From 2030f029e558318fc593fd9b6b2c09fbc2f70720 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 01/20] omamobile-device-tegu 0.1.3: the Pixel 9a device package, storage guard and USB networking The thin package the 9a has run since the first native boots: the storage guard's udev policy, the usb0 address, NetworkManager kept off the link, the device descriptor, the mirrorlist and reboot-to, plus the torch, DMA-heap and feedbackd rules. It starts no hardware and marks no slot. --- packages/omamobile-device-tegu/LICENSE | 21 +++++ packages/omamobile-device-tegu/PKGBUILD | 27 +++++++ packages/omamobile-device-tegu/README.md | 26 +++++++ packages/omamobile-device-tegu/build | 8 ++ packages/omamobile-device-tegu/device.conf | 35 +++++++++ .../etc/omamobile-mirrorlist | 6 ++ packages/omamobile-device-tegu/install | 29 +++++++ .../omamobile-device-tegu.install | 24 ++++++ .../omamobile-device-tegu/src/reboot-to.c | 24 ++++++ packages/omamobile-device-tegu/test/all | 76 +++++++++++++++++++ .../conf.d/10-omamobile-usb.conf | 5 ++ .../systemd/network/20-usb-gadget.network | 14 ++++ .../udev/rules.d/10-omamobile-storage.rules | 8 ++ .../udev/rules.d/20-omamobile-torch.rules | 5 ++ .../udev/rules.d/70-omamobile-dma-heap.rules | 8 ++ .../udev/rules.d/72-omamobile-feedbackd.rules | 7 ++ packages/omamobile-device-tegu/version | 1 + 17 files changed, 324 insertions(+) create mode 100644 packages/omamobile-device-tegu/LICENSE create mode 100644 packages/omamobile-device-tegu/PKGBUILD create mode 100644 packages/omamobile-device-tegu/README.md create mode 100755 packages/omamobile-device-tegu/build create mode 100644 packages/omamobile-device-tegu/device.conf create mode 100644 packages/omamobile-device-tegu/etc/omamobile-mirrorlist create mode 100755 packages/omamobile-device-tegu/install create mode 100644 packages/omamobile-device-tegu/omamobile-device-tegu.install create mode 100644 packages/omamobile-device-tegu/src/reboot-to.c create mode 100755 packages/omamobile-device-tegu/test/all create mode 100644 packages/omamobile-device-tegu/vendor/NetworkManager/conf.d/10-omamobile-usb.conf create mode 100644 packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/10-omamobile-storage.rules create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/20-omamobile-torch.rules create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/70-omamobile-dma-heap.rules create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/72-omamobile-feedbackd.rules create mode 100644 packages/omamobile-device-tegu/version diff --git a/packages/omamobile-device-tegu/LICENSE b/packages/omamobile-device-tegu/LICENSE new file mode 100644 index 000000000..aecd59624 --- /dev/null +++ b/packages/omamobile-device-tegu/LICENSE @@ -0,0 +1,21 @@ +MIT License + +Copyright (c) 2026 Derek Ross + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. diff --git a/packages/omamobile-device-tegu/PKGBUILD b/packages/omamobile-device-tegu/PKGBUILD new file mode 100644 index 000000000..11e98e6f8 --- /dev/null +++ b/packages/omamobile-device-tegu/PKGBUILD @@ -0,0 +1,27 @@ +# SPDX-License-Identifier: MIT +# Maintainer: Derek Ross +pkgname=omamobile-device-tegu +pkgver=$(cat "${startdir:-.}/version") +pkgrel=1 +pkgdesc="Omarchy Mobile support for the Google Pixel 9a (tegu): storage guard and USB networking" +arch=(aarch64) +url="https://github.com/derekross/omarchy-mobile" +license=(MIT) +depends=(bash systemd omamobile-keyring) +provides=(omamobile-device) +conflicts=(omamobile-device) +backup=(etc/pacman.d/omamobile-mirrorlist) +install=omamobile-device-tegu.install +makedepends=(gcc) + +build() { + "$startdir/build" "$srcdir/build" +} + +check() { + "$startdir/test/all" +} + +package() { + "$startdir/install" "$pkgdir" "$srcdir/build" +} diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md new file mode 100644 index 000000000..770a586d6 --- /dev/null +++ b/packages/omamobile-device-tegu/README.md @@ -0,0 +1,26 @@ + +# Pixel 9a device policy + +Experimental tegu support: the unchanged storage guard's udev policy, USB +networking and the device descriptor. The package does not start firmware, +radio, audio, charging, fingerprint or other unqualified hardware services. +It contains no vendor files or per-phone data. + +The CS40L26 feedbackd rule is reused unchanged from the Pixel7 Pro package. +It identifies an already registered force-feedback input and grants desktop +access; it does not load the driver, power hardware or send a vibration. + +The LM3644 torch rule grants the existing video group write access only to +`white:torch/brightness` owned by that driver, for the normal quick-settings +`brightnessctl` path. It neither loads a driver nor turns on the light. The +kernel board current cap remains authoritative. On the local d816 kernel, +the owner has confirmed rear-light on/off through the existing tile at the +single tested minimum step; brightness tuning is still separate. + +The qualified storage path and remaining standalone-boot limitation are in +`device/tegu/UFS-PLAN.md` and `BOOT-PLAN.md`. The temporary recovery module +is a bring-up tool, not production reboot support. `reboot-to` by itself +does not provide the tested guarded bootloader route on this kernel. + +Build with `makepkg`; `test/all` checks the staged storage and USB policy. +The package is self-contained and provides/conflicts with `omamobile-device`. diff --git a/packages/omamobile-device-tegu/build b/packages/omamobile-device-tegu/build new file mode 100755 index 000000000..5dacf6c60 --- /dev/null +++ b/packages/omamobile-device-tegu/build @@ -0,0 +1,8 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Compile reboot-to into BUILDDIR (default: build). +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")" +out=${1:-build} +mkdir -p "$out" +${CC:-cc} ${CFLAGS:--O2} -Wall -Wextra -o "$out/reboot-to" src/reboot-to.c ${LDFLAGS:-} diff --git a/packages/omamobile-device-tegu/device.conf b/packages/omamobile-device-tegu/device.conf new file mode 100644 index 000000000..80bc14ffc --- /dev/null +++ b/packages/omamobile-device-tegu/device.conf @@ -0,0 +1,35 @@ +# SPDX-License-Identifier: MIT +# Pixel 9a first-light descriptor. See NOTES.md. +# Data only: this is a staged port, not an installable release. +CODENAME=tegu +NAME="Pixel 9a" +USB_PRODUCT="Pixel 9a (M1)" +SOC=zumapro +COMPATIBLE="google,zumapro-tegu" +PACKAGE=omamobile-device-tegu + +DTB=exynos/google/zumapro-tegu.dtb +# Deliberately no SOC_IDS: this loader uses /soc_compatible matching. +# Stock IPOP candidates: product_id 0x9875, major 0, minor 0/1, pkg_mode 1 +# (NOTES.md). Shared support is reviewed separately in #27. +# SOC_COMPATIBLE="a0-ipop:0x9875:0:0:1 a1-ipop:0x9875:0:1:1" +# Keep both selector keys unset: #27 accepts either one. Activate only +# with the public DTB and required AVB footer values in a later PR. +# Do not use a dummy soc_id or borrow gs201's values to bypass the check. + +DTBO_SOURCE=tegu-m1-overlay.dts +DTBO_BOARD_ID=0x60b06 +DTBO_BOARD_REV=0x10000 +VENDOR_CMDLINE="console=tty0 clk_ignore_unused pd_ignore_unused printk.devkmsg=on bootconfig" +VENDOR_PAGESIZE=0x800 +# Official BP3A init_boot header; both OTA/AVB and stock-A checks: NOTES.md. +OS_VERSION=16.0.0 +OS_PATCH_LEVEL=2025-09 +BOOT_METHOD=android-v4 +BOOT_IMAGES="boot init_boot vendor_boot vendor_kernel_boot dtbo" +UFS_RW="boot_b init_boot_b vendor_boot_b vendor_kernel_boot_b dtbo_b userdata" + +INITRAMFS_INIT=/initramfs/init +PANEL_SIZE=1080x2424 +# No BRINGUP_PROBE or SIMPLEFB_TRIG_CON: those gs201 display accesses do +# not apply here. The prototype's G4 trigger handling needs its own change. diff --git a/packages/omamobile-device-tegu/etc/omamobile-mirrorlist b/packages/omamobile-device-tegu/etc/omamobile-mirrorlist new file mode 100644 index 000000000..b272b8fca --- /dev/null +++ b/packages/omamobile-device-tegu/etc/omamobile-mirrorlist @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: MIT +# Where pacman gets Omarchy Mobile's packages ([omamobile] in /etc/pacman.conf). +# While the project is private, the repository lives on the phone and is +# filled from the build machine (tools/repo-push). Once it's public, this +# becomes a URL, delivered by a package update. +Server = file:///var/lib/omamobile/repo/$arch diff --git a/packages/omamobile-device-tegu/install b/packages/omamobile-device-tegu/install new file mode 100755 index 000000000..93d1117ba --- /dev/null +++ b/packages/omamobile-device-tegu/install @@ -0,0 +1,29 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Stage omamobile-device-tegu into DESTDIR. Nothing is enabled that would +# mark a slot successful or start the radio, the charger, or a camera. +# ./install DESTDIR [BUILDDIR] +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")" +destination=${1:?Usage: ./install DESTDIR [BUILDDIR]} +build=${2:-build} +[[ $destination == /* && $destination != "/" ]] || { echo "Use an absolute staging directory" >&2; exit 1; } +install -Dm755 "$build/reboot-to" "$destination/usr/bin/reboot-to" +install -Dm644 vendor/udev/rules.d/10-omamobile-storage.rules \ + "$destination/usr/lib/udev/rules.d/10-omamobile-storage.rules" +install -Dm644 vendor/udev/rules.d/72-omamobile-feedbackd.rules \ + "$destination/usr/lib/udev/rules.d/72-omamobile-feedbackd.rules" +install -Dm644 vendor/udev/rules.d/20-omamobile-torch.rules \ + "$destination/usr/lib/udev/rules.d/20-omamobile-torch.rules" +install -Dm644 vendor/udev/rules.d/70-omamobile-dma-heap.rules \ + "$destination/usr/lib/udev/rules.d/70-omamobile-dma-heap.rules" +install -Dm644 vendor/systemd/network/20-usb-gadget.network \ + "$destination/usr/lib/systemd/network/20-usb-gadget.network" +install -Dm644 vendor/NetworkManager/conf.d/10-omamobile-usb.conf \ + "$destination/usr/lib/NetworkManager/conf.d/10-omamobile-usb.conf" +install -Dm644 device.conf \ + "$destination/usr/share/omamobile/device.conf" +install -Dm644 etc/omamobile-mirrorlist \ + "$destination/etc/pacman.d/omamobile-mirrorlist" +install -Dm644 LICENSE "$destination/usr/share/licenses/omamobile-device-tegu/LICENSE" +install -Dm644 version "$destination/usr/share/omamobile/device-tegu/version" diff --git a/packages/omamobile-device-tegu/omamobile-device-tegu.install b/packages/omamobile-device-tegu/omamobile-device-tegu.install new file mode 100644 index 000000000..60517fa32 --- /dev/null +++ b/packages/omamobile-device-tegu/omamobile-device-tegu.install @@ -0,0 +1,24 @@ +# SPDX-License-Identifier: MIT +# Add the [omamobile] repository to pacman.conf, before [omarchy] so its +# builds (e.g. Mesa with Mali-G710 support) win. Added once; never removed +# automatically, so a user's edits stay. +add_repository() { + local conf=/etc/pacman.conf + grep -q '^\[omamobile\]' "$conf" && return + local entry='[omamobile]\nInclude = /etc/pacman.d/omamobile-mirrorlist\n' + if grep -q '^\[omarchy\]' "$conf"; then + sed -i "0,/^\[omarchy\]/s//${entry//\//\\/}\n[omarchy]/" "$conf" + elif grep -q '^\[core\]' "$conf"; then + sed -i "0,/^\[core\]/s//${entry//\//\\/}\n[core]/" "$conf" + else + printf '\n[omamobile]\nInclude = /etc/pacman.d/omamobile-mirrorlist\n' >>"$conf" + fi +} + +post_install() { + add_repository +} + +post_upgrade() { + add_repository +} diff --git a/packages/omamobile-device-tegu/src/reboot-to.c b/packages/omamobile-device-tegu/src/reboot-to.c new file mode 100644 index 000000000..f79e948ed --- /dev/null +++ b/packages/omamobile-device-tegu/src/reboot-to.c @@ -0,0 +1,24 @@ +// SPDX-License-Identifier: MIT +// reboot-to: reboot with a reason the Pixel bootloader understands, +// e.g. "reboot-to bootloader" lands in fastboot. +// The kernel's syscon-reboot-mode driver stores the reason in the PMU +// before the reset; busybox reboot can't pass one. +#include +#include +#include + +/* From (stable ABI); musl's sysroot has no kernel headers */ +#define LINUX_REBOOT_MAGIC1 0xfee1dead +#define LINUX_REBOOT_MAGIC2 672274793 +#define LINUX_REBOOT_CMD_RESTART2 0xA1B2C3D4 + +int main(int argc, char **argv) +{ + const char *reason = argc > 1 ? argv[1] : "bootloader"; + + sync(); + syscall(SYS_reboot, LINUX_REBOOT_MAGIC1, LINUX_REBOOT_MAGIC2, + LINUX_REBOOT_CMD_RESTART2, reason); + perror("reboot"); + return 1; +} diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all new file mode 100755 index 000000000..885263848 --- /dev/null +++ b/packages/omamobile-device-tegu/test/all @@ -0,0 +1,76 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Tests for omamobile-device-tegu. reboot-to is compiled from +# src/reboot-to.c in this directory, so a copied package still builds. +# Do not execute reboot-to. +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +fail() { printf 'not ok - %s\n' "$1" >&2; exit 1; } +pass() { printf 'ok - %s\n' "$1"; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT + +bash -n build || fail "build parses" +bash -n install || fail "install parses" +pass "scripts parse" + +grep -q '\.\./\.\.' build install && fail "build or install reaches outside the package directory" +pass "build and install stay inside the package directory" + +if [[ -f ../omamobile-device-cheetah/src/reboot-to.c ]]; then + cmp -s src/reboot-to.c ../omamobile-device-cheetah/src/reboot-to.c \ + || fail "reboot-to.c drifted from the cheetah package" + pass "reboot-to.c matches the cheetah source" + cmp -s vendor/udev/rules.d/10-omamobile-storage.rules \ + ../omamobile-device-cheetah/vendor/udev/rules.d/10-omamobile-storage.rules \ + || fail "the storage rule drifted from the cheetah package" + pass "the storage rule matches the cheetah package" + cmp -s etc/omamobile-mirrorlist ../omamobile-device-cheetah/etc/omamobile-mirrorlist \ + || fail "the mirrorlist drifted from the cheetah package" + pass "the mirrorlist matches the cheetah package" + cmp -s omamobile-device-tegu.install \ + ../omamobile-device-cheetah/omamobile-device-cheetah.install \ + || fail "the pacman install script drifted from the cheetah package" + pass "the pacman install script matches the cheetah package" +fi +grep -q 'omamobile-keyring' PKGBUILD || fail "pacman needs omamobile-keyring" +grep -q '^install=omamobile-device-tegu.install$' PKGBUILD \ + || fail "the pacman install script is not named" +[[ -f etc/omamobile-mirrorlist && -f omamobile-device-tegu.install ]] \ + || fail "the pacman mirrorlist or install script is missing" +pass "pacman gets the keyring, the mirrorlist, and the install script" +if [[ -f ../../device/tegu/device.conf ]]; then + cmp -s device.conf ../../device/tegu/device.conf \ + || fail "device.conf drifted from the descriptor" + pass "device.conf matches the descriptor" +fi + +./build "$tmp/build" || fail "reboot-to compiles" +[[ -x $tmp/build/reboot-to ]] || fail "reboot-to is not executable" +pass "reboot-to compiles" + +./install "$tmp/root" "$tmp/build" +for f in usr/bin/reboot-to \ + usr/lib/udev/rules.d/10-omamobile-storage.rules \ + usr/lib/systemd/network/20-usb-gadget.network \ + usr/lib/NetworkManager/conf.d/10-omamobile-usb.conf \ + usr/share/omamobile/device.conf \ + usr/share/omamobile/device-tegu/version \ + usr/share/licenses/omamobile-device-tegu/LICENSE \ + etc/pacman.d/omamobile-mirrorlist; do + [[ -f $tmp/root/$f ]] || fail "install stages $f" +done +grep -q 'KERNEL=="sda", MODE="0000"' "$tmp/root/usr/lib/udev/rules.d/10-omamobile-storage.rules" \ + || fail "the whole-disk node is made inaccessible" +grep -q 'ID_PART_ENTRY_NAME}!="userdata"' "$tmp/root/usr/lib/udev/rules.d/10-omamobile-storage.rules" \ + || fail "userdata is the only partition desktop tools may see" +grep -q 'Address=172.16.42.1/24' "$tmp/root/usr/lib/systemd/network/20-usb-gadget.network" \ + || fail "usb0 keeps the initramfs address" +grep -q '^PACKAGE=omamobile-device-tegu$' "$tmp/root/usr/share/omamobile/device.conf" \ + || fail "the installed descriptor is tegu" +for absent in usr/bin/omamobile-boot-success usr/bin/omamobile-bootctl \ + usr/lib/systemd/system/omamobile-boot-success.service \ + usr/bin/omamobile-wifi-start usr/bin/omamobile-charge-guard \ + usr/bin/omamobile-camera-start; do + [[ ! -e $tmp/root/$absent ]] || fail "install includes $absent" +done +pass "install stages the udev rule, the USB link, and the mirrorlist, and nothing that marks a slot or starts hardware" diff --git a/packages/omamobile-device-tegu/vendor/NetworkManager/conf.d/10-omamobile-usb.conf b/packages/omamobile-device-tegu/vendor/NetworkManager/conf.d/10-omamobile-usb.conf new file mode 100644 index 000000000..252de6e15 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/NetworkManager/conf.d/10-omamobile-usb.conf @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: the USB gadget link to the host (usb0) is configured by +# systemd-networkd; NetworkManager manages everything else (Wi-Fi later). +[keyfile] +unmanaged-devices=interface-name:usb0 diff --git a/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network b/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network new file mode 100644 index 000000000..4e30191f6 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network @@ -0,0 +1,14 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: USB networking to the host. The initramfs sets up the USB +# gadget (CDC NCM) before handing over to Arch; this gives it an address. +# The host shares its internet over this link (NetworkManager "shared" +# mode on 172.16.42.2), so it is also the gateway. DNS goes to public resolvers. +[Match] +Name=usb0 + +[Network] +Address=172.16.42.1/24 +Gateway=172.16.42.2 +DNS=1.1.1.1 +DNS=9.9.9.9 +ConfigureWithoutCarrier=yes diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/10-omamobile-storage.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/10-omamobile-storage.rules new file mode 100644 index 000000000..89c02cf30 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/10-omamobile-storage.rules @@ -0,0 +1,8 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile storage guard, continued in Arch. The initramfs marks every UFS +# partition read-only except slot B's boot images and userdata, and removes +# the whole-disk node of the main LUN. udev would recreate /dev/sda, so make +# the whole-disk node inaccessible and keep desktop tools from touching any +# Android partition. +SUBSYSTEM=="block", KERNEL=="sda", MODE="0000", ENV{UDISKS_IGNORE}="1" +SUBSYSTEM=="block", KERNEL=="sd[a-d][0-9]*", ENV{ID_PART_ENTRY_NAME}!="userdata", ENV{UDISKS_IGNORE}="1" diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/20-omamobile-torch.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/20-omamobile-torch.rules new file mode 100644 index 000000000..7104ff66c --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/20-omamobile-torch.rules @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: MIT +# The rear torch tile uses brightnessctl as the desktop user. Grant the +# existing video group access only to this driver's brightness attribute; +# the kernel's board current cap still limits every request. +ACTION=="add", SUBSYSTEM=="leds", KERNEL=="white:torch", DRIVERS=="lm3644", RUN+="/usr/bin/chgrp video /sys%p/brightness", RUN+="/usr/bin/chmod g+w /sys%p/brightness" diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/70-omamobile-dma-heap.