Skip to content
2 changes: 1 addition & 1 deletion AGENTS.md
Original file line number Diff line number Diff line change
Expand Up @@ -172,7 +172,7 @@ Each entry gives the gotcha, then the reason or the fix. Details are in the link

### Suspend and modem

- **On the 9a, NFC must not poll through s2idle.** A suspend with neard polling reset the palmtop (2026-10-27); with polling stopped, 35 s2idle cycles with RTC wakes that day all woke. The cause isn't known: nxp-nci has no PM callbacks, but neither does the 7 Pro's st-nci, so the controller's interrupt or the I2C bus during the freeze are the next suspects. A sleep hook on the palmtop, `omamobile-tegu-nfc` (not in the repository yet; it comes with tegu's device package), stops neard's poll loop before a systemd sleep and restarts it after; a sleep written straight to `/sys/power/state` (`rtcwake -m mem`) skips the hook, so stop polling by hand first. [tegu MODEM-PLAN](device/tegu/MODEM-PLAN.md) "Suspend with the modem on"
- **On the 9a, NFC must not poll through s2idle.** A suspend with neard polling reset the palmtop (2026-10-27); with polling stopped, 35 s2idle cycles with RTC wakes that day all woke (`7.3.0-rc6-omamobile-tegu-s2mpg15-m1`). The cause isn't known: nxp-nci has no PM callbacks, but neither does the 7 Pro's st-nci, so the controller's interrupt or the I2C bus during the freeze are the next suspects. A sleep hook on the palmtop, `omamobile-tegu-nfc` (not in the repository yet; it comes with tegu's device package), stops neard's poll loop before a systemd sleep and restarts it after; a sleep written straight to `/sys/power/state` (`rtcwake -m mem`) skips the hook, so stop polling by hand first. [tegu MODEM-PLAN](device/tegu/MODEM-PLAN.md) "Suspend with the modem on"
- **UFS must be at `spm_lvl` 5 in system suspend.** At the core default 3, UFS failed its first hibern8 exit on resume and needed a host reset (it recovered, no I/O errors); at 5 the first real s2idle freeze resumed clean. gs201 now defaults to 5 in the kernel (S3a, `a7e5e1d95`), so a suspend test only checks that `spm_lvl` reads 5; on a kernel without that commit, `echo 5` first. [SUSPEND-PLAN](device/cheetah/SUSPEND-PLAN.md) S1-S3
- **Only s2idle, never `deep`.** The firmware advertises PSCI SYSTEM_SUSPEND, so mainline offers `deep`, but Linux has no gs201 SoC sleep sequence (SYS_SLEEP), and the phone wouldn't wake. `mem_sleep_default=s2idle` in the bootargs and device-cheetah's `sleep.conf` drop-in keep it on s2idle. [SUSPEND-PLAN](device/cheetah/SUSPEND-PLAN.md) §1, §5
- **Some unattended RTC cycles wake early on the PMIC interrupt** (2 of 20 in S4, 4 of 31 in S5c, at 13-16 s instead of 30 s). Which PMIC event it is hasn't been read yet; count it, don't treat it as a failed cycle. [SUSPEND-PLAN](device/cheetah/SUSPEND-PLAN.md) S4, S5c
Expand Down
7 changes: 6 additions & 1 deletion device/tegu/MODEM-PLAN.md
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,12 @@ The 9a's GNSS is the modem's own: stock's `cp_shmem` carries a
over SPI (`gnss@0`, `samsung,gnss-spi`, interrupt on gpa5-4). So GNSS needs
the modem up, but not a SIM. Following Derek's advice on issue #51, the
modem comes up for GNSS first, and the SIM (a prepaid physical one on a
T-Mobile-network MVNO, PIN off, activated on slot A) comes later.
T-Mobile-network MVNO) comes later. It is activated in another phone or
through the carrier's web activation, with its PIN turned off there, then
put in the 9a with the 9a off. Not on slot A: stock there overwrites
`userdata` before it reaches recovery ([NOTES, 2026-10-26 and 2026-10-28](NOTES.md#2026-10-26-and-2026-10-28-slot-a-fallbacks-that-cost-userdata)), so booting it loses the Linux
root. The backup check and a
fresh read-only backup before the first Linux modem session stay as below.

