From 8493c99e8b5c72e184222c21906eb1f9b8305874 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Tue, 27 Oct 2026 06:28:03 -0700 Subject: [PATCH 1/7] TOUCH-PLAN (panther): the FT3658 answers once its supplies are on The Pixel 7's touch controller read only zeros because ABL leaves its supplies off: S2MPG12 LDO25M (VDD) and LDO26M (AVDD), as on the Pixel Fold. With linux-omarchy-mobile#27's S2MPG12 LDO driver on the local kernel branch, the touch driver switches them on in Google's order and the controller answers: chip ID 0x56, firmware 0x16. LDO26M holds 3.3 V when ABL hands over, above stock's 2.5-3.0 V range; the plan stopped there (T2) and the DT now fixes it at 3.0 V, which the regulator core set while the rail was off (T2b). Only the two touch rails changed state, at the voltages expected; unbind asserts reset and drops both (T5). Touch events (T4) wait for someone at the phone; a read-only register dump shows the chip awake and scanning. NOTES gets the devinfo 3.19 evidence so far for L-36: every getvar and slot-byte reading agrees with cheetah's layout. Co-Authored-By: Claude Opus 5.5 --- device/panther/NOTES.md | 21 +++ device/panther/TOUCH-PLAN.md | 262 +++++++++++++++++++++++++++++++++++ 2 files changed, 283 insertions(+) create mode 100644 device/panther/TOUCH-PLAN.md diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index a4dd6ce46..7e687138b 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -52,3 +52,24 @@ Captured on stock `CP3A.261005.002.A1` (Android 17, security patch 2026-10-05) o | Charger, TCPC, gauge | MAX77759, as on cheetah | [V] live DT | | Reserved memory | Every stock `reserved-memory` region that Google's tree describes matches gs201.dtsi. `sec_dram`, `sec_pt` and `pkvm_guest_firmware` are injected by the bootloader (not in Google's DTS); `cp_aoc_rmem` (0x197fd000) is AoC SRAM, not DRAM | [V] live DT against the built cheetah DTB; Google `gs201-rmem.dtsi` | | Framebuffer | Not described in Google's tree: ABL injects it. Not copied from cheetah's `0xfac00000`; to be read from DPP0 on the first boot | [I] | + +## 2026-10-27: devinfo slot bytes against ABL (L-36 evidence) + +devinfo is version **3.19** here (cheetah's bootctl is verified on 3.15 +only). Only the version and the slot bytes (0x30-0x37) were read, from +Linux, read-only. Byte layout per slot: retry count, then flags (bit 0 +unbootable, bit 1 successful, bit 2 active, bit 3 fastboot_ok). + +| When | `fastboot getvar` | devinfo 0x30-0x37 | +| --- | --- | --- | +| stock capture, before any of our boots | A: retry 2, successful, bootable; B: retry 3, not successful, bootable | not read | +| one boot after `--set-active=b` (boot 7) | not read | `02 0a 00 00 02 04 00 00` | +| before the T2 flash (after boots 7 and 8) | B: retry 1, not successful, bootable | not read | +| after the T2 flash (`--set-active=b`) and two boots (T2, T2b) | not read | `02 0a 00 00 01 04 00 00` | + +Every reading agrees with the layout: slot A `02 0a` is retry 2, +successful, fastboot_ok, as getvar says; slot B counts down from 3 by one +per boot, ABL's own change, and `--set-active` resets it. Still missing +for L-36: a getvar and a byte read of the same state with no boot in +between (the next time the phone is in fastboot, before anything is +flashed). diff --git a/device/panther/TOUCH-PLAN.md b/device/panther/TOUCH-PLAN.md new file mode 100644 index 000000000..77037306c --- /dev/null +++ b/device/panther/TOUCH-PLAN.md @@ -0,0 +1,262 @@ + +# Pixel 7 (panther) touch: the supplies, then the chip ID + +Status (2026-10-27): T0-T3 and T5 passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on. T4 (touches) waits for the owner. + +The touch controller is a FocalTech FT3658 (chip ID 0x56) on USI0 SPI, +wired as the Pixel 7 Pro's S3908. A local driver +(`drivers/input/touchscreen/focaltech_ft3658.c`, kernel branch +`panther-first-light`) reads only zeros from it. BRINGUP-M1.md (boots 7 +and 8) ruled out the driver's timing, the SPI mode and speed, the +controller, the AP/AoC bus select (`gph1-0` reads AP) and the reset line. +What is left are the supplies. This plan switches them on through the +S2MPG12 regulator driver, then reads the chip ID. + +## What is known + +| Fact | Evidence | +| --- | --- | +| VDD is S2MPG12 LDO25M, AVDD is LDO26M, and the touch controller is their only consumer | stock live DT: `vdd-supply` and `avdd-supply` of `focaltech,ts` are the only references to `S2MPG12_LDO25` and `S2MPG12_LDO26` [V] | +| Stock ranges: LDO25M 1.6-1.95 V, LDO26M 2.5-3.0 V | stock live DT `regulator-min/max-microvolt` [V] | +| The Pixel Fold uses the same two LDOs for its cover touch at 1.8 V and 3.3 V. Panther's stock range for LDO26M stops at 3.0 V, so **felix's voltages are not copied** | `gs201-felix.dts` on `omamobile` [V] | +| On the Fold, ABL leaves both off, and the cover controller answers once its driver enables them | device/felix/NOTES.md, 2026-10-23 [V] | +| On panther, ABL hands the touch over held in reset (`gpp23-2` output low) | BRINGUP-M1.md, boot 8 [V] | +| On panther, ABL leaves LDO25M/LDO26M off | [I], read in T2 | +| Google's power-on: reset low, AVDD on, VDD on, 1 ms, reset high; power-off: reset low, 1 ms, AVDD off, VDD off. No voltage is set | Google's GPL `focaltech_core.c`, `fts_power_source_ctrl()` [V] | +| The S2MPG12 LDO driver exists: linux-omarchy-mobile#27, merged into `omamobile` but not in the `omamobile-int` snapshot. It registered all 28 LDOs read-only on the Pixel 7 Pro, and drives the Fold's cover touch rails | device/cheetah/NOTES.md, 2026-10-23; device/felix/NOTES.md [V] | +| #27's PM regmap can write only the interrupt masks and L1M_CTRL-LDO_CTRL2 (0x2b-0x49); buck control and SW_RESET stay outside | #27, `mfd: sec: s2mpg12: open the LDO control registers` [V] | +| Without `regulator-min/max-microvolt`, the regulator core never sets a voltage; without `regulator-always-on`, a consumer may switch the rail | `drivers/regulator/of_regulator.c` (`apply_uV`, `REGULATOR_CHANGE_STATUS`) [V] | +| The core's unused-rail cleanup runs once, about 30 s after boot, and only switches off rails that are on, unused and switchable | `drivers/regulator/core.c` [V] | + +## The changes (local kernel branch, laptop only) + +- **K1.