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/70-omamobile-dma-heap.rules new file mode 100644 index 000000000..2341cae32 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/70-omamobile-dma-heap.rules @@ -0,0 +1,8 @@ +# SPDX-License-Identifier: MIT +# libcamera's software ISP (and so PipeWire's libcamera plugin, which runs in +# the user's WirePlumber) allocates its output buffers from the DMA-BUF heaps. +# They are root-only by default, so without this a user session saw the +# cameras but got no debayered stream ("Could not open any dma-buf provider", +# device/cheetah/CAMERA-PLAN.md CAM5.4). Give the video group and the seat's +# user the general heaps; the camera and display pools stay the kernel's. +SUBSYSTEM=="dma_heap", KERNEL=="system|reserved|default_cma_region", GROUP="video", MODE="0660", TAG+="uaccess" diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/72-omamobile-feedbackd.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/72-omamobile-feedbackd.rules new file mode 100644 index 000000000..88ad325a3 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/72-omamobile-feedbackd.rules @@ -0,0 +1,7 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: let feedbackd drive the Cirrus Logic CS40L26 haptics +# (cs40l26-vibra), in the same form as feedbackd's own 72-feedbackd.rules. +# Before 73-seat-late.rules, which turns the uaccess tag into an ACL. +ACTION=="remove", GOTO="omamobile_feedbackd_end" +SUBSYSTEM=="input", KERNEL=="event*", ENV{ID_INPUT}=="1", SUBSYSTEMS=="input", ATTRS{name}=="cs40l26_vibra", TAG+="uaccess", ENV{FEEDBACKD_TYPE}="vibra" +LABEL="omamobile_feedbackd_end" diff --git a/packages/omamobile-device-tegu/version b/packages/omamobile-device-tegu/version new file mode 100644 index 000000000..b1e80bb24 --- /dev/null +++ b/packages/omamobile-device-tegu/version @@ -0,0 +1 @@ +0.1.3 From f2a514b447c0a22374c09f900d87b948f7d589ec Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 02/20] omamobile-device-tegu: the host is a fallback route behind Wi-Fi The usb0 gateway becomes a default route at metric 2048, so Wi-Fi (NetworkManager's 600) wins when it is up and the host's shared connection still serves a first setup over USB. The copy the palmtop runs. --- .../vendor/systemd/network/20-usb-gadget.network | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network b/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network index 4e30191f6..0dc179bd0 100644 --- a/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network +++ b/packages/omamobile-device-tegu/vendor/systemd/network/20-usb-gadget.network @@ -8,7 +8,13 @@ Name=usb0 [Network] Address=172.16.42.1/24 -Gateway=172.16.42.2 DNS=1.1.1.1 DNS=9.9.9.9 ConfigureWithoutCarrier=yes + +# Prefer Wi-Fi (NetworkManager's metric 600) when it is available. Keep the +# host's shared connection as a fallback for initial setup over USB. +[Route] +Destination=0.0.0.0/0 +Gateway=172.16.42.2 +Metric=2048 From e50bfc579cc8a24b28c659e459b9a9cd677871c2 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 03/20] omamobile-device-tegu: USB replug rescue with the guarded TCPC helper The MAX77779 opens its USB data switch on unplug and leaves it open when VBUS returns, so the palmtop never re-enumerates. omamobile-tegu-usb-rescue closes it while the gadget stays unconfigured, through tegu-tcpc-switch, which checks the chip's identity and VBUS before its one write and never forces an address a driver holds. It exits once a Type-C driver owns the chip, as the cheetah rescue does. The install script now stages bin/ and vendor/ with cheetah's loops, so the units' wants links come from the package. --- packages/omamobile-device-tegu/PKGBUILD | 5 +- .../bin/omamobile-tegu-usb-rescue | 54 ++++++ packages/omamobile-device-tegu/build | 3 +- packages/omamobile-device-tegu/install | 37 ++-- .../src/tegu-tcpc-switch.c | 167 ++++++++++++++++++ .../omamobile-tegu-usb-rescue.service | 1 + .../system/omamobile-tegu-usb-rescue.service | 15 ++ 7 files changed, 259 insertions(+), 23 deletions(-) create mode 100755 packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue create mode 100644 packages/omamobile-device-tegu/src/tegu-tcpc-switch.c create mode 120000 packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-usb-rescue.service create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-usb-rescue.service diff --git a/packages/omamobile-device-tegu/PKGBUILD b/packages/omamobile-device-tegu/PKGBUILD index 11e98e6f8..3520c3d44 100644 --- a/packages/omamobile-device-tegu/PKGBUILD +++ b/packages/omamobile-device-tegu/PKGBUILD @@ -3,16 +3,17 @@ pkgname=omamobile-device-tegu pkgver=$(cat "${startdir:-.}/version") pkgrel=1 -pkgdesc="Omarchy Mobile support for the Google Pixel 9a (tegu): storage guard and USB networking" +pkgdesc="Omarchy Mobile support for the Google Pixel 9a (tegu): storage guard, USB networking, hang-safe graphics start" arch=(aarch64) url="https://github.com/derekross/omarchy-mobile" license=(MIT) -depends=(bash systemd omamobile-keyring) +depends=(bash systemd kmod i2c-tools bluez-utils omamobile-keyring) provides=(omamobile-device) conflicts=(omamobile-device) backup=(etc/pacman.d/omamobile-mirrorlist) install=omamobile-device-tegu.install makedepends=(gcc) +checkdepends=(diffutils) build() { "$startdir/build" "$srcdir/build" diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue b/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue new file mode 100755 index 000000000..71a8c29a9 --- /dev/null +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue @@ -0,0 +1,54 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Pixel 9a: close the MAX77779's USB data switch after a cable replug. The +# chip opens USBSW_CTRL (0x93) on unplug and leaves it open when VBUS +# returns, so the phone never re-enumerates (device/tegu/NOTES.md 2026-10-26). +# While the gadget is not configured and the switch is open, tegu-tcpc-switch +# connect (identity and VBUS checked, I2C_SLAVE never forced) writes 0x09 +# back. A Type-C driver that owns the chip (tcpci_maxim, since kernel #124) +# closes the switch itself: then this exits and leaves the chip to it. +# omamobile-tegu-usb-rescue [--once] (--once: one pass, for the tests) +# OMAMOBILE_TEGU_ROOT: a scratch root for the package tests; +# OMAMOBILE_TEGU_RESCUE_GRACE: seconds the gadget may stay unconfigured +# before the switch is read (default 4). +set -uo pipefail + +root=${OMAMOBILE_TEGU_ROOT:-} +if [[ -n $root ]]; then + [[ $root == /* && $root != / && -d $root ]] || exit 2 +fi +tool=${TEGU_TCPC_SWITCH:-/usr/bin/tegu-tcpc-switch} +grace=${OMAMOBILE_TEGU_RESCUE_GRACE:-4} +once=0 +[[ ${1:-} == --once ]] && once=1 +log() { echo "omamobile-tegu-usb-rescue: $*"; } + +[[ -x $tool ]] || { log "no $tool"; exit 0; } +tcpm_owns() { local d; for d in "$root"/sys/bus/i2c/devices/*-0025; do [[ -e $d/driver ]] && return 0; done; return 1; } +gadget_state() { cat "$root"/sys/class/udc/*/state 2>/dev/null | head -1; } + +open_since=0 +log "watching the gadget state and the TCPC data switch" +while :; do + if tcpm_owns; then + log "a Type-C driver owns the chip; nothing to do" + exit 0 + fi + state=$(gadget_state) + if [[ $state == configured ]]; then + open_since=0 + else + now=$(date +%s) + (( open_since == 0 )) && open_since=$now + if (( now - open_since >= grace )); then + sw=$("$tool" switch 2>/dev/null | tail -1) + if [[ -n $sw && $sw != 0x09 ]]; then + log "gadget ${state:-absent}, USBSW_CTRL $sw: closing the data switch" + "$tool" connect 2>&1 | tail -1 + open_since=$now + fi + fi + fi + ((once)) && exit 0 + sleep 2 +done diff --git a/packages/omamobile-device-tegu/build b/packages/omamobile-device-tegu/build index 5dacf6c60..dbbae3421 100755 --- a/packages/omamobile-device-tegu/build +++ b/packages/omamobile-device-tegu/build @@ -1,8 +1,9 @@ #!/bin/bash # SPDX-License-Identifier: MIT -# Compile reboot-to into BUILDDIR (default: build). +# Compile the C tools into BUILDDIR (default: build). set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")" out=${1:-build} mkdir -p "$out" ${CC:-cc} ${CFLAGS:--O2} -Wall -Wextra -o "$out/reboot-to" src/reboot-to.c ${LDFLAGS:-} +${CC:-cc} ${CFLAGS:--O2} -Wall -Wextra -o "$out/tegu-tcpc-switch" src/tegu-tcpc-switch.c ${LDFLAGS:-} diff --git a/packages/omamobile-device-tegu/install b/packages/omamobile-device-tegu/install index 93d1117ba..ed1d3f8ec 100755 --- a/packages/omamobile-device-tegu/install +++ b/packages/omamobile-device-tegu/install @@ -1,29 +1,26 @@ #!