## Safety

Expand Down
139 changes: 139 additions & 0 deletions device/tegu/NOTES.md
Original file line number Diff line number Diff line change
Expand Up @@ -117,6 +117,145 @@ The laptop keeps the command log. The 10-28 harness copy differed from the

Only byte 0x35 changed (bit 0x02), and slot A was untouched.

## 2026-10-28: what the device table's claims rest on

Each line is a run on this 9a, on `7.3.0-rc6-omamobile-tegu-s2mpg15-m1-g991042e2eff3-dirty`
unless it says otherwise; the kernel pull request has the details.

- **Display:** the published kernel drives the panel at 60 Hz. 120 Hz runs
on the palmtop with kernel #88's two files (DECON timers per panel), in
review.
- **Touch, sleep fix (kernel #82):** 20 s2idle cycles with RTC wakes in a
row, no `bad message header`, touch working after each; without it touch
broke at the 13th.
- **Buttons, Power wakes s2idle:** checked by hand on 2026-10-27, Power from
a freeze back to the desktop.
- **Bluetooth, BCM4383 (kernel #94, merged):** `hci_uart` reloaded with that
exact `hci_bcm.c`: the controller came up as `hci0` on its REG_ON latch.
Pairing (A2DP) ran with the older form of the code.
- **Charging, status and inhibit (kernel #92):** the driver bound at 0x69;
status, and inhibit through `charge_behaviour`, read and set on the
palmtop.
- **Haptics, CS40L26 with VPBR (kernel #95, merged):** binds, loads its
firmware with brownout prevention at 3.008 V, plays effects.
- **Torch, LM3644 (DT in kernel #96):** 25 brightness steps up to 75 mA on
the torch LED.
- **Suspend, Power and RTC wakes:** Power by hand as above; RTC wakes in the
#82 run and in the 2026-10-28 sleep tests, where one of four never woke
(recorded in the camera switching entry below).

## 2026-10-28: camera switching, pstore, idle hints, sleep memory rate

Kernel `7.3.0-rc6-omamobile-tegu-s2mpg15-m1-g991042e2eff3-dirty`.

- **Camera switching works** (kernel #103). The first capture after a
switch away from the ultrawide or front camera came back empty about
half the time. The LoCH's EN_STATUS (0x0a8) stays active with the old
link in its low bits after a stream that stopped cleanly, and a disabled
LoCH ignores the abort, so the next stream on another link got no
interrupts. Enabling it for the abort fixes it: nine captures, every
switch order, all write frames.
- **pstore** (kernel #106): ramoops on stock's region survives a watchdog
reset on zumapro (the bootloader writes its own log at the start of the
console zone).
- **Core and cluster power-down hints** (the CPU_INFORM register Google's
driver writes at every idle entry): powering down cluster 0 stops ACPM
answering its DVFS channel, and the palmtop resets (pstore shows it);
Google never powers down cluster 0. Clusters 1 and 2 are stable, but
neither hint saves anything measurable, screen on or off. Dropped.
- **Memory rate in s2idle:** with the cable in and charging paused,
10-minute RTC-woken freezes drew 371 and 373 mA at MIF 2028 MHz and
314 mA at 421 MHz. One later freeze never woke (in s2idle the watchdog
is stopped); a hand restart brings it back.
- **Link errors** on every camera are ERR_WRONG_CONFIG: packets on VC 0
with a data type no channel takes (an embedded-data line a frame on the
front and ultrawide, about 527 PDAF packets a frame on the main camera).
- **Fingerprint:** the trusted app's secure storage is up, with Trusty's
stores at the paths stock uses (`0`, `persist/0`) but on loop devices of
Linux-owned images under `/var/lib/omamobile-fp`, not on stock's
`trusty_userdata` and `trusty_persist` (omarchy-mobile #101;
FINGERPRINT-PLAN D1 and F4b). The factory
calibration lives with Google: the bootloader names the file (its
`get-cal-url` answer) and it comes from a manufacturing bucket. The panel's
id isn't the key (omarchy-mobile #101).
- **GNSS:** the S2MPG15's three GNSS rails are off at hand-off and stock
turns them on; they're now described (kernel #101) and on.