** Cherry-pick #27's three driver commits from `omamobile` onto + `panther-first-light`: `dt-bindings: mfd: samsung,s2mpg12: describe the + LDO regulators`, `mfd: sec: s2mpg12: open the LDO control registers`, + `regulator: s2mps11: add S2MPG12 LDOs`. They apply cleanly (checked in + a temporary worktree). +- **K2. DTS.** `&acpm_ipc { pmic-1 }` as felix's (`samsung,s2mpg12-pmic`, + GIC SPI 72, `google,vgpio2ap-sysreg`), with only `ldo25m` + (`vdd-ts`) and `ldo26m` (`avdd-ts`) under `regulators`, **with no + voltage properties and no `regulator-always-on`**: the PMIC keeps + whatever voltage it holds. The touch node gets `vdd-supply` and + `avdd-supply`. The bootargs blacklist `sec_acpm` too, so the regulators + register only when loaded by hand, after the 30 s cleanup. +- **K3. Driver.** Google's sequence instead of `devm_regulator_bulk_get_enable()` + (which switches both at once, in no set order): reset held asserted, + AVDD on, VDD on, 1 ms, release reset, then the 200 ms boot wait. A devm + action powers off in the reverse order, reset asserted first, so unbind + never drops the rails under a running chip. +- **K4. Checks.** checkpatch, a W=1 build, and the DTB decompiled: the two + LDO nodes with no voltage or `always-on` property, the touch supplies, + `modprobe.blacklist=focaltech_ft3658,sec_acpm`. + +## Gates + +Slot B's retry budget (BRINGUP-M1.md): three boots per `--set-active`, +and nothing marks the slot successful. T2 is one boot. + +### T0. Read-only, before any new image + +Hash `efs`, `efs_backup`, `modem_userdata`, `persist`, `modem_a` and +`modem_b` on the phone and compare them with +`~/android/reference/panther-backup-20261026-linux/SHA256SUMS`. The sums +stay out of the repository. + +Then read slot B's five boot partitions (boot 8, the last good boot) into +`~/android/mainline/known-good-boot8/`, as the rollback set. + +**Pass:** every sum matches. **Stop:** any mismatch. **Risk:** none. + +### T1. Build and check the images (laptop) + +K1-K4, then `DEVICE=panther bin/omamobile kernel`, `modules`, `images`. + +**Pass:** K4's checks. **Risk:** none. + +### T2. Boot, then register the regulators read-only + +`DEVICE=panther bin/omamobile flash` from fastboot by the buttons (it +restores the retry budget), or `push` if the budget allows. On the +laptop, start a log first: `tools/phone-ssh 'dmesg -w' > /t2.log &`. +On the phone, at least 40 s after boot: + +```sh +mount -t debugfs none /sys/kernel/debug 2>/dev/null +cat /sys/kernel/debug/regulator/regulator_summary > /tmp/reg-0.txt +modprobe sec_acpm +sleep 2 +cat /sys/kernel/debug/regulator/regulator_summary > /tmp/reg-1.txt +diff /tmp/reg-0.txt /tmp/reg-1.txt +grep -A1 -E 'vdd-ts|avdd-ts' /tmp/reg-1.txt +``` + +**Expect / pass:** `s2mpg12-regulator` binds; `vdd-ts` and `avdd-ts` +appear with no users, their voltages inside the stock ranges (1.6-1.95 V +and 2.5-3.0 V); nothing else in the summary changes; no +`ufshcd_err_handler`; USB and SSH stay up. Record whether the two rails +read **disabled** (the hypothesis) or enabled. + +**Stop:** a voltage outside the stock range; any rail other than the two +appearing as switched; a UFS error; USB dropping. Do not load the touch +driver. + +**Rollback:** `rmmod` is not used on the PMIC. Flash +`~/android/mainline/known-good-boot8/` to slot B from fastboot. + +**Risk:** low. The driver registers the rails and switches none. + +### T3. Load the touch driver + +```sh +cp /sys/kernel/debug/regulator/regulator_summary /tmp/reg-2.txt +modprobe focaltech_ft3658 +cp /sys/kernel/debug/regulator/regulator_summary /tmp/reg-3.txt +diff /tmp/reg-2.txt /tmp/reg-3.txt +dmesg | grep -E 'ft3658|s2mpg12' +``` + +**Expect / pass:** `FT3658, chip id 0x56, firmware 0x..`; the diff shows +only `vdd-ts` and `avdd-ts` enabled, one user each, voltages unchanged; +a `FocalTech FT3658` input device. + +**Stop:** no chip ID (unbind at once, T5); a rail other than those two +changing; a voltage changing; a UFS error; USB dropping; an interrupt +storm (`/proc/interrupts` climbing with no touch). + +**Risk:** medium. This is the first write to the PMIC on this phone: the +enable bits of LDO25M and LDO26M, at the voltages the PMIC already holds. + +### T4. Touches + +Capture `event` of the new input device with `timeout 30 cat +/dev/input/eventN > /tmp/ts.bin` while the owner touches the four +corners, then decode it; then unlock the lock screen with the PIN on the +on-screen keyboard. + +**Pass:** the corners read near (0, 0), (1079, 0), (0, 2399), +(1079, 2399); slots go up on release; the lock screen accepts the PIN. + +**Stop:** axes swapped or inverted (record and fix in the DT with +`touchscreen-swapped-x-y` / `-inverted-x/y`, a new boot), an interrupt +storm. + +### T5. Unbind + +```sh +echo spi0.0 > /sys/bus/spi/drivers/ft3658/unbind +grep -A1 -E 'vdd-ts|avdd-ts' /sys/kernel/debug/regulator/regulator_summary +``` + +**Pass:** both rails disabled, no users; `gpp23-2` low. +**Stop:** either rail still enabled, or another rail changed. + +## After the gates + +If T3-T5 pass: autoload (drop the two blacklist entries, a boot of its +own), the binding for `focaltech,ft3658`, and the kernel pull request +against `omamobile` (board DTS split by subsystem, new nodes disabled, +the PMIC node as felix's). If T2 shows the rails already **on**, the +supplies were not the cause: stop and go back to the SPI side. + +## Results + +### 2026-10-27: T0 + +- `modem_a` and `modem_b` match the 2026-10-26 backup. `efs`, + `efs_backup`, `modem_userdata` and `persist` do not: boot 6's fallback + ran stock recovery on slot A, which mounted them. A byte compare against + the backup shows only F2FS checkpoint areas (`efs`, `efs_backup`, + `modem_userdata`) and the ext4 superblock and journal (`persist`, which + now reads "needs journal recovery"); `devinfo` differs in slot B's retry + byte (0x34) and 0x6e-0x6f. A fresh backup, + `~/android/reference/panther-backup-20261027-linux`, is the baseline + from here. +- Slot B's boot 