/bin/bash # SPDX-License-Identifier: MIT -# Stage omamobile-device-tegu into DESTDIR. Nothing is enabled that would -# mark a slot successful or start the radio, the charger, or a camera. -# ./install DESTDIR [BUILDDIR] +# Stage omamobile-device-tegu into DESTDIR. Never activates anything: the +# graphics start and the USB rescue are enabled by vendor wants links, as +# omarchy-mac does; the AoC start ships with no link (see README.md). +# ./install DESTDIR [BUILDDIR] (BUILDDIR holds the compiled C tools) set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")" destination=${1:?Usage: ./install DESTDIR [BUILDDIR]} build=${2:-build} [[ $destination == /* && $destination != "/" ]] || { echo "Use an absolute staging directory" >&2; exit 1; } -install -Dm755 "$build/reboot-to" "$destination/usr/bin/reboot-to" -install -Dm644 vendor/udev/rules.d/10-omamobile-storage.rules \ - "$destination/usr/lib/udev/rules.d/10-omamobile-storage.rules" -install -Dm644 vendor/udev/rules.d/72-omamobile-feedbackd.rules \ - "$destination/usr/lib/udev/rules.d/72-omamobile-feedbackd.rules" -install -Dm644 vendor/udev/rules.d/20-omamobile-torch.rules \ - "$destination/usr/lib/udev/rules.d/20-omamobile-torch.rules" -install -Dm644 vendor/udev/rules.d/70-omamobile-dma-heap.rules \ - "$destination/usr/lib/udev/rules.d/70-omamobile-dma-heap.rules" -install -Dm644 vendor/systemd/network/20-usb-gadget.network \ - "$destination/usr/lib/systemd/network/20-usb-gadget.network" -install -Dm644 vendor/NetworkManager/conf.d/10-omamobile-usb.conf \ - "$destination/usr/lib/NetworkManager/conf.d/10-omamobile-usb.conf" -install -Dm644 device.conf \ - "$destination/usr/share/omamobile/device.conf" -install -Dm644 etc/omamobile-mirrorlist \ - "$destination/etc/pacman.d/omamobile-mirrorlist" +for file in bin/*; do install -Dm755 "$file" "$destination/usr/bin/${file##*/}"; done +for tool in reboot-to tegu-tcpc-switch; do install -Dm755 "$build/$tool" "$destination/usr/bin/$tool"; done +while IFS= read -r -d '' file; do + mode=644 + [[ $file == vendor/systemd/system-sleep/* ]] && mode=755 + install -Dm$mode "$file" "$destination/usr/lib/${file#vendor/}" +done < <(find vendor -type f -print0) +while IFS= read -r -d '' file; do + install -d "$destination/usr/lib/$(dirname "${file#vendor/}")" + ln -sfn "$(readlink "$file")" "$destination/usr/lib/${file#vendor/}" +done < <(find vendor -type l -print0) +install -Dm644 device.conf "$destination/usr/share/omamobile/device.conf" +install -Dm644 etc/omamobile-mirrorlist "$destination/etc/pacman.d/omamobile-mirrorlist" install -Dm644 LICENSE "$destination/usr/share/licenses/omamobile-device-tegu/LICENSE" install -Dm644 version "$destination/usr/share/omamobile/device-tegu/version" diff --git a/packages/omamobile-device-tegu/src/tegu-tcpc-switch.c b/packages/omamobile-device-tegu/src/tegu-tcpc-switch.c new file mode 100644 index 000000000..ebfc0c2b2 --- /dev/null +++ b/packages/omamobile-device-tegu/src/tegu-tcpc-switch.c @@ -0,0 +1,167 @@ +// SPDX-License-Identifier: MIT +/* Fixed MAX77779 observations and one guarded USB data-switch restoration. */ +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define TCPC_ADDR 0x25 + +static void fail(const char *message) +{ + perror(message); + exit(EXIT_FAILURE); +} + +static uint8_t read_byte(int fd, uint8_t reg) +{ + uint8_t value; + struct i2c_msg messages[] = { + { .addr = TCPC_ADDR, .len = 1, .buf = ® }, + { .addr = TCPC_ADDR, .flags = I2C_M_RD, .len = 1, .buf = &value }, + }; + struct i2c_rdwr_ioctl_data transfer = { + .msgs = messages, + .nmsgs = 2, + }; + int ret; + + ret = ioctl(fd, I2C_RDWR, &transfer); + if (ret != 2) { + if (ret >= 0) + errno = EIO; + fail("one-byte TCPC read"); + } + return value; +} + +static void connect_data(int fd) +{ + uint8_t bytes[] = { 0x93, 0x09 }; + struct i2c_msg message = { + .addr = TCPC_ADDR, .len = sizeof(bytes), .buf = bytes, + }; + struct i2c_rdwr_ioctl_data transfer = { + .msgs = &message, .nmsgs = 1, + }; + uint8_t value; + int ret; + + if (read_byte(fd, 0x02) != 0x79 || read_byte(fd, 0x00) != 0x56 || + read_byte(fd, 0x01) != 0x04 || !(read_byte(fd, 0x1e) & 4)) { + fputs("MAX77779 identity or present VBUS not confirmed; no write\n", stderr); + exit(EXIT_FAILURE); + } + value = read_byte(fd, 0x93); + if (value == 0x09) + return; + if (value != 0x00) { + fputs("Unexpected USB switch state; no write\n", stderr); + exit(EXIT_FAILURE); + } + if (!(read_byte(fd, 0x1e) & 4)) { + fputs("VBUS disappeared; no write\n", stderr); + exit(EXIT_FAILURE); + } + fputs("Fixed USB data-switch restoration: register 0x93 <- 0x09\n", stderr); + ret = ioctl(fd, I2C_RDWR, &transfer); + if (ret != 1) { + if (ret >= 0) + errno = EIO; + fail("fixed TCPC data-switch write"); + } + if (read_byte(fd, 0x93) != 0x09) { + fputs("USB data-switch readback failed\n", stderr); + exit(EXIT_FAILURE); + } +} + +int main(int argc, char **argv) +{ + static const struct { + const char *name; + uint8_t reg; + } reads[] = { + { "vendor-low", 0x00 }, + { "vendor-high", 0x01 }, + { "product", 0x02 }, + { "power", 0x1e }, + { "switch", 0x93 }, + }; + char path[PATH_MAX], compatible[128] = { 0 }; + FILE *file; + unsigned long functions; + size_t i; + int fd, selected = -1; + int connect = argc == 2 && !strcmp(argv[1], "connect"); + + if (argc == 2) { + for (i = 0; i < sizeof(reads) / sizeof(reads[0]); i++) { + if (!strcmp(argv[1], reads[i].name)) + selected = (int)i; + } + } + if (selected < 0 && !connect) { + fputs("Usage: tegu-tcpc-switch vendor-low|vendor-high|product|power|switch|connect\n", + stderr); + return 64; + } + file = fopen("/proc/device-tree/compatible", "rb"); + if (!file) + fail("machine compatible"); + if (!fread(compatible, 1, sizeof(compatible) - 1, file) || + strcmp(compatible, "google,zumapro-tegu")) { + fputs("Expected Pixel 9a compatible\n", stderr); + return EXIT_FAILURE; + } + fclose(file); + /* + * The USI6 adapter's number depends on probe order (it was i2c-0 on + * the curated boots and is not after the integration DT): find the + * adapter whose device is 108d0000.i2c instead of assuming i2c-0. + */ + { + char dev[64] = { 0 }; + int n; + + for (n = 0; n < 64 && !dev[0]; n++) { + char link[64]; + + snprintf(link, sizeof(link), "/sys/bus/i2c/devices/i2c-%d", n); + if (realpath(link, path) && strstr(path, "/108d0000.i2c/i2c-")) + snprintf(dev, sizeof(dev), "/dev/i2c-%d", n); + } + if (!dev[0]) { + fputs("Expected USI6 I2C adapter (108d0000.i2c)\n", stderr); + return EXIT_FAILURE; + } + fd = open(dev, O_RDWR | O_CLOEXEC); + } + if (fd < 0) + fail("open qualified I2C adapter"); + /* I2C_SLAVE, never I2C_SLAVE_FORCE: refuse an address owned by a driver. */ + if (ioctl(fd, I2C_SLAVE, TCPC_ADDR) < 0) + fail("claim unbound TCPC address"); + if (ioctl(fd, I2C_FUNCS, &functions) < 0) + fail("I2C adapter functionality"); + if (!(functions & I2C_FUNC_I2C)) { + fputs("I2C combined transfers unavailable\n", stderr); + return EXIT_FAILURE; + } + if (connect) { + connect_data(fd); + puts("0x09"); + } else { + printf("0x%02x\n", read_byte(fd, reads[selected].reg)); + } + close(fd); + return EXIT_SUCCESS; +} diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-usb-rescue.service b/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-usb-rescue.service new file mode 120000 index 000000000..5619acf9c --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-usb-rescue.service @@ -0,0 +1 @@ +../omamobile-tegu-usb-rescue.service \ No newline at end of file diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-usb-rescue.service b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-usb-rescue.service new file mode 100644 index 000000000..76ea62860 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-usb-rescue.service @@ -0,0 +1,15 @@ +# SPDX-License-Identifier: MIT +# Close the USB data switch after each replug (omamobile-tegu-usb-rescue). +# Exits by itself when a Type-C driver owns the MAX77779. +[Unit] +Description=Reconnect USB after a replug (MAX77779 data switch) +After=systemd-modules-load.service + +[Service] +ExecStartPre=-/usr/bin/modprobe i2c-dev +ExecStart=/usr/bin/omamobile-tegu-usb-rescue +Restart=on-failure +RestartSec=5 + +[Install] +WantedBy=multi-user.target From d6200e20cac6296b8d97cad60e24296f1a541f8c Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 04/20] omamobile-device-tegu: hang-safe graphics start for the GPU, display, haptics, torch, gauge and Bluetooth The curated start of the bring-up, as a package guard in cheetah's form: a marker written and synced before the first modprobe, panthor, gs_decon, cs40l26_vibra, leds_lm3644, max77779_battery and hci_uart in that order, the marker naming the step so a hang is easy to place, and the next boot keeping all six off until --retry. No file-hash manifest, EGL smoke test, mask checks or bootloader dead-man: the marker is the safety net. The drivers are blacklisted for autoload without install lines, the display options (DSI commands, 120 Hz, the panel really off on DPMS) and the GPU DVFS ceiling move to modprobe.d, and the display manager waits for the guard. --- .../bin/omamobile-tegu-graphics-start | 132 ++++++++++++++++++ .../modprobe.d/omamobile-tegu-clk-acpm.conf | 5 + .../modprobe.d/omamobile-tegu-display.conf | 7 + .../modprobe.d/omamobile-tegu-guards.conf | 16 +++ .../10-omamobile-tegu.conf | 7 + .../omamobile-tegu-graphics.service | 1 + .../system/omamobile-tegu-graphics.service | 21 +++ 7 files changed, 189 insertions(+) create mode 100755 packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start create mode 100644 packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-clk-acpm.conf create mode 100644 packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf create mode 100644 packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-guards.conf create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/display-manager.service.d/10-omamobile-tegu.conf create mode 120000 packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-graphics.service create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start new file mode 100755 index 000000000..05155041c --- /dev/null +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start @@ -0,0 +1,132 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Bring up the Pixel 9a's GPU, display, haptics, torch, battery gauge and +# Bluetooth controller at boot, hang-safe. All six drivers are in the kernel +# tree and blacklisted for autoload (modprobe.d/omamobile-tegu-guards.conf), +# so this script is the only thing that loads them, in this order, after a +# marker has been written and synced. If the phone hangs in one of the steps, +# the next boot finds the marker, leaves all of it alone and says so; after +# reading the previous boot's log, run "omamobile-tegu-graphics-start --retry". +# +# This is the curated start of the bring-up (device/tegu/bringup) without its +# file-hash manifest, EGL smoke test, mask checks, sddm start and bootloader +# dead-man: the marker is the safety net, as on the Pixel 7 Pro. +# omamobile-tegu-graphics-start [--retry] +# OMAMOBILE_TEGU_ROOT: a scratch root for the package tests (fake /sys, /proc). +set -uo pipefail + +root=${OMAMOBILE_TEGU_ROOT:-} +if [[ -n $root ]]; then + [[ $root == /* && $root != / && -d $root ]] || exit 2 +fi +state=$root/var/lib/omamobile +marker=$state/tegu-graphics-started +dt=$root/proc/device-tree +gpu=$root/sys/bus/platform/devices/1f000000.gpu +decon=$root/sys/bus/platform/devices/19470000.display-controller +torch=$root/sys/class/leds/white:torch +gauge=$root/sys/class/power_supply/max77779-battery +haptics_node=$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43 +torch_node=$dt/soc/usi@108b00c0/i2c@108b0000/led-controller@63 +bluetooth_node=$dt/soc/usi@155d00c0/serial@155d0000/bluetooth + +log() { echo "omamobile-tegu-graphics: $*"; } +loaded() { [[ -d $root/sys/module/$1 ]]; } +# A device-tree node is enabled when it exists and has no status or "okay". +enabled() { + [[ -d $1 ]] || return 1 + [[ -e $1/status ]] || return 0 + case $(tr -d '\0' < "$1/status") in okay|ok) return 0 ;; *) return 1 ;; esac +} +# Wait for a device to bind to a driver: device dir, driver name, quarter seconds. +bound() { + local i + for ((i = 0; i < $3; i++)); do + [[ $(readlink -f "$1/driver" 2>/dev/null) == */$2 ]] && return 0 + sleep 0.25 + done + return 1 +} + +case ${1:-} in + '') ;; + --retry) rm -f "$marker" ;; + *) log "usage: omamobile-tegu-graphics-start [--retry]"; exit 2 ;; +esac + +if [[ -e $marker ]]; then + log "the last start ($(cat "$marker" 2>/dev/null)) never finished; the GPU, display, haptics, torch, gauge and Bluetooth stay off this boot." + log "after checking the previous boot's log, retry with: omamobile-tegu-graphics-start --retry" + exit 0 +fi +if [[ ! -e $gpu ]]; then + log "no GPU in this DTB; nothing to do" + exit 0 +fi +# Once per boot: a loaded driver is never unloaded or retried (AGENTS.md). +if loaded panthor || loaded gs_decon; then + log "already started this boot" + exit 0 +fi + +mkdir -p "$state" +# The marker names the step in progress, so a hang is easy to place. +phase() { printf '%s %s\n' "$(date -Is)" "$1" > "$marker"; sync -f "$marker" 2>/dev/null || sync; } +finish() { rm -f "$marker"; sync -f "$state" 2>/dev/null || sync; } +fail() { log "$*"; log "the marker stays: the next boot skips this start until --retry"; exit 1; } + +phase gpu +modprobe panthor || fail "panthor failed" +bound "$gpu" panthor 60 || fail "the GPU did not bind to panthor" + +# stage=2 is the full driver (the lower stages are bring-up probes); the +# refresh rate and the panel-off behaviour come from modprobe.d. +phase display +modprobe gs_decon stage=2 fbdev=0 || fail "gs_decon failed" +bound "$decon" gs-decon 60 || fail "the display controller did not bind to gs-decon" +up="GPU, display" + +if enabled "$haptics_node"; then + phase haptics + modprobe cs40l26_vibra || fail "cs40l26_vibra failed" + up+=", haptics" +else + log "no haptics in this DTB; skipped" +fi + +if enabled "$torch_node"; then + phase torch + modprobe leds_lm3644 || fail "leds_lm3644 failed" + for _ in $(seq 40); do [[ -e $torch/brightness ]] && break; sleep 0.25; done + if [[ ! -e $torch/brightness ]]; then + log "no white:torch LED after the driver loaded (see dmesg)" + elif [[ $(cat "$torch/brightness") != 0 ]]; then + fail "the torch came up lit; it must start off" + else + up+=", torch" + fi +else + log "no torch in this DTB; skipped" +fi + +phase gauge +modprobe max77779_battery || fail "max77779_battery failed" +for _ in $(seq 40); do [[ -e $gauge/capacity ]] && break; sleep 0.25; done +if [[ -e $gauge/capacity ]]; then + up+=", gauge $(cat "$gauge/capacity") %" +else + log "no max77779-battery supply after the driver loaded (see dmesg)" +fi + +# The address and the power-on are omamobile-bt-addr@'s, started by udev. +if enabled "$bluetooth_node"; then + phase bluetooth + modprobe hci_uart || fail "hci_uart failed" + up+=", Bluetooth" +else + log "no Bluetooth in this DTB; skipped" +fi + +finish +log "up: $up" +exit 0 diff --git a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-clk-acpm.conf b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-clk-acpm.conf new file mode 100644 index 000000000..59c99ab55 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-clk-acpm.conf @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: MIT +# GPU DVFS top level through the ACPM clocks (kernel #84): 890 MHz, the +# highest level the bring-up ran under its thermal throttle; 940 MHz is +# still untested. +options clk_acpm zumapro_gpu_max_khz=890000 diff --git a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf new file mode 100644 index 000000000..0ad529694 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf @@ -0,0 +1,7 @@ +# SPDX-License-Identifier: MIT +# The panel's brightness, sleep and refresh rate go over DSI commands. +options gs_dsim_cmd allow_tx=1 +# 120 Hz (kernel #88) and the panel really off on DPMS, as the owner's phone +# runs it. omamobile-tegu-graphics-start adds stage=2 (the full driver) and +# fbdev=0 on its modprobe line. +options gs_decon max_refresh=120 screen_off=2 diff --git a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-guards.conf b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-guards.conf new file mode 100644 index 000000000..c1f91e335 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-guards.conf @@ -0,0 +1,16 @@ +# SPDX-License-Identifier: MIT +# Loaded only by the package's hang-safe scripts, in their order, after a +# marker is written: omamobile-tegu-graphics-start (the first six) and, once +# a kernel ships them in tree, the Wi-Fi and NFC starts. Blacklisting only +# stops udev from autoloading them ahead of the guard; "modprobe panthor" +# still works. No "install X /bin/false" lines: those break the guard's own +# modprobe. +blacklist panthor +blacklist gs_decon +blacklist cs40l26_vibra +blacklist leds_lm3644 +blacklist max77779_battery +blacklist hci_uart +blacklist pcie_zumapro +blacklist brcmfmac +blacklist nxp_nci_i2c diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/display-manager.service.d/10-omamobile-tegu.conf b/packages/omamobile-device-tegu/vendor/systemd/system/display-manager.service.d/10-omamobile-tegu.conf new file mode 100644 index 000000000..919c7e7cd --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/display-manager.service.d/10-omamobile-tegu.conf @@ -0,0 +1,7 @@ +# SPDX-License-Identifier: MIT +# The display manager needs the GPU and the display controller, which the +# hang-safe graphics start loads; without this it would come up on no DRM +# device and give up. +[Unit] +Wants=omamobile-tegu-graphics.service +After=omamobile-tegu-graphics.service diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-graphics.service b/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-graphics.service new file mode 120000 index 000000000..f93023ce6 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/multi-user.target.wants/omamobile-tegu-graphics.service @@ -0,0 +1 @@ +../omamobile-tegu-graphics.service \ No newline at end of file diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service new file mode 100644 index 000000000..a8a3ada55 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service @@ -0,0 +1,21 @@ +# SPDX-License-Identifier: MIT +# The GPU, display, haptics, torch, gauge and Bluetooth controller at boot, +# enabled by the vendor multi-user.target.wants link. The drivers are +# blacklisted for autoload (modprobe.d/omamobile-tegu-guards.conf), so this +# hang-safe script is the only thing that loads them; if a start hangs, the +# next boot leaves them off (see the script). Runs before the display +# manager (display-manager.service.d/10-omamobile-tegu.conf). +[Unit] +Description=Bring up the Pixel 9a GPU, display, haptics, torch, gauge and Bluetooth (hang-safe) +After=local-fs.target systemd-modules-load.service +Before=display-manager.service +ConditionPathExists=/sys/bus/platform/devices/1f000000.gpu + +[Service] +Type=oneshot +ExecStart=/usr/bin/omamobile-tegu-graphics-start +RemainAfterExit=yes +TimeoutStartSec=120 + +[Install] +WantedBy=multi-user.target From ad60a3d86aadfbd8273e544754bd4c606fcaf66b Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 05/20] omamobile-device-tegu: the Bluetooth address from the bootloader cheetah's omamobile-bt-addr, started by udev when hci0 appears, with the waits the 9a needed: a controller that has never been powered is powered on once before Set Public Address, the index is awaited after it, and the address is checked against the device tree before the script reports success. --- .../bin/omamobile-bt-addr | 69 +++++++++++++++++++ .../systemd/system/omamobile-bt-addr@.service | 11 +++ .../udev/rules.d/80-omamobile-bt-addr.rules | 4 ++ 3 files changed, 84 insertions(+) create mode 100755 packages/omamobile-device-tegu/bin/omamobile-bt-addr create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/omamobile-bt-addr@.service create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/80-omamobile-bt-addr.rules diff --git a/packages/omamobile-device-tegu/bin/omamobile-bt-addr b/packages/omamobile-device-tegu/bin/omamobile-bt-addr new file mode 100755 index 000000000..e52d00f60 --- /dev/null +++ b/packages/omamobile-device-tegu/bin/omamobile-bt-addr @@ -0,0 +1,69 