### Reserved memory beyond Google's map (kernel #80)

`zumapro-reserved-memory.dtsi` in #80 carries Google's fixed regions from
`zumapro-rmem.dtsi`, zuma's xhci pools, and three more:

- `abl@f8800000`: predeclared in stock's own factory base DTB as 32 MiB,
which ABL shrinks to 16 MiB at hand-off.
- `0x913a0000` (+0x60000) and `0xfd7ff000` (+0x1000): two ranges the early
bring-up kept reserved, next to GXP and the debug tail. They come from the
private DRAM map the first boots were built from, not from a public
source.

Keeping them reserved is the safe direction. What would free the two gaps
is stock's live `/proc/iomem` and reserved-memory nodes compared with ours;
this palmtop can't produce that (stock on slot A overwrites `userdata`
before it reaches recovery: [2026-10-26 and 2026-10-28](#2026-10-26-and-2026-10-28-slot-a-fallbacks-that-cost-userdata)),
so it needs a second 9a or a reinstall of stock.

## 2026-10-27: cameras, NFC, USB-C, storage power, fingerprint, drain

Kernel `7.3.0-rc6-omamobile-tegu-s2mpg15-m1-g991042e2eff3-dirty` (the
integration tree; the pieces are kernel PRs, linked below) for everything
here. The 35 s2idle cycles in AGENTS.md's NFC entry ran on this release too.

**Cameras (kernel #63, #103).** `cam -l`:

1: Internal back camera (/base/soc/usi@108c00c0/i2c@108c0000/camera@56)
2: Internal back camera (/base/soc/usi@108900c0/i2c@10890000/camera@1a)
3: Internal front camera (/base/soc/usi@108a00c0/i2c@108a0000/camera@1a)

`tools/phone-smoke`: `| cameras | ok | 3 listed by libcamera |`. With the
ultrawide's test pattern on, LoCH 0 on link 1 (CTL 0x01000200, RESOL
0x0c001070, 5312 bytes a line, format 4) wrote three 4208x3072 frames
(48955392 bytes) showing the sensor's eight colour bars, full width and
height; the solid-colour pattern (R 1023, Gr 0, B 341, Gb 682) reads back as
BGGR. The main camera (S5KGN8, C-PHY) delivers whole 4000x3000 frames once
gs201-csis raises the INT domain and the ISP front end's NOC clock while it
streams. The ultrawide and the front stream at about 29 fps through the
software ISP. The tested `phy-gs201-mipi-dcphy-rx.c` is #63's (ed3dacc7f958);
the tested `gs201-csis.c` (acd8d7d0987a) differs from #63's in one comment.
Open: the first capture after a camera switch is often empty.

**NFC (kernel #100).** The ST54L reads a contactless card through the
kernel and neard once nxp-nci sends ST's CORE_SET_CONFIG after init: an
ISO-DEP target on NFC-A, 15 activations in 45 s with the card held, neard
lists a tag, with no periodic discovery restart. NFC polling costs about
60 mA, so omamobile-shell 0.2.33 (PR #99 here) stops it with the screen off.