8 images are saved in `~/android/mainline/known-good-boot8/`. + +### 2026-10-27: T1 + +- K1: #27's three commits cherry-picked (`-x`) onto `panther-first-light`. +- K2, K3: two local commits (driver power sequence; the touch supplies). +- K4: checkpatch `--strict` clean; `W=1` builds of `focaltech_ft3658.o`, + `s2mps11.o` and `sec-acpm.o` with no warning; the DTB has `ldo25m` + (`vdd-ts`) and `ldo26m` (`avdd-ts`) with no voltage or `always-on` + property, the touch node's two supplies, and + `modprobe.blacklist=focaltech_ft3658,sec_acpm`. +- Kernel `7.3.0-rc6-omamobile-push1-g13a1ccd9ae5d` (a new release: its + modules must be installed before the boot). + +### 2026-10-27: T2 (stopped on the voltage rule) + +Kernel `7.3.0-rc6-omamobile-push1-g13a1ccd9ae5d`, flashed from fastboot +(before: `slot-retry-count:b` 1, `slot-successful:b` no, +`slot-unbootable:b` no, matching the devinfo bytes). SSH 6.3 s into the +boot; `sec_acpm` and `focaltech_ft3658` not loaded (the blacklist works). + +- `modprobe sec_acpm` at 45 s: `s2mpg12-regulator` registered all 28 + LDOs with no users; the regulator summary only gained entries, nothing + changed or disappeared. No UFS error; USB up. The PMIC's RTC registered + as `rtc0` and set the system clock, so the clock is right from here. +- **The hypothesis holds:** `vdd-ts` (LDO25M) and `avdd-ts` (LDO26M) are + **disabled**. ABL leaves the touch supplies off on this phone too; that + is why the FT3658 read zeros. +- **Stop:** `vdd-ts` holds 1.8 V, inside stock's 1.6-1.95 V, but + `avdd-ts` holds **3.3 V**, above stock's 2.5-3.0 V. Stock's regulator + core, with those constraints, lowers it to the range's maximum before + the touch driver enables it, so on Android the FT3658 runs at 3.0 V + [I: from the core's `machine_constraints_voltage()`, not measured]. The + touch driver was not loaded. +- The other LDOs, for the record (read-only): enabled 1M-12M, 14M-18M, + 21M (3.3 V), 27M (3.025 V); disabled 13M, 19M, 20M, 22M-24M, 28M. + +### T2b (proposed): AVDD at 3.0 V before it is switched on + +One DT change: `regulator-min-microvolt` and `regulator-max-microvolt` +3000000 on `ldo26m` only. When `sec_acpm` registers the rail, the core +moves the selector of `L26M_CTRL` (0x45, bits 5:0, group 5: 1.8 V + +25 mV steps) from 60 (3.3 V) to 48 (3.0 V), with the enable bits (7:6) +untouched and the rail still off. LDO25M stays without constraints. + +**Pass:** after `modprobe sec_acpm`, `avdd-ts` reads 3000000 µV and +disabled; nothing else changed. **Stop:** any other value, any other rail +changing. **Risk:** low to medium: the first PMIC write on this phone, one +voltage selector of an LDO that is off and has no other consumer. + +### 2026-10-27: T2b + +Kernel `7.3.0-rc6-omamobile-push1-g0e2cc7da7352` through `push`; SSH +5.4 s into the boot. `modprobe sec_acpm` at 46 s: +`avdd-ts: Bringing 3300000uV into 3000000-3000000uV`. `avdd-ts` reads +3000000 µV and **disabled**, `vdd-ts` 1800000 µV and disabled, both with +no users. The other 26 LDOs read exactly as in T2. No UFS error, oops or +BUG. **Passed.** + +### 2026-10-27: T3 + +Same boot as T2b. `modprobe focaltech_ft3658`: +`FT3658, chip id 0x56, firmware 0x16` 0.2 s after the load, and the input +device `FocalTech FT3658` (`event1`). The regulator summary changed in +`vdd-ts` and `avdd-ts` only: enabled, one user each (`spi0.0-vdd`, +`spi0.0-avdd`), 1800000 and 3000000 µV as before. One interrupt at the +load, none in the next 5 s with no touch. No UFS error; USB up. +**Passed.** The supplies were the cause. + +### 2026-10-27: T4, first try (no result) + +A 60 s capture of `event1` came back empty and the interrupt count stayed +at 1: the owner was away and nobody touched the screen. To repeat. + +- A read-only register dump from a debug build of the module (not + committed; loaded with `insmod`, then replaced by the normal module): + work mode (0x00) 0x00, normal; power mode (0xA5) 0x00, active; monitor + (0x86) 0x01; heatmap (0x9E/0xED) 1/0, uncompressed, the firmware + default; continuous report (0xE7) 1; vendor ID (0xA8) 0x8d; the scan + counter (0x91) climbed 0x2d, 0x40, 0x54 at 100 ms steps. The chip is + awake and scanning; with no touch it raises no interrupt. + +### 2026-10-27: T5 + +`echo spi0.0 > /sys/bus/spi/drivers/ft3658/unbind`: `vdd-ts` and +`avdd-ts` disabled with no users, `gpp23-2` driven low (reset asserted), +the input device gone, no other rail changed. **Passed.** The driver was +loaded again afterwards, so the phone is left with touch powered. From f53af4452136e0b1bbd9e2b5003f387cc6a52f81 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Tue, 27 Oct 2026 20:22:26 -0700 Subject: [PATCH 2/7] TOUCH-PLAN (panther): T4 passed, touches and the lock screen PIN The corners land where they should on the 1080x2400 panel, so the board file needs no swap or inversion flags; the owner unlocked the phone by touch. All gates passed; autoload is next. Co-Authored-By: Claude Opus 5.5 --- device/panther/TOUCH-PLAN.md | 21 ++++++++++++++++++++- 1 file changed, 20 insertions(+), 1 deletion(-) diff --git a/device/panther/TOUCH-PLAN.md b/device/panther/TOUCH-PLAN.md index 77037306c..3787d14ea 100644 --- a/device/panther/TOUCH-PLAN.md +++ b/device/panther/TOUCH-PLAN.md @@ -1,7 +1,7 @@ # Pixel 7 (panther) touch: the supplies, then the chip ID -Status (2026-10-27): T0-T3 and T5 passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on. T4 (touches) waits for the owner. +Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Next: autoload (see After the gates). The touch controller is a FocalTech FT3658 (chip ID 0x56) on USI0 SPI, wired as the Pixel 7 Pro's S3908. A local driver @@ -254,6 +254,25 @@ at 1: the owner was away and nobody touched the screen. To repeat. counter (0x91) climbed 0x2d, 0x40, 0x54 at 100 ms steps. The chip is awake and scanning; with no touch it raises no interrupt. +### 2026-10-27: T4 + +A 60 s capture of `event1` (`timeout 60 cat`, decoded on the phone) while +the owner touched the four corners and swiped down the middle, three +times: 1632 interrupts, 17 strokes, no `report read failed`. + +| Touch | First point (x, y) | +| --- | --- | +| Top left | (44, 19) | +| Top right | (1071, 45) | +| Bottom left | (34, 2389) | +| Bottom right | (1033, 2333) | +| Swipe down, centre | (576, 22) to (598, 1092) | + +x spans 29-1072 and y 5-2389 on the 1080x2400 panel: no swap or +inversion, the DT needs no `touchscreen-*` flags. The owner then locked +the screen and entered the PIN by touch in the Omarchy lock screen. +**Passed.