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Give the Bluetooth controller the phone's real address, as Android does. +# The bootloader writes it into the device tree at every boot +# (/chosen/config/bt_addr); the BCM4383 otherwise comes up with a +# placeholder (43:89:00:00:..). +# omamobile-bt-addr hciN (started by udev through omamobile-bt-addr@.service) +set -euo pipefail + +hci=${1:?Usage: omamobile-bt-addr hciN} +index=${hci#hci} +[[ $index =~ ^[0-9]+$ ]] || { echo "not a controller: $hci" >&2; exit 1; } +source=/proc/device-tree/chosen/config/bt_addr +[[ -s $source ]] || exit 0 +addr=$(tr -d '\0' < "$source" | tr 'a-f' 'A-F') + +# A unicast, globally administered address only, like the Wi-Fi one. +[[ $addr =~ ^[0-9A-F]{2}(:[0-9A-F]{2}){5}$ ]] || exit 0 +(( (16#${addr:0:2} & 3) == 0 )) || exit 0 + +# udev starts us when the controller appears, before hci_bcm has finished +# downloading its firmware (about 7 s); it answers info only after that. +for _ in {1..60}; do + info=$(btmgmt --index "$index" info 2> /dev/null || true) + [[ $info == *"current settings:"* ]] && break + sleep 0.5 +done +[[ $info == *"current settings:"* ]] || { echo "$hci never finished its setup" >&2; exit 1; } + +# Set Public Address needs the controller powered off. Set on a controller +# that has never been powered, though, the kernel withdraws the index to +# reconfigure it and it never comes back (seen at every boot). So power it +# on once first, then cycle it around the change, and leave it on. +powered=$(grep -c 'current settings:.*powered' <<< "$info" || true) +if ((!powered)); then + btmgmt --index "$index" power on > /dev/null + info=$(btmgmt --index "$index" info 2> /dev/null || true) + grep -q 'current settings:.*powered' <<< "$info" || { + echo "$hci did not power on before address configuration" >&2 + exit 1 + } +fi +btmgmt --index "$index" power off > /dev/null +btmgmt --index "$index" public-addr "$addr" > /dev/null + +# Set Public Address acknowledges before HCI_CONFIG work finishes. The +# management index is absent until that work has added it again. btmgmt +# can exit zero even for Invalid Index, so wait for a real Read Info reply. +for _ in {1..100}; do + info=$(btmgmt --index "$index" info 2> /dev/null || true) + [[ $info == *"current settings:"* ]] && break + sleep 0.1 +done +[[ $info == *"current settings:"* ]] || { + echo "$hci did not return after address configuration" >&2 + exit 1 +} + +btmgmt --index "$index" power on > /dev/null +info=$(btmgmt --index "$index" info 2> /dev/null || true) +grep -q 'current settings:.*powered' <<< "$info" || { + echo "$hci did not power on after address configuration" >&2 + exit 1 +} +[[ ${info^^} == *"ADDR $addr "* ]] || { + echo "$hci address does not match the bootloader" >&2 + exit 1 +} +exit 0 diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-bt-addr@.service b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-bt-addr@.service new file mode 100644 index 000000000..e6d7b652d --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-bt-addr@.service @@ -0,0 +1,11 @@ +# SPDX-License-Identifier: MIT +# Started by 80-omamobile-bt-addr.rules for the BCM4383 Bluetooth controller. +# A service rather than a udev RUN, because udev's sandbox has no Bluetooth +# sockets. +[Unit] +Description=Set the Bluetooth address of %i from the bootloader +Before=bluetooth.service + +[Service] +Type=oneshot +ExecStart=/usr/bin/omamobile-bt-addr %i diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/80-omamobile-bt-addr.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/80-omamobile-bt-addr.rules new file mode 100644 index 000000000..e53bd1bbb --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/80-omamobile-bt-addr.rules @@ -0,0 +1,4 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: give the BCM4383 Bluetooth controller its real address +# (omamobile-bt-addr@.service). +ACTION=="add", SUBSYSTEM=="bluetooth", KERNEL=="hci[0-9]*", DRIVERS=="hci_uart_bcm", TAG+="systemd", ENV{SYSTEMD_WANTS}+="omamobile-bt-addr@%k.service" From 47053b35efce02776f2cafe364703e96b82d2d6e Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 06/20] omamobile-device-tegu: a charge limit from /etc/omamobile/charge-limit The bring-up script wrote the owner's charge limit into the charger by hand at every start. A udev rule now runs omamobile-tegu-charge-limit when the MAX77779 charger appears: a percentage from 1 to 100 goes to charge_control_end_threshold, anything else is refused and leaves the charger charging to 100 %. --- .../bin/omamobile-tegu-charge-limit | 26 +++++++++++++++++++ .../rules.d/75-omamobile-tegu-charger.rules | 4 +++ 2 files changed, 30 insertions(+) create mode 100755 packages/omamobile-device-tegu/bin/omamobile-tegu-charge-limit create mode 100644 packages/omamobile-device-tegu/vendor/udev/rules.d/75-omamobile-tegu-charger.rules diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-charge-limit b/packages/omamobile-device-tegu/bin/omamobile-tegu-charge-limit new file mode 100755 index 000000000..ed95353a3 --- /dev/null +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-charge-limit @@ -0,0 +1,26 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Pixel 9a charge limit: /etc/omamobile/charge-limit holds the percentage at +# which charging stops (say 80); the charger starts again 5 below it. No +# file, or not a number from 1 to 100: charge to 100 %. Run by udev when the +# MAX77779 charger appears (75-omamobile-tegu-charger.rules), and by hand +# after editing the file. +set -uo pipefail + +root=${OMAMOBILE_TEGU_ROOT:-} +if [[ -n $root ]]; then + [[ $root == /* && $root != / && -d $root ]] || exit 2 +fi +file=$root/etc/omamobile/charge-limit +threshold=$root/sys/class/power_supply/max77779-charger/charge_control_end_threshold +log() { echo "omamobile-tegu-charge-limit: $*"; } + +[[ -s $file ]] || exit 0 +[[ -w $threshold ]] || { log "no charger threshold to set"; exit 0; } +limit=$(tr -d ' \t\n' < "$file") +if [[ ! $limit =~ ^[0-9]+$ ]] || (( 10#$limit < 1 || 10#$limit > 100 )); then + log "$file is not a percentage from 1 to 100; no limit" + exit 1 +fi +echo "$((10#$limit))" > "$threshold" || { log "the charger refused $limit"; exit 1; } +log "charge limit $(cat "$threshold") %" diff --git a/packages/omamobile-device-tegu/vendor/udev/rules.d/75-omamobile-tegu-charger.rules b/packages/omamobile-device-tegu/vendor/udev/rules.d/75-omamobile-tegu-charger.rules new file mode 100644 index 000000000..ed60bab49 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/udev/rules.d/75-omamobile-tegu-charger.rules @@ -0,0 +1,4 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: apply /etc/omamobile/charge-limit when the MAX77779 charger +# appears (omamobile-tegu-charge-limit). +ACTION=="add", SUBSYSTEM=="power_supply", KERNEL=="max77779-charger", RUN+="/usr/bin/omamobile-tegu-charge-limit" From 8c4deacce56c556e5b39784b947df8e05ca64a39 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 07/20] omamobile-device-tegu: sleep hooks for Wi-Fi and NFC, s2idle, the reboot watchdog, module cleanup held off The Wi-Fi hook keeps brcmfmac loaded across s2idle when the owner's flag says so and unloads it around sleeps otherwise, or after a sleep that lost the chip; it now loads the in-tree module instead of a file in /var/lib. The NFC hook stops neard's poll loop and powers the ST54L off before a sleep, since it does not answer on I2C after a wake. Suspend means s2idle and a hung shutdown resets the phone after a minute, as on cheetah; Arch's module cleanup is held off while the package exists, because our kernels come as tarballs. --- .../systemd/sleep.conf.d/10-omamobile.conf | 11 +++ .../systemd/system-sleep/omamobile-tegu-nfc | 48 +++++++++++++ .../systemd/system-sleep/omamobile-tegu-wifi | 70 +++++++++++++++++++ .../10-omamobile-reboot-watchdog.conf | 7 ++ .../10-omamobile-tegu.conf | 6 ++ 5 files changed, 142 insertions(+) create mode 100644 packages/omamobile-device-tegu/vendor/systemd/sleep.conf.d/10-omamobile.conf create mode 100755 packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc create mode 100755 packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-wifi create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf diff --git a/packages/omamobile-device-tegu/vendor/systemd/sleep.conf.d/10-omamobile.conf b/packages/omamobile-device-tegu/vendor/systemd/sleep.conf.d/10-omamobile.conf new file mode 100644 index 000000000..d5f37e34c --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/sleep.conf.d/10-omamobile.conf @@ -0,0 +1,11 @@ +# SPDX-License-Identifier: MIT +# The firmware advertises PSCI SYSTEM_SUSPEND, so the kernel offers "deep" +# sleep, but Linux has no gs201 SoC sleep sequence yet: entering it without +# one doesn't wake. Suspend means suspend-to-idle until SYS_SLEEP exists +# (device/cheetah/SUSPEND-PLAN.md, sections 1 and 5). There is no swap to +# hibernate to. +[Sleep] +MemorySleepMode=s2idle +AllowHibernation=no +AllowHybridSleep=no +AllowSuspendThenHibernate=no diff --git a/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc new file mode 100755 index 000000000..f05ac8b83 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc @@ -0,0 +1,48 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# systemd-sleep hook for the Pixel 9a's NFC (ST54L on nxp-nci): after an +# s2idle wake the controller doesn't answer on I2C until the adapter is +# powered off and on (kernel #132). Before the sleep: stop the poll loop, +# power the adapter off and check that Powered reads false (the kernel +# refuses a device-down while polling, EBUSY). If it doesn't, leave it with +# polling