**USB-C.** tcpci_maxim (merged, kernel #97) loads at boot: a PD contract at
5 V 3 A, replug back within 1 s. Host mode: VBUS comes from the RT9470 on
USI10 (Google's ext-otg-only use case; USI10 clocks in kernel #102), the
TCPC sees VBUS, dwc3 goes host, but a flash drive never connects: the
eUSB2 repeater (hsi2c9 0x3e) doesn't take the host role yet. Plug-in wakes
s2idle once the TCPC's interrupt is armed.

**Storage power (kernel #99).** UFS idle clock gating: the cores spend
60-68 % of an idle minute in C2 with the core's latency request left on
(0 % with the clocks held on), no UFS errors under load.

**Fingerprint.** The sensor's 3.3 V is S2MPG15 LDO15S, off at hand-off
(kernel #101); with it on the trusted app finds the sensor and the HAL
reports it at (540, 1769), radius 113. Init still ends without factory
calibration: `mfg_data` is all zeros on this palmtop and the TA's
calibration objects don't open (FINGERPRINT-PLAN.md).

**Drain, cable out.** Awake with the screen off: 556 mA before UFS gating,
409 mA after it at MIF 2028 MHz and 366 mA at 421 MHz (ACPM lowers MIF on
request); NFC polling about 60 mA of that; Wi-Fi off and the CPUs at their
lowest changed little. s2idle: about 140 mA over 59 min. MIF stays at the
bootloader's 2028 MHz until something asks for less.

**Recovery.** A reboot stuck in `device_shutdown` needs Power held about
30 s, then Power; with a task stuck in D, reboot with `echo b >
/proc/sysrq-trigger` instead of `systemctl reboot`.