** + ### 2026-10-27: T5 `echo spi0.0 > /sys/bus/spi/drivers/ft3658/unbind`: `vdd-ts` and From a6fa81f216bc65e1015a18d1bd99d867fe893da0 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Tue, 27 Oct 2026 20:40:25 -0700 Subject: [PATCH 3/7] TOUCH-PLAN (panther): the touch loads on its own at boot The bootargs no longer keep the FT3658 and the S2MPG12 regulators out, and the touch comes up 3.6 s into the boot; the lock screen took the PIN by touch. NOTES.md records the next slot-B getvar for L-36, and two things found on the way: Omarchy's shut down reboots the phone with the cable in, and linux-modules-cleanup drops a modules-push tree when another kernel boots first. Co-Authored-By: Claude Opus 5.5 --- device/panther/NOTES.md | 36 ++++++++++++++++++++++++++++++++---- device/panther/TOUCH-PLAN.md | 15 +++++++++++++-- 2 files changed, 45 insertions(+), 6 deletions(-) diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index 7e687138b..d4e204cb6 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -66,10 +66,38 @@ unbootable, bit 1 successful, bit 2 active, bit 3 fastboot_ok). | one boot after `--set-active=b` (boot 7) | not read | `02 0a 00 00 02 04 00 00` | | before the T2 flash (after boots 7 and 8) | B: retry 1, not successful, bootable | not read | | after the T2 flash (`--set-active=b`) and two boots (T2, T2b) | not read | `02 0a 00 00 01 04 00 00` | +| one more boot of slot B (an unplanned reboot), then fastboot | current slot **A**; A: retry 2, successful, bootable; B: retry 0, not successful, bootable | not read | Every reading agrees with the layout: slot A `02 0a` is retry 2, successful, fastboot_ok, as getvar says; slot B counts down from 3 by one -per boot, ABL's own change, and `--set-active` resets it. Still missing -for L-36: a getvar and a byte read of the same state with no boot in -between (the next time the phone is in fastboot, before anything is -flashed). +per boot, ABL's own change, and `--set-active` resets it. The last row is +the next step of the count (1 to 0), and shows ABL's choice once B has no +retries left: A becomes current, B is not marked unbootable. Still +missing for L-36: a getvar and a byte read of the same state with no boot +in between; a byte read needs Linux, so it has to be taken right before a +planned reboot to fastboot. + +## 2026-10-27: Omarchy's power off reboots the phone + +"Shut down" in the Omarchy menu, with the USB cable to the laptop +plugged in, rebooted into slot B instead of powering off (and spent a +slot B retry). Holding the power key forced it off. Not looked into yet: +the bootloader may be powering the phone back on because the cable is in. +Until then, reach fastboot from a forced power off. + +## 2026-10-27: modules-push and linux-modules-cleanup + +`kernel-modules-hook` (in the root image) enables +`linux-modules-cleanup.service`, which on every boot moves each +`/usr/lib/modules/` that is not the running kernel and that no +package owns into `/usr/lib/modules/.old`. A tree from +`bin/omamobile modules-push` is owned by no package, so a boot of any +other kernel before the new one (here, the unplanned reboot above) loses +it, and the new kernel then boots with no modules. `.old` kept only the +old release, so the pushed tree was not recoverable from it. Fix used: +`modules-push` again from the new kernel. Loading the missing modules +afterwards with `udevadm trigger` raced (ACPM probed before its mailbox, +`15 channels, 0 usable`, so `sec-pmic-acpm` failed with -ENODEV before +any PMIC write); a reboot loaded everything in order. To raise with the +tooling: push right before the reboot, or have `modules-push` warn. + diff --git a/device/panther/TOUCH-PLAN.md b/device/panther/TOUCH-PLAN.md index 3787d14ea..e7e796e69 100644 --- a/device/panther/TOUCH-PLAN.md +++ b/device/panther/TOUCH-PLAN.md @@ -1,7 +1,7 @@ # Pixel 7 (panther) touch: the supplies, then the chip ID -Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Next: autoload (see After the gates). +Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Autoload passed too: the touch comes up on its own at boot. The touch controller is a FocalTech FT3658 (chip ID 0x56) on USI0 SPI, wired as the Pixel 7 Pro's S3908. A local driver @@ -154,7 +154,7 @@ grep -A1 -E 'vdd-ts|avdd-ts' /sys/kernel/debug/regulator/regulator_summary ## After the gates If T3-T5 pass: autoload (drop the two blacklist entries, a boot of its -own), the binding for `focaltech,ft3658`, and the kernel pull request +own; done, see the results), the binding for `focaltech,ft3658`, and the kernel pull request against `omamobile` (board DTS split by subsystem, new nodes disabled, the PMIC node as felix's). If T2 shows the rails already **on**, the supplies were not the cause: stop and go back to the SPI side. @@ -279,3 +279,14 @@ the screen and entered the PIN by touch in the Omarchy lock screen. `avdd-ts` disabled with no users, `gpp23-2` driven low (reset asserted), the input device gone, no other rail changed. **Passed.** The driver was loaded again afterwards, so the phone is left with touch powered. + +### 2026-10-27: autoload + +Kernel `7.3.0-rc6-omamobile-push1-gdca2c17c8c19`: the bootargs without +`modprobe.blacklist`, nothing else changed. Flashed from fastboot. The +first boot came up with no modules (`linux-modules-cleanup`, see +NOTES.md); after `modules-push` again and a reboot over SSH: ACPM +`15 channels, 14 usable` at 3.3 s, `FT3658, chip id 0x56, firmware 0x16` +and the input device at 3.6 s, `vdd-ts` 1.8 V and `avdd-ts` 3.0 V +enabled with one user each, unchanged at 63 s (after the unused-rail +cleanup). The owner unlocked the lock screen by touch. **Passed.