IBATTERM 0x105B 09, COMP_OVR1 (VBAT_DET override) 00, BOOT_DONE 80, 0x1643 0b,
|
||
USB 0x1308/0x1309 02/90, USB RT 03, BAT_IF 0x1208 83, BAT RT 03, USB IRQ cfg 0x1311/15/16 02.
|
||
|
||
## Charger driver writes (`diff-probe.txt`, full PMIC diff vs baseline)
|
||
| reg | before → after | meaning |
|
||
|---|---|---|
|
||
| 0x1044 IBAT_MAX | 00 → **04** | **450 mA** (500 mA requested; 90 mA hardware step: 500000/90000 − 1 = 4) |
|
||
| 0x1045 IBAT_SAFE | 0a → **04** | 450 mA |
|
||
| 0x1049 CHG_CTRL | 90 → **a0** | CHG_EN + PSTG_EN (bit4 cleared, bit5 set) |
|
||
| **0x1060 TCHG_MAX_EN** | **80 → 00** | **upstream disables the charger safety timer** (0x1061 stays 1d) |
|
||
| 0x1040 VDD_MAX / 0x1041 VDD_SAFE | 08 → 08 (written, same value) | 4.20 V |
|
||
Untouched: VIN_MIN, IBATTERM/termination (0x105B 09), COMP_OVR1/VBAT_DET (00), BOOT_DONE (80), WDOG, BAT_IF, USB IRQ config (same as with pm8916_usbin).
|
||
All other diffs are dynamic data (ADC 0x31/32/33xx, BMS 0x40xx data, RTC). Exactly as predicted from the driver source.
|
||
|
||
sysfs `pm8916-lbc-chgr` (type USB): online 1, constant_charge_current 450000, constant_charge_voltage_max 4200000. BMS status → Charging.
|
||
|
||
## Observation (`ram1-mon.txt`, `ram1-mon-table.txt`; 22 samples × 30 s, uptime 54–686 s, then stopped on purpose)
|
||
- State: CC/fast charge the whole time (CHG_STATUS 05, CHGR RT 20 = FAST_CHG_ON), USB path 02/90, USB RT 03, BAT RT 03, online 1.
|
||
- **VADC VBAT: 3.80 → 3.8235 V**; after the noisy LTE/Wi-Fi start-up a steady rise 3.8086 (295 s) → 3.8235 (686 s) ≈ +2.3 mV/min.
|
||
BMS FIFO 13316 → 13433 (diagnostic; BMS voltage_now still ×300 µV, ≈5.5 % high, not used for decisions).
|
||
- USB meter (operator): 0.357 A during kernel boot, then **0.355 A, peaks 0.38 A**.
|
||
- PMIC die temp 44 → 48–49 °C (plateau), CPU 45–47 °C. No ALERT. VBAT far below 4.22 V.
|
||
- CV and termination were not reached (estimated 2.5–4 h more from VBAT 3.82 V).
|
||
|
||
## BOOT_DONE / the earlier ~0.08 A
|
||
| boot | BOOT_DONE | IBAT_MAX | USB meter | VBAT trend |
|
||
|---|---|---|---|---|
|
||
| cache (lk2nd extlinux), class A | 00 | 00 (90 mA) | 0.15–0.41 A | charging, terminated after ~45 min (B10A) |
|
||
| RAM via lk1st fastboot, no LBC driver (B10B-0/1) | 80 | 00 (90 mA) | ≈0.08 A | flat / draining under load (ram1 drained to ≈3.7 V in ~7.7 h) |
|
||
| RAM via lk1st fastboot + pm8916_lbc (B10B-2) | 80 (before and after probe) | 04 (450 mA) | ≈0.36 A | rising ≈2.3 mV/min |
|
||
With BOOT_DONE = 1 the input/charge current follows IBAT_MAX: the ~0.08 A in B10B-0/1 was IBAT_MAX = 90 mA under BOOT_DONE, not another
|
||
cause. Without BOOT_DONE (cache boots) the LBC evidently ignores the 90 mA field (autonomous current). BOOT_DONE = 0 with IBAT_MAX = 450 mA
|
||
could not be tested in a RAM boot (lk1st always sets BOOT_DONE in fastboot). The USB meter shows the whole board, not the battery current;
|
||
the conclusion rests on meter + VADC VBAT trend together.
|
||
|
||
## Other subsystems
|
||
LTE START OK 74.1 s, ping via wwan0 10/10 (snapshot 3/4 during start-up); Wi-Fi AP enabled 78.8 s, Pronto running; NCM/telnet and
|
||
gadget (UDC configured) fine with the LBC extcon; display errors 0 (fbcon/panel up); eMMC w=0.
|
||
|
||
## Not verified (caveat)
|
||
Hardware termination and the (expected) absence of automatic recharge with upstream pm8916_lbc were **not observed in this run**
|
||
(low initial SOC, multi-hour duration, and the driver disables the safety timer, so an unattended multi-hour RAM charge was not
|
||
accepted). Absence of software recharge is established by the code audit (B10A-DRIVER-AUDIT: only the usb_vbus IRQ is handled).
|
||
The full cycle termination → VBAT drop → restart is to be tested in B10B-3 with a recharge supervisor.
|
||
|
||
## After the test
|
||
Monitor stopped (host and board side). The LBC registers written by the driver (IBAT_MAX 450 mA, timer off, CHG_CTRL a0) and
|
||
BOOT_DONE live in the PMIC retained domain and survive a warm/class-B reboot → return to SBL defaults needs a **full power-off
|
||
(battery out ≥ 90 s) = class A**.
|
||
|
||
## Bench reset after B10B-2 (class-A power cycle, no RESET; `post-classA-check.txt`)
|
||
Operator: USB out → battery out → ≈2.5 min → battery in → USB in. Boot 14:39 = normal cache B9C (`7.2.7-aurora-b9c`), RTC 3 s → **class A**.
|
||
Read-only snapshot: CHG_STATUS 05, VDD_MAX/SAFE 08/08, **IBAT_MAX 00**, IBAT_SAFE 0a, VIN_MIN 04, **CHG_CTRL 90**, IBATTERM 09,
|
||
**TCHG_MAX_EN/TCHG_MAX 80/1d**, COMP_OVR1 00, USB 0x1308 02, **BOOT_DONE 0x1642 00**, 0x1643 00, PON_REASON1 10, POFF2 00.
|
||
→ All LBC registers back at SBL1 defaults; the bench is clean.
|