## 2026-10-26: audio (speakers, microphones, A2DP) with omamobile-audio 0.5.0

Kernel `7.3.0-rc6-omamobile-tegu-s2mpg15-m1-g991042e2eff3-dirty`: the private
Expand Down
2 changes: 1 addition & 1 deletion docs/DEVICES.md
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ The status of each device, as verified on real hardware. Rows in the matrix have
| Pixel 6 | oriole | gs101 (Tensor G1) | WIP ([#2](https://github.com/derekross/omarchy-mobile/issues/2): descriptor and boot images from the factory image and a retail phone; not booted yet) | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no |
| Pixel 4 | flame | SM8150 (Snapdragon 855) | testing (temporary RAM boot; [notes](../device/flame/BRINGUP-M1.md)) | testing (RAM probe; all UFS read-only) | testing (1080x2280 console; software desktop in separate Arch payload) | testing (separate Arch payload; FD640 contexts, llvmpipe desktop) | testing (separate Arch payload; events and PIN unlock verified) | no | testing (NCM + ACM + SSH; sustained transfers can disconnect) | testing (guarded RAM scan; association untested) | no | no | no | no | no | no | - | no | no | no | no | no | no | - |
| Pixel 4 XL ([notes](../device/coral/NOTES.md)) | coral | SM8150 (Snapdragon 855) | WIP (direct ABL v2 boot on the test handset's slot A; slot-B install and §3.1 entry pending; [notes](../device/coral/BRINGUP-M1.md)) | testing (read-only `blockdev` guard + RAM `overlayfs`) | testing (1440x3040@60Hz S6E3HC2 DSC 1.1 KMS, DPMS, brightness) | testing (Adreno 640 / FD640 Mesa Freedreno, Hyprland) | testing (FTS SPI multitouch + read-only flash guard) | testing (power via `qcom-pon` + volume via `gpio-keys`) | testing (CDC ECM + ACM + SSH; stays up with host USB switched off) | testing (`WCN3990` `ath10k_snoc` + `rmtfs -r`, WPA2/DHCP/HTTPS; radio sessions ran before the modem baseline) | WIP (`WCN3990` `hci_qca` `hci0` up + BLE scan; audio/pairing open) | WIP (PM8150B ADC `VBAT_SNS` voltage + FG temp; aftermarket pack on test unit: no reply from `1-0036`) | testing (PM8150B SMB5 charger with `4.20 V` / `500 mA` kernel clamp) | testing (`CS40L25A` I2C click waveform) | testing (PM8150L flash LEDs + quick-settings tile) | no | - | testing (QDSP6 + dual `CS35L36` amps capped at `-26 dB` + `IA8508` SPI mic DSP) | testing (`SLPI` + `fastrpc` + RAM-backed `hexagonrpcd` + QMI SSC: 7 sensors + Soli radar) | WIP (`MPSS` boots with `rmtfs -r` + patched `tqftpserv`, SIM detected; no IPA/voice yet; radio sessions ran before the modem baseline) | no | testing (`CAMSS` + `IMX363`, `IMX481`, `IMX520`, dual `OV7251` NIR; `libcamera` softISP) | no | - |
| Pixel 9a ([notes](../device/tegu/NOTES.md)) | tegu | zumapro (Tensor G4) | WIP ([#28](https://github.com/derekross/omarchy-mobile/issues/28): the SoC's clocks and bindings are merged into linux-omarchy-mobile; the SoC and board device tree is in review, [kernel #80](https://github.com/derekross/linux-omarchy-mobile/pull/80)) | WIP | WIP (DSI panel, 1080x2424; unpublished kernel) | WIP (Panthor, the desktop renders; unpublished kernel) | WIP | WIP (gpio-keys; unpublished kernel) | WIP | WIP (brcmfmac joins WPA2; driver in review, [kernel #56](https://github.com/derekross/linux-omarchy-mobile/pull/56)) | WIP (powered; unpublished kernel) | WIP (the gauge reads; unpublished kernel) | no | no | no | no | WIP (audio and sensors; unpublished kernel) | WIP (speakers capped, mics, A2DP; unpublished kernel) | WIP (accelerometer, light, proximity; unpublished kernel) | no | no | no | no | WIP (Trusty's TA runs and the HAL sees the sensor, unpublished kernel; enrollment waits for the factory calibration; [plan](../device/tegu/FINGERPRINT-PLAN.md)) |