** From e78dca7bfbf4330a720229b8f993a05ce21d9e36 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Tue, 27 Oct 2026 23:31:35 -0700 Subject: [PATCH 4/7] NOTES (panther): a second stock fallback, and a new backup Getting to fastboot spent slot B's last retry again and stock recovery ran; the reset was declined and a fresh read-only backup taken. Co-Authored-By: Claude Opus 5.5 --- device/panther/NOTES.md | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index d4e204cb6..9dd1892c6 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -77,6 +77,22 @@ missing for L-36: a getvar and a byte read of the same state with no boot in between; a byte read needs Linux, so it has to be taken right before a planned reboot to fastboot. +## 2026-10-27: a second stock fallback, and a new backup + +Getting to fastboot from Arch went the same way twice: the phone came back +up in slot B instead of powering off, spent slot B's last retry, and the +next boot fell to slot A, whose stock recovery started. The data reset was +declined both times; `adb reboot bootloader` from recovery reached +fastboot (getvar: current slot A, B retry 0, not unbootable), and +`bin/omamobile flash` brought slot B back. A fresh read-only backup was +taken from Linux right after (each partition `ro=1`, found by `PARTNAME`, +checksum-matched against the phone): against the 2026-10-27 baseline, +`modem_a` and `modem_b` match, `persist`, `efs`, `efs_backup` and +`modem_userdata` differ in 100-750 bytes each, `devinfo` in 2. That copy is +the baseline now. To reach fastboot without spending a retry, hold Power +and Volume Down on the running phone until the bootloader shows, without +shutting down first. + ## 2026-10-27: Omarchy's power off reboots the phone "Shut down" in the Omarchy menu, with the USB cable to the laptop From b8e981052d38bd56cf02f2e9300c832b43c9dc77 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Thu, 29 Oct 2026 04:46:08 -0700 Subject: [PATCH 5/7] NOTES (panther): newest first The 2026-10-27 entries had landed oldest first below the stock facts, while the header says newest first. Put them in order after the 2026-10-28 entry, and note on the 10-27 power-off entry that it reboots without the cable too. Co-Authored-By: Claude Opus 5.5 --- device/panther/NOTES.md | 130 ++++++++++++++++++++-------------------- 1 file changed, 65 insertions(+), 65 deletions(-) diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index 9dd1892c6..203d805cf 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -15,6 +15,71 @@ Test kernels from local branches on slot B, each one change on top of the touch - **Power off reboots the phone.** Omarchy's "Power off" (or `systemctl poweroff`) restarts the phone within a second, which spends a slot B retry. With the reboot watchdog off, systemd-shutdown runs to the end. A shutdown hook that saved the kernel log to `userdata` ten times a second, then powered off itself (sysrq `o`, the same `kernel_power_off()`), saved nothing after the trigger: no `sysrq: Power Off`, no device `shutdown` lines, no `reboot: Power down`. So the reset comes in the first tenth of a second of the kernel's power off, or before UFS can write another frame. With `syscon-poweroff` disabled in the test DT that leaves PSCI SYSTEM_OFF, which the firmware seems to turn into a restart (the boot reason reads `watchdog/apc`). The Pixel 7 Pro has never powered off from Linux either ([SUSPEND-PLAN](../cheetah/SUSPEND-PLAN.md) P7). To get to fastboot, hold Power and Volume Down from a running phone instead; that costs no retry. - **USB doesn't come back after a replug.** A replug while Linux runs leaves the phone off the bus, as on the other Tensor phones (`AGENTS.md`): the MAX77759 opens its USB2 data switch, and nothing closes it. The Pixel 7 Pro's `omamobile-usb-rescue` needs the MAX77759's I2C bus (USI13, `10d60000`), which this board file doesn't enable yet, so for now a replug means Power and Volume Down, then a flash. +## 2026-10-27: modules-push and linux-modules-cleanup + +`kernel-modules-hook` (in the root image) enables +`linux-modules-cleanup.service`, which on every boot moves each +`/usr/lib/modules/` that is not the running kernel and that no +package owns into `/usr/lib/modules/.old`. A tree from +`bin/omamobile modules-push` is owned by no package, so a boot of any +other kernel before the new one (here, the unplanned reboot above) loses +it, and the new kernel then boots with no modules. `.old` kept only the +old release, so the pushed tree was not recoverable from it. Fix used: +`modules-push` again from the new kernel. Loading the missing modules +afterwards with `udevadm trigger` raced (ACPM probed before its mailbox, +`15 channels, 0 usable`, so `sec-pmic-acpm` failed with -ENODEV before +any PMIC write); a reboot loaded everything in order. To raise with the +tooling: push right before the reboot, or have `modules-push` warn. + +## 2026-10-27: Omarchy's power off reboots the phone + +"Shut down" in the Omarchy menu, with the USB cable to the laptop +plugged in, rebooted into slot B instead of powering off (and spent a +slot B retry). Holding the power key forced it off. Not looked into yet: +the bootloader may be powering the phone back on because the cable is in. +Until then, reach fastboot from a forced power off. (2026-10-28: it +reboots without the cable too, from the kernel's power off; see above.) + +## 2026-10-27: a second stock fallback, and a new backup + +Getting to fastboot from Arch went the same way twice: the phone came back +up in slot B instead of powering off, spent slot B's last retry, and the +next boot fell to slot A, whose stock recovery started. The data reset was +declined both times; `adb reboot bootloader` from recovery reached +fastboot (getvar: current slot A, B retry 0, not unbootable), and +`bin/omamobile flash` brought slot B back. A fresh read-only backup was +taken from Linux right after (each partition `ro=1`, found by `PARTNAME`, +checksum-matched against the phone): against the 2026-10-27 baseline, +`modem_a` and `modem_b` match, `persist`, `efs`, `efs_backup` and +`modem_userdata` differ in 100-750 bytes each, `devinfo` in 2. That copy is +the baseline now. To reach fastboot without spending a retry, hold Power +and Volume Down on the running phone until the bootloader shows, without +shutting down first. + +## 2026-10-27: devinfo slot bytes against ABL (L-36 evidence) + +devinfo is version **3.19** here (cheetah's bootctl is verified on 3.15 +only). Only the version and the slot bytes (0x30-0x37) were read, from +Linux, read-only. Byte layout per slot: retry count, then flags (bit 0 +unbootable, bit 1 successful, bit 2 active, bit 3 fastboot_ok). + +| When | `fastboot getvar` | devinfo 0x30-0x37 | +| --- | --- | --- | +| stock capture, before any of our boots | A: retry 2, successful, bootable; B: retry 3, not successful, bootable | not read | +| one boot after `--set-active=b` (boot 7) | not read | `02 0a 00 00 02 04 00 00` | +| before the T2 flash (after boots 7 and 8) | B: retry 1, not successful, bootable | not read | +| after the T2 flash (`--set-active=b`) and two boots (T2, T2b) | not read | `02 0a 00 00 01 04 00 00` | +| one more boot of slot B (an unplanned reboot), then fastboot | current slot **A**; A: retry 2, successful, bootable; B: retry 0, not successful, bootable | not read | + +Every reading agrees with the layout: slot A `02 0a` is retry 2, +successful, fastboot_ok, as getvar says; slot B counts down from 3 by one +per boot, ABL's own change, and `--set-active` resets it. The last row is +the next step of the count (1 to 0), and shows ABL's choice once B has no +retries left: A becomes current, B is not marked unbootable. Still +missing for L-36: a getvar and a byte read of the same state with no boot +in between; a byte read needs Linux, so it has to be taken right before a +planned reboot to fastboot. + ## 2026-10-27: boot 6's fallback was a stock boot, and a new backup When slot B ran out of retries (boot 6, [BRINGUP-M1.md](BRINGUP-M1.md)), ABL booted stock Android on slot A, which went to recovery because its `userdata` is now Arch. The data reset that recovery offers was **declined**: the phone was taken to fastboot with `adb reboot bootloader` and slot B flashed again. @@ -52,68 +117,3 @@ Captured on stock `CP3A.261005.002.A1` (Android 17, security patch 2026-10-05) o | Charger, TCPC, gauge | MAX77759, as on cheetah | [V] live DT | | Reserved memory | Every stock `reserved-memory` region that Google's tree describes matches gs201.dtsi. `sec_dram`, `sec_pt` and `pkvm_guest_firmware` are injected by the bootloader (not in Google's DTS); `cp_aoc_rmem` (0x197fd000) is AoC SRAM, not DRAM | [V] live DT against the built cheetah DTB; Google `gs201-rmem.dtsi` | | Framebuffer | Not described in Google's tree: ABL injects it. Not copied from cheetah's `0xfac00000`; to be read from DPP0 on the first boot | [I] | - -## 2026-10-27: devinfo slot bytes against ABL (L-36 evidence) - -devinfo is version **3.19** here (cheetah's bootctl is verified on 3.15 -only). Only the version and the slot bytes (0x30-0x37) were read, from -Linux, read-only. Byte layout per slot: retry count, then flags (bit 0 -unbootable, bit 1 successful, bit 2 active, bit 3 fastboot_ok). - -| When | `fastboot getvar` | devinfo 0x30-0x37 | -| --- | --- | --- | -| stock capture, before any of our boots | A: retry 2, successful, bootable; B: retry 3, not successful, bootable | not read | -| one boot after `--set-active=b` (boot 7) | not read | `02 0a 00 00 02 04 00 00` | -| before the T2 flash (after boots 7 and 8) | B: retry 1, not successful, bootable | not read | -| after the T2 flash (`--set-active=b`) and two boots (T2, T2b) | not read | `02 0a 00 00 01 04 00 00` | -| one more boot of slot B (an unplanned reboot), then fastboot | current slot **A**; A: retry 2, successful, bootable; B: retry 0, not successful, bootable | not read | - -Every reading agrees with the layout: slot A `02 0a` is retry 2, -successful, fastboot_ok, as getvar says; slot B counts down from 3 by one -per boot, ABL's own change, and `--set-active` resets it. The last row is -the next step of the count (1 to 0), and shows ABL's choice once B has no -retries left: A becomes current, B is not marked unbootable. Still -missing for L-36: a getvar and a byte read of the same state with no boot -in between; a byte read needs Linux, so it has to be taken right before a -planned reboot to fastboot. - -## 2026-10-27: a second stock fallback, and a new backup - -Getting to fastboot from Arch went the same way twice: the phone came back -up in slot B instead of powering off, spent slot B's last retry, and the -next boot fell to slot A, whose stock recovery started. The data reset was -declined both times; `adb reboot bootloader` from recovery reached -fastboot (getvar: current slot A, B retry 0, not unbootable), and -`bin/omamobile flash` brought slot B back. A fresh read-only backup was -taken from Linux right after (each partition `ro=1`, found by `PARTNAME`, -checksum-matched against the phone): against the 2026-10-27 baseline, -`modem_a` and `modem_b` match, `persist`, `efs`, `efs_backup` and -`modem_userdata` differ in 100-750 bytes each, `devinfo` in 2. That copy is -the baseline now. To reach fastboot without spending a retry, hold Power -and Volume Down on the running phone until the bootloader shows, without -shutting down first. - -## 2026-10-27: Omarchy's power off reboots the phone - -"Shut down" in the Omarchy menu, with the USB cable to the laptop -plugged in, rebooted into slot B instead of powering off (and spent a -slot B retry). Holding the power key forced it off. Not looked into yet: -the bootloader may be powering the phone back on because the cable is in. -Until then, reach fastboot from a forced power off. - -## 2026-10-27: modules-push and linux-modules-cleanup - -`kernel-modules-hook` (in the root image) enables -`linux-modules-cleanup.service`, which on every boot moves each -`/usr/lib/modules/` that is not the running kernel and that no -package owns into `/usr/lib/modules/.old`. A tree from -`bin/omamobile modules-push` is owned by no package, so a boot of any -other kernel before the new one (here, the unplanned reboot above) loses -it, and the new kernel then boots with no modules. `.old` kept only the -old release, so the pushed tree was not recoverable from it. Fix used: -`modules-push` again from the new kernel. Loading the missing modules -afterwards with `udevadm trigger` raced (ACPM probed before its mailbox, -`15 channels, 0 usable`, so `sec-pmic-acpm` failed with -ENODEV before -any PMIC write); a reboot loaded everything in order. To raise with the -tooling: push right before the reboot, or have `modules-push` warn. - From 8e7075252a5816d6757995e7dc3341debdd17f83 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Thu, 29 Oct 2026 04:46:08 -0700 Subject: [PATCH 6/7] device/panther: build the FT3658 touch driver, and record T6 Add panther's kernel.config with CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m, as derekross asked in linux-omarchy-mobile#109, for the touch node in review in linux-omarchy-mobile#134. NOTES records T6, one boot of that node with the driver's review fixes (#133), suspend-to-idle and LDO25M at 1.8 V: the touch loads at boot, works across the panel and unlocks the PIN screen, and the GPU and the haptics keep working. TOUCH-PLAN's status and the Pixel 7 row's Touch cell say so. Co-Authored-By: Claude Opus 5.5 --- device/panther/NOTES.md | 9 +++++++++ device/panther/TOUCH-PLAN.md | 2 +- device/panther/kernel.config | 5 +++++ 3 files changed, 15 insertions(+), 1 deletion(-) create mode 100644 device/panther/kernel.config diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index 203d805cf..4e3a4819b 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -5,6 +5,15 @@ Findings from real hardware, newest first. [BRINGUP-M1.md](BRINGUP-M1.md) is the The board file (`gs201-panther.dts`) was merged as [linux-omarchy-mobile#108](https://github.com/derekross/linux-omarchy-mobile/pull/108), and the FT3658 driver as [#109](https://github.com/derekross/linux-omarchy-mobile/pull/109); the test kernels below add one local commit each on top. +## 2026-10-29: the touch node, suspend-to-idle and LDO25M (T6) + +One boot of the board branch in review ([linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134): the GPU, the haptics, `mem_sleep_default=s2idle`, LDO25M fixed at 1.8 V and the final touch node), merged with the driver's review fixes ([#133](https://github.com/derekross/linux-omarchy-mobile/pull/133): DMA-safe SPI buffers, `fsleep()`), with `CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m` from this directory's `kernel.config`. The modules were pushed before the boot. + +- `/sys/power/mem_sleep` reads `[s2idle] deep`. +- `vdd-ts` reads 1.8 V and `avdd-ts` 3.0 V, each with the touch controller as its only consumer. +- `focaltech_ft3658` loads at boot by itself: chip ID 0x56, firmware 0x16. Touches across the panel, the four corners included, drive Hyprland and unlock the PIN screen, with no SPI or CRC errors. +- The GPU and the haptics keep working (a 300 ms rumble was felt), and ACPM has 14 of its 15 channels usable. + ## 2026-10-28: the GPU and the haptics, power off, and USB after a replug Test kernels from local branches on slot B, each one change on top of the touch test kernel; one boot each. diff --git a/device/panther/TOUCH-PLAN.md b/device/panther/TOUCH-PLAN.md index e7e796e69..ed7874e6c 100644 --- a/device/panther/TOUCH-PLAN.md +++ b/device/panther/TOUCH-PLAN.md @@ -1,7 +1,7 @@ # Pixel 7 (panther) touch: the supplies, then the chip ID -Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Autoload passed too: the touch comes up on its own at boot. +Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Autoload passed too: the touch comes up on its own at boot. T6 (2026-10-29) booted the final node in review, [linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134), with the driver's review fixes, [#133](https://github.com/derekross/linux-omarchy-mobile/pull/133): it passes the same way ([NOTES.md](NOTES.md)). The touch controller is a FocalTech FT3658 (chip ID 0x56) on USI0 SPI, wired as the Pixel 7 Pro's S3908. A local driver diff --git a/device/panther/kernel.config b/device/panther/kernel.config new file mode 100644 index 000000000..073f41a8a --- /dev/null +++ b/device/panther/kernel.config @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: MIT +# Panther-only kernel options, merged after kernel/configs/omamobile.config. +# The FocalTech FT3658 touchscreen on USI0 SPI (TOUCH-PLAN.md), a module +# while it is brought up; the board's node is kernel #134. +CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m From d9bb1d546b1a321bb35388a425dbdb7dba960363 Mon Sep 17 00:00:00 2001 From: Alejandro Garcia Date: Sat, 31 Oct 2026 13:53:57 -0700 Subject: [PATCH 7/7] device/panther: the FT3658 option in the shared fragment, and #149's review points Move CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m to kernel/configs/omamobile.config, next to the other touchscreen options, and drop device/panther/kernel.config: that file is for options that would break another device, and the release kernel never reads it; the driver binds only to its own compatible. NOTES' T6 entry names its kernel release and how it relates to the PR heads (#134 at ace1ba57fbe7, #133 at fdc4f4403103: the build's local merge has the same tree), and that the option came from the shared fragment. The second stock fallback entry says where the changed bytes sit: the same areas as after boot 6. TOUCH-PLAN's status line carries T6's date, and BRINGUP-M1's status says the FT3658 never answered because of its supplies, pointing at TOUCH-PLAN's result. Co-Authored-By: Claude Opus 5.5 --- device/panther/BRINGUP-M1.md | 4 +++- device/panther/NOTES.md | 11 +++++++++-- device/panther/TOUCH-PLAN.md | 2 +- device/panther/kernel.config | 5 ----- kernel/configs/omamobile.config | 4 ++++ 5 files changed, 17 insertions(+), 9 deletions(-) delete mode 100644 device/panther/kernel.config diff --git a/device/panther/BRINGUP-M1.md b/device/panther/BRINGUP-M1.md index 47a412461..9234e50c9 100644 --- a/device/panther/BRINGUP-M1.md +++ b/device/panther/BRINGUP-M1.md @@ -6,7 +6,9 @@ Hyprland and the Omarchy lock screen, on simpledrm