| Pixel 9a ([notes](../device/tegu/NOTES.md)) | tegu | zumapro (Tensor G4) | WIP ([#28](https://github.com/derekross/omarchy-mobile/issues/28): clocks, ACPM, panel and TCPC merged; the SoC and board device tree is in review, [kernel #80](https://github.com/derekross/linux-omarchy-mobile/pull/80)) | WIP (UFS root; reset across suspend merged, idle clock gating [kernel #99](https://github.com/derekross/linux-omarchy-mobile/pull/99)) | WIP (DSI panel 1080x2424 at 60 Hz, 120 Hz in review; panel driver merged, display DT [kernel #90](https://github.com/derekross/linux-omarchy-mobile/pull/90), 120 Hz [kernel #88](https://github.com/derekross/linux-omarchy-mobile/pull/88)) | WIP (Panthor to 890 MHz; DT [kernel #85](https://github.com/derekross/linux-omarchy-mobile/pull/85), [kernel #87](https://github.com/derekross/linux-omarchy-mobile/pull/87)) | WIP (DT [kernel #81](https://github.com/derekross/linux-omarchy-mobile/pull/81), sleep fix [kernel #82](https://github.com/derekross/linux-omarchy-mobile/pull/82)) | WIP (gpio-keys; Power wakes s2idle) | WIP (gadget and replug; host mode: VBUS on, no enumeration yet; repeater driver [kernel #107](https://github.com/derekross/linux-omarchy-mobile/pull/107)) | WIP (brcmfmac joins WPA2; driver in review, [kernel #56](https://github.com/derekross/linux-omarchy-mobile/pull/56)) | WIP (BCM4383 merged, [kernel #94](https://github.com/derekross/linux-omarchy-mobile/pull/94); BlueZ, A2DP) | WIP (the gauge reads; unpublished kernel) | WIP (status and inhibit [kernel #92](https://github.com/derekross/linux-omarchy-mobile/pull/92); PD contract, tcpci_maxim merged, [kernel #97](https://github.com/derekross/linux-omarchy-mobile/pull/97)) | WIP (CS40L26, VPBR merged, [kernel #95](https://github.com/derekross/linux-omarchy-mobile/pull/95); DT [kernel #96](https://github.com/derekross/linux-omarchy-mobile/pull/96)) | WIP (LM3644, 25 steps; DT [kernel #96](https://github.com/derekross/linux-omarchy-mobile/pull/96)) | no (firmware and CHPP run; the engine faulted with its three rails off at hand-off, now on, [kernel #101](https://github.com/derekross/linux-omarchy-mobile/pull/101)) | WIP (audio and sensors; unpublished kernel) | WIP (speakers capped, mics, A2DP; unpublished kernel) | WIP (accelerometer, light, proximity; unpublished kernel) | WIP (ONLINE with IPC, no SIM yet) | WIP (reads cards, [kernel #100](https://github.com/derekross/linux-omarchy-mobile/pull/100), clock [kernel #104](https://github.com/derekross/linux-omarchy-mobile/pull/104)) | WIP (all three stream through libcamera, switching between them works; [kernel #63](https://github.com/derekross/linux-omarchy-mobile/pull/63), [kernel #103](https://github.com/derekross/linux-omarchy-mobile/pull/103)) | WIP (s2idle; Power and RTC alarm wake; drain still high) | WIP (sensor, trusted app and secure storage up, its stores on loop devices of images in `userdata`, [#101](https://github.com/derekross/omarchy-mobile/pull/101); factory calibration fetched, not loaded yet) |
| Pixel 8 Pro ([notes](../device/husky/NOTES.md)) | husky | zuma (Tensor G3) | WIP ([#1](https://github.com/derekross/omarchy-mobile/issues/1): our kernel boots from slot B; the zuma device tree is merged ([kernel #13](https://github.com/derekross/linux-omarchy-mobile/pull/13))) | testing (UFS at HS-G4 behind the read-only guard; Arch on `userdata`; [UFS-PLAN](../device/husky/UFS-PLAN.md)) | WIP (simple-framebuffer console) | no | no | no | testing (USB and SSH) | no | no | no | no | no | no | no | no | no | no | no | no | no | no | no |
| Pixel 7 ([notes](../device/panther/NOTES.md)) | panther | gs201 (Tensor G2) | WIP (slot B, not marked successful; Arch and Hyprland boot; [BRINGUP-M1](../device/panther/BRINGUP-M1.md); board file merged, [kernel #108](https://github.com/derekross/linux-omarchy-mobile/pull/108)) | testing (UFS behind the read-only guard; Arch on `userdata`; merged, [kernel #108](https://github.com/derekross/linux-omarchy-mobile/pull/108)) | WIP (simple-framebuffer: console and the Hyprland lock screen) | testing (Panthor on the Mali-G710; Hyprland renders on it; in review, [kernel #129](https://github.com/derekross/linux-omarchy-mobile/pull/129)) | WIP (FocalTech FT3658 driver merged, [kernel #109](https://github.com/derekross/linux-omarchy-mobile/pull/109); the touch node is next) | testing (gpio-keys: power, volume; merged, [kernel #108](https://github.com/derekross/linux-omarchy-mobile/pull/108)) | testing (USB and SSH; a replug needs a reboot; merged, [kernel #108](https://github.com/derekross/linux-omarchy-mobile/pull/108)) | no | no | no | no | testing (CS40L26 with the phone's own firmware, `tools/extract-vendor-firmware -d panther`; in review, [kernel #130](https://github.com/derekross/linux-omarchy-mobile/pull/130)) | no | no | no | no | no | no | no | no | no | no |
| *Template* | *codename* | *SoC* | | | | | | | | | | | | | | | | | | | | | | |
Expand Down