over the bootloader's framebuffer. SSH answers at `172.16.42.1`, and `bin/omamobile push` rewrites slot B and reboots into it with no fastboot. The power and volume keys work (the power key wakes the screen); there was no touch yet: -the FocalTech FT3658 on SPI0 had no mainline driver. Kernel 7.3.0-rc6, +the FocalTech FT3658 on SPI0 never answered in boots 6 to 8, with the +driver loaded, because its supplies were off; [TOUCH-PLAN.md](TOUCH-PLAN.md) +switched them on, and touch works from its T4 on. Kernel 7.3.0-rc6, local branch `panther-first-light` on `omamobile-int`; the board file was merged into `omamobile` as [linux-omarchy-mobile#108](https://github.com/derekross/linux-omarchy-mobile/pull/108), and the FT3658 driver as diff --git a/device/panther/NOTES.md b/device/panther/NOTES.md index 4e3a4819b..f3803faa2 100644 --- a/device/panther/NOTES.md +++ b/device/panther/NOTES.md @@ -7,7 +7,7 @@ The board file (`gs201-panther.dts`) was merged as [linux-omarchy-mobile#108](ht ## 2026-10-29: the touch node, suspend-to-idle and LDO25M (T6) -One boot of the board branch in review ([linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134): the GPU, the haptics, `mem_sleep_default=s2idle`, LDO25M fixed at 1.8 V and the final touch node), merged with the driver's review fixes ([#133](https://github.com/derekross/linux-omarchy-mobile/pull/133): DMA-safe SPI buffers, `fsleep()`), with `CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m` from this directory's `kernel.config`. The modules were pushed before the boot. +One boot of kernel `7.3.0-rc6-omamobile-push1-g8d46192a301e`: the board branch of [linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134) (the GPU, the haptics, `mem_sleep_default=s2idle`, LDO25M fixed at 1.8 V and the final touch node, since merged) merged with the driver's review fixes, [#133](https://github.com/derekross/linux-omarchy-mobile/pull/133) (DMA-safe SPI buffers, `fsleep()`), with `CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m` from `kernel/configs/omamobile.config`. The build's local merge, `8d46192a3`, predates the commit messages' test paragraphs and sign-offs; its tree is the same as the merge of the PR heads, #134 at `ace1ba57fbe7` and #133 at `fdc4f4403103` (both `0eb5d6eb`). The modules were pushed before the boot. - `/sys/power/mem_sleep` reads `[s2idle] deep`. - `vdd-ts` reads 1.8 V and `avdd-ts` 3.0 V, each with the touch controller as its only consumer. @@ -60,7 +60,14 @@ fastboot (getvar: current slot A, B retry 0, not unbootable), and taken from Linux right after (each partition `ro=1`, found by `PARTNAME`, checksum-matched against the phone): against the 2026-10-27 baseline, `modem_a` and `modem_b` match, `persist`, `efs`, `efs_backup` and -`modem_userdata` differ in 100-750 bytes each, `devinfo` in 2. That copy is +`modem_userdata` differ in 100-750 bytes each, `devinfo` in 2. Those bytes +sit in the same places as after boot 6 (compared block by block, both +backups read-only): in each F2FS partition, the checkpoint (the other of +its two copies this time) and the root inode's node block written just +before it; in `persist`, the ext4 superblock, the first five journal blocks +and one inode table block, where only the ctime of `/sensors` and +`/sensors/registry` moved, the same two inodes as after boot 6. No data +block changed in any of them, so recovery's mount explains it. That copy is the baseline now. To reach fastboot without spending a retry, hold Power and Volume Down on the running phone until the bootloader shows, without shutting down first. diff --git a/device/panther/TOUCH-PLAN.md b/device/panther/TOUCH-PLAN.md index ed7874e6c..3093b4eee 100644 --- a/device/panther/TOUCH-PLAN.md +++ b/device/panther/TOUCH-PLAN.md @@ -1,7 +1,7 @@ # Pixel 7 (panther) touch: the supplies, then the chip ID -Status (2026-10-27): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Autoload passed too: the touch comes up on its own at boot. T6 (2026-10-29) booted the final node in review, [linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134), with the driver's review fixes, [#133](https://github.com/derekross/linux-omarchy-mobile/pull/133): it passes the same way ([NOTES.md](NOTES.md)). +Status (2026-10-29): all gates passed (T2 stopped on its voltage rule, which T2b fixed): the FT3658 answers with chip ID 0x56 once its supplies are on, reports touches with the axes right, and the lock screen PIN was entered by touch. Autoload passed too: the touch comes up on its own at boot. T6 (2026-10-29) booted the final node, [linux-omarchy-mobile#134](https://github.com/derekross/linux-omarchy-mobile/pull/134), since merged, with the driver's review fixes, [#133](https://github.com/derekross/linux-omarchy-mobile/pull/133): it passes the same way ([NOTES.md](NOTES.md)). The touch controller is a FocalTech FT3658 (chip ID 0x56) on USI0 SPI, wired as the Pixel 7 Pro's S3908. A local driver diff --git a/device/panther/kernel.config b/device/panther/kernel.config deleted file mode 100644 index 073f41a8a..000000000 --- a/device/panther/kernel.config +++ /dev/null @@ -1,5 +0,0 @@ -# SPDX-License-Identifier: MIT -# Panther-only kernel options, merged after kernel/configs/omamobile.config. -# The FocalTech FT3658 touchscreen on USI0 SPI (TOUCH-PLAN.md), a module -# while it is brought up; the board's node is kernel #134. -CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m diff --git a/kernel/configs/omamobile.config b/kernel/configs/omamobile.config index 5bb1b64c3..4ba602d8a 100644 --- a/kernel/configs/omamobile.config +++ b/kernel/configs/omamobile.config @@ -99,6 +99,10 @@ CONFIG_EXYNOS_USI=y CONFIG_SPI_S3C64XX=y CONFIG_INPUT_TOUCHSCREEN=y CONFIG_TOUCHSCREEN_SYNAPTICS_TCM_SPI=y +# Pixel 7 touchscreen: a FocalTech FT3658 on USI0 SPI +# (device/panther/TOUCH-PLAN.md). A module; it binds only to its own +# compatible. +CONFIG_TOUCHSCREEN_FOCALTECH_FT3658=m # Pixel Fold cover touch id reader (device/felix/COVER-TOUCH-PLAN.md). # Module. The node stays disabled; cover-touch.dtso enables it, and # omamobile-device-felix blacklists salixp so it does not autoload.