stopped. After the wake: power it on, check, and restart the poll +# loop if it was polling. Logs every step. Nothing happens without neard. +# OMAMOBILE_TEGU_ROOT: a scratch root for the package tests. +root=${OMAMOBILE_TEGU_ROOT:-} +A=/org/neard/nfc0 +D=$root/run/omamobile +L=$root/var/log/omamobile/nfc-sleep.log +mkdir -p "$D" "${L%/*}" +log() { echo "$(date -Is) $1: $2" >> "$L"; } +prop() { busctl get-property org.neard $A org.neard.Adapter "$1" 2>/dev/null; } +is_true() { prop "$1" | grep -c true >/dev/null; } +case "$1" in + pre) + rm -f "$D/nfc-was-polling" "$D/nfc-was-powered" + is_true Polling && : > "$D/nfc-was-polling" + if is_true Powered; then + : > "$D/nfc-was-powered" + out=$(busctl call org.neard $A org.neard.Adapter StopPollLoop 2>&1) || log pre "StopPollLoop: $out" + for i in 1 2 3 4 5; do is_true Polling || break; sleep 0.2; done + out=$(busctl set-property org.neard $A org.neard.Adapter Powered b false 2>&1) || log pre "Powered false: $out" + for i in 1 2 3 4 5; do is_true Powered || break; sleep 0.2; done + log pre "polling=$(prop Polling) powered=$(prop Powered)" + sync + else + log pre "adapter already off" + fi + ;; + post) + if [[ -e $D/nfc-was-powered ]]; then + rm -f "$D/nfc-was-powered" + out=$(busctl set-property org.neard $A org.neard.Adapter Powered b true 2>&1) || log post "Powered true: $out" + for i in 1 2 3 4 5 6 7 8 9 10; do is_true Powered && break; sleep 0.3; done + fi + if [[ -e $D/nfc-was-polling ]]; then + rm -f "$D/nfc-was-polling" + out=$(busctl call org.neard $A org.neard.Adapter StartPollLoop s Initiator 2>&1) || log post "StartPollLoop: $out" + fi + log post "polling=$(prop Polling) powered=$(prop Powered)" + ;; +esac +exit 0 diff --git a/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-wifi b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-wifi new file mode 100755 index 000000000..d1d8bf663 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-wifi @@ -0,0 +1,70 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# systemd-sleep hook for the Pixel 9a's BCM4383 (device/tegu/WIFI-PLAN.md). +# +# Wi-Fi survives s2idle loaded: brcmfmac sends D3, the chip loses its +# firmware in sleep, and on resume brcmfmac finds that and probes it again +# ("device lost its state in sleep, probing again"); iwd reconnects within +# seconds. So when /etc/omamobile/wifi-keep-across-sleep exists the hook +# leaves Wi-Fi alone and only reloads it afterwards if it didn't come back. +# Without the flag, or after a sleep that lost Wi-Fi, it unloads brcmfmac +# before the sleep and reloads it afterwards. Nothing happens on a boot +# where Wi-Fi was never loaded. +# OMAMOBILE_TEGU_ROOT: a scratch root for the package tests. +root=${OMAMOBILE_TEGU_ROOT:-} +KEEP=$root/etc/omamobile/wifi-keep-across-sleep +RUNDIR=$root/run/omamobile +FALLBACK=$RUNDIR/wifi-unload-across-sleep +UNLOADED=$RUNDIR/wifi-unloaded-for-sleep + +loaded() { lsmod | grep -c '^brcmfmac ' >/dev/null; } + +wifi_up() { + local i + for i in $(seq 1 20); do + ls "$root"/sys/bus/pci/drivers/brcmfmac/ 2>/dev/null | grep -c '^0000:' >/dev/null && + ls "$root"/sys/class/ieee80211/ 2>/dev/null | grep -c . >/dev/null && return 0 + sleep 1 + done + return 1 +} + +unload() { + systemctl stop iwd + timeout 20 rmmod brcmfmac_wcc brcmfmac 2>/dev/null || true +} + +reload() { + loaded || modprobe brcmfmac + systemctl start iwd +} + +mkdir -p "$RUNDIR" +case "$1" in + pre) + rm -f "$UNLOADED" + loaded || exit 0 + if [[ -e $KEEP && ! -e $FALLBACK ]]; then + exit 0 + fi + unload + : > "$UNLOADED" + ;; + post) + if [[ -e $UNLOADED ]]; then + rm -f "$UNLOADED" + reload + exit 0 + fi + loaded || exit 0 + if wifi_up; then + exit 0 + fi + # Wi-Fi didn't come back: unload around sleeps for the rest of this boot. + : > "$FALLBACK" + logger -t omamobile-tegu-wifi "Wi-Fi lost across sleep; unloading it around sleeps until reboot" + unload + reload + ;; +esac +exit 0 diff --git a/packages/omamobile-device-tegu/vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf b/packages/omamobile-device-tegu/vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf new file mode 100644 index 000000000..9764faba8 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf @@ -0,0 +1,7 @@ +# SPDX-License-Identifier: MIT +# Omarchy Mobile: if shutdown hangs (a driver stuck in remove(), say), let the +# hardware watchdog reset the phone after a minute instead of leaving it +# on the shutdown splash until someone holds Power. systemd's default is +# 10 minutes. Nothing holds the watchdog while the system runs. +[Manager] +RebootWatchdogSec=1min diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf b/packages/omamobile-device-tegu/vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf new file mode 100644 index 000000000..cc48a8901 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: MIT +[Unit] +# Our slot-B kernels come as modules tarballs (bin/omamobile push), not as +# pacman-owned kernel packages. Arch's cleanup would remove the running, +# next-boot and rollback module trees; keep them while this package exists. +ConditionPathExists=!/usr/share/omamobile/device-tegu/version From 7074a01e0a8c9f055a7f270d1f9b053a08e23d96 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 08/20] omamobile-device-tegu: the AoC start, shipped disabled cheetah's hang-safe AoC start with the 9a's names: Trusty, the GSA, the AoC drivers with aoc_channel_dev for the sensors' USF channel, aoc-tegu.bin, then a wait for a stable service count. The unit has no wants link: the kernel's device tree does not enable the AoC, GSA and Trusty nodes yet (the bring-up turns them on with out-of-tree overlay modules), so the script would find nothing to do until it does. --- .../bin/omamobile-tegu-aoc-start | 120 ++++++++++++++++++ .../vendor/modprobe.d/omamobile-tegu-aoc.conf | 13 ++ .../systemd/system/omamobile-tegu-aoc.service | 20 +++ 3 files changed, 153 insertions(+) create mode 100755 packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start create mode 100644 packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-aoc.conf create mode 100644 packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-aoc.service diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start new file mode 100755 index 000000000..9bdb54821 --- /dev/null +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start @@ -0,0 +1,120 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# Boot the Pixel 9a's AoC (audio and sensor DSP), hang-safe, in the order +# the bring-up proved (omamobile-tegu-audio-up, docs/HANDOFF.md S5) and the +# Pixel 7 Pro's omamobile-aoc-start uses: +# 1. Trusty (trusty_core, trusty_log, trusty_ipc, then trusty_virtio): +# the AoC is started by Trusty's hwmgr.aoc service. +# 2. The GSA mailbox (gsa), which authenticates the image. +# 3. The AoC System MMU, mailboxes, core and service front ends +# (aoc_channel_dev for the sensors' USF channel and the sound card's +# aoc_alsa_dev before aoc_char_dev, which takes every service still free). +# 4. The firmware load (aoc-tegu.bin), then wait for the 93 services and +# the sensor channel. +# The TAS2572 amplifiers come up from the device tree through the tas2781 +# codec driver; the card's UCM profile (omamobile-audio) does the routing. +# +# This waits for a kernel whose device tree enables the AoC, GSA and Trusty +# nodes: today they are turned on by out-of-tree overlay modules that no +# package can ship, so the unit has no wants link (README.md). +# A marker is written and synced before any of it, and removed at the end. +# If a boot hangs here, the next one finds the marker and leaves the AoC off; +# "omamobile-tegu-aoc-start --retry" tries again. +# omamobile-tegu-aoc-start [--retry] +set -uo pipefail + +root=${OMAMOBILE_TEGU_ROOT:-} +if [[ -n $root ]]; then + [[ $root == /* && $root != / && -d $root ]] || exit 2 +fi +state=$root/var/lib/omamobile +marker=$state/tegu-aoc-started +# Set right before the firmware load, gone after a reboot (tmpfs) +loaded=$root/run/omamobile/tegu-aoc-firmware-loaded +node=$root/sys/bus/platform/devices/154b0000.aoc +trusty=$root/proc/device-tree/trusty/status + +log() { echo "omamobile-tegu-aoc: $*"; } + +case ${1:-} in + '') ;; + --retry) rm -f "$marker" ;; + *) log "usage: omamobile-tegu-aoc-start [--retry]"; exit 2 ;; +esac + +if [[ -e $marker ]]; then + log "the last start ($(cat "$marker" 2>/dev/null)) never finished; the AoC stays off this boot." + log "after checking the previous boot's log, retry with: omamobile-tegu-aoc-start --retry" + exit 0 +fi +if [[ ! -e $node ]]; then + log "the AoC is not enabled in this DTB; nothing to do" + exit 0 +fi +# The AoC starts once per boot: a second firmware load takes the running AoC +# down (the GSA refuses the image while it runs, and there is no AoC reset). +if grep -q '^Services : [1-9]' "$node/services" 2>/dev/null; then + log "already online: $(head -1 "$node/services")" + exit 0 +fi +if [[ -e $loaded ]]; then + log "the AoC was already started this boot but is not online; reboot to start it again" + exit 1 +fi +if [[ ! -e $trusty ]] || [[ $(tr -d '\0' < "$trusty") != okay ]]; then + log "Trusty is not enabled in this DTB; the AoC cannot be started" + exit 1 +fi +fw= +for d in "$root/usr/lib/firmware/updates" "$root/usr/lib/firmware"; do + [[ -e $d/aoc-tegu.bin ]] && fw=$d/aoc-tegu.bin && break +done +if [[ -z $fw ]]; then + log "no aoc-tegu.bin (install it from the phone's vendor partition: tools/extract-vendor-firmware)" + exit 1 +fi + +mkdir -p "$state" +date -Is > "$marker" +sync -f "$marker" 2>/dev/null || sync + +finish() { rm -f "$marker"; sync -f "$state" 2>/dev/null || sync; } +fail() { log "$*"; finish; exit 1; } + +# 1. Trusty +for m in trusty_core trusty_log trusty_ipc trusty_virtio; do + modprobe "$m" || fail "$m failed" +done +for _ in $(seq 50); do [[ -e $root/dev/trusty-ipc-dev0 ]] && break; sleep 0.2; done +[[ -e $root/dev/trusty-ipc-dev0 ]] || fail "no /dev/trusty-ipc-dev0" + +# 2. GSA +modprobe gsa || fail "gsa failed" + +# 3. AoC drivers +for m in gs201_aoc_sysmmu mailbox_wc aoc_core aoc_channel_dev aoc_alsa_dev aoc_char_dev; do + modprobe "$m" || fail "$m failed" +done +[[ -w $node/firmware ]] || fail "the AoC core did not bind ($node/firmware missing)" + +# 4. Firmware; the core loads, authenticates and starts it in the background. +# The bring-up saw the service count flap before it settled, so wait for it +# to hold for 3 s. +mkdir -p "${loaded%/*}" && date -Is > "$loaded" +echo aoc-tegu.bin > "$node/firmware" || fail "firmware load request failed" +stable=0 +for _ in $(seq 150); do + if grep -q '^Services : [1-9]' "$node/services" 2>/dev/null; then + stable=$((stable + 1)) + else + stable=0 + fi + if ((stable >= 15)); then + log "up: $(head -1 "$node/services"), firmware $(head -1 "$node/revision" 2>/dev/null)" + for _ in $(seq 25); do [[ -e $root/dev/acd-com.google.usf ]] && break; sleep 0.2; done + [[ -e $root/dev/acd-com.google.usf ]] || log "no /dev/acd-com.google.usf (sensors unavailable)" + finish; exit 0 + fi + sleep 0.2 +done +fail "no stable AoC services after 30 s (see dmesg and /dev/acd-logging)" diff --git a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-aoc.conf b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-aoc.conf new file mode 100644 index 000000000..ef470fc35 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-aoc.conf @@ -0,0 +1,13 @@ +# SPDX-License-Identifier: MIT +# AoC: only omamobile-tegu-aoc-start (or a bring-up runbook) loads these, in +# order. Explicit modprobe still works. +blacklist gsa +blacklist gs201_aoc_sysmmu +blacklist mailbox_wc +blacklist aoc_core +blacklist aoc_channel_dev +blacklist aoc_char_dev +# The sound card's driver would load at boot and pull aoc_core, gsa and Trusty +# in ahead of the order above; the start script loads it with the others. +blacklist aoc_alsa_dev +blacklist aoc_alsa_dev_util diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-aoc.service b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-aoc.service new file mode 100644 index 000000000..87966fd84 --- /dev/null +++ b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-aoc.service @@ -0,0 +1,20 @@ +# SPDX-License-Identifier: MIT +# The AoC (audio and sensor DSP) at boot, through the hang-safe +# omamobile-tegu-aoc-start. Not enabled by the package: the Pixel 9a kernel +# does not yet enable the AoC, GSA and Trusty nodes in its device tree (they +# are turned on by out-of-tree overlay modules in the bring-up), so the +# script would find nothing to do. Once a kernel enables them: +# systemctl enable omamobile-tegu-aoc.service +# The AoC modules are blacklisted for autoload (modprobe.d/omamobile-tegu-aoc.conf). +[Unit] +Description=Boot the Pixel 9a AoC audio and sensor DSP (hang-safe) +After=local-fs.target systemd-modules-load.service +ConditionPathExists=/sys/bus/platform/devices/154b0000.aoc + +[Service] +Type=oneshot +ExecStart=/usr/bin/omamobile-tegu-aoc-start +RemainAfterExit=yes + +[Install] +WantedBy=multi-user.target From dcad2309e30703f87bc0bc5d09863a6c4a9f7c81 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:15:04 -0647 Subject: [PATCH 09/20] omamobile-device-tegu 0.2.0: tests in a scratch root, the README, the version test/all checks what is staged, that only the two guards are enabled, that every guarded driver is blacklisted without install lines, and that no blob or address is in the package; the scripts run in a scratch root with fake modprobe, sync, busctl and friends (test/graphics-start, aoc-start, usb-rescue, sleep-hooks, charge-limit), no phone needed. The README says what the kernel command line must stop carrying for the package to take over, why the AoC start ships disabled, which firmware the phone needs, and what to take off the first 9a when moving from the bring-up. --- packages/omamobile-device-tegu/README.md | 153 +++++++++++++++--- packages/omamobile-device-tegu/test/all | 105 ++++++++++-- packages/omamobile-device-tegu/test/aoc-start | 83 ++++++++++ .../omamobile-device-tegu/test/charge-limit | 32 ++++ .../omamobile-device-tegu/test/graphics-start | 111 +++++++++++++ .../omamobile-device-tegu/test/sleep-hooks | 90 +++++++++++ .../omamobile-device-tegu/test/usb-rescue | 45 ++++++ packages/omamobile-device-tegu/version | 2 +- 8 files changed, 583 insertions(+), 38 deletions(-) create mode 100755 packages/omamobile-device-tegu/test/aoc-start create mode 100755 packages/omamobile-device-tegu/test/charge-limit create mode 100755 packages/omamobile-device-tegu/test/graphics-start create mode 100755 packages/omamobile-device-tegu/test/sleep-hooks create mode 100755 packages/omamobile-device-tegu/test/usb-rescue diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index 770a586d6..914530a90 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -1,26 +1,129 @@ -# Pixel 9a device policy - -Experimental tegu support: the unchanged storage guard's udev policy, USB -networking and the device descriptor. The package does not start firmware, -radio, audio, charging, fingerprint or other unqualified hardware services. -It contains no vendor files or per-phone data. - -The CS40L26 feedbackd rule is reused unchanged from the Pixel7 Pro package. -It identifies an already registered force-feedback input and grants desktop -access; it does not load the driver, power hardware or send a vibration. - -The LM3644 torch rule grants the existing video group write access only to -`white:torch/brightness` owned by that driver, for the normal quick-settings -`brightnessctl` path. It neither loads a driver nor turns on the light. The -kernel board current cap remains authoritative. On the local d816 kernel, -the owner has confirmed rear-light on/off through the existing tile at the -single tested minimum step; brightness tuning is still separate. - -The qualified storage path and remaining standalone-boot limitation are in -`device/tegu/UFS-PLAN.md` and `BOOT-PLAN.md`. The temporary recovery module -is a bring-up tool, not production reboot support. `reboot-to` by itself -does not provide the tested guarded bootloader route on this kernel. - -Build with `makepkg`; `test/all` checks the staged storage and USB policy. -The package is self-contained and provides/conflicts with `omamobile-device`. +# omamobile-device-tegu + +Omarchy Mobile's system support for the Google Pixel 9a (codename tegu, +Tensor G4 / zumapro). It sits beside Omarchy's own packages the way +`omarchy-mac` adds Apple Silicon support, and does for the 9a what +`omamobile-device-cheetah` does for the Pixel 7 Pro, with one difference: +the 9a's kernel is still being upstreamed, so part of its bring-up cannot be +packaged yet. What is here is what runs on in-tree drivers. + +| Path | Installs to | What it does | +| --- | --- | --- | +| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` tries again. The display manager waits for it | +| `vendor/modprobe.d/omamobile-tegu-guards.conf` | `/usr/lib/modprobe.d` | Keeps those drivers (and the Wi-Fi, PCIe and NFC ones, for the day they are in tree) from autoloading ahead of the guard. No `install` lines: they would break the guard's own modprobe | +| `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | DSI commands for brightness and sleep, 120 Hz, the panel really off on DPMS; the GPU DVFS ceiling at 890 MHz | +| `bin/omamobile-tegu-aoc-start` + `vendor/systemd/system/omamobile-tegu-aoc.service` + `vendor/modprobe.d/omamobile-tegu-aoc.conf` | `/usr/bin`, `/usr/lib/systemd/system` (not enabled), `/usr/lib/modprobe.d` | Boots the AoC (audio and sensor DSP) hang-safe: Trusty, the GSA, the AoC drivers, `aoc-tegu.bin`, then waits for its 93 services. Ships disabled: see below | +| `bin/omamobile-bt-addr` + `vendor/systemd/system/omamobile-bt-addr@.service` + `vendor/udev/rules.d/80-omamobile-bt-addr.rules` | `/usr/bin`, `/usr/lib/systemd/system`, `/usr/lib/udev/rules.d` | Gives the BCM4383 its real Bluetooth address from the bootloader's device tree when the controller appears, powering it on once first (set on a never-powered controller, the kernel withdraws the index for good). The Pixel 7 Pro's script, with the extra waits the 9a needed | +| `bin/omamobile-tegu-usb-rescue` + `src/tegu-tcpc-switch.c` + `vendor/systemd/system/omamobile-tegu-usb-rescue.service` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Closes the MAX77779's USB data switch after a cable replug while no Type-C driver owns the chip; with one loaded it exits. The helper reads the chip's identity and VBUS before its one write and never forces an address a driver holds | +| `bin/omamobile-tegu-charge-limit` + `vendor/udev/rules.d/75-omamobile-tegu-charger.rules` | `/usr/bin`, `/usr/lib/udev/rules.d` | Applies `/etc/omamobile/charge-limit` (a percentage, say `80`) to the charger when it appears; charging stops there and starts again 5 below. No file: charge to 100 % | +| `vendor/systemd/system-sleep/omamobile-tegu-wifi` | `/usr/lib/systemd/system-sleep` | Keeps Wi-Fi loaded across s2idle when `/etc/omamobile/wifi-keep-across-sleep` exists (the chip reloads its firmware on resume and iwd reconnects), otherwise unloads and reloads brcmfmac around the sleep; after a sleep that lost Wi-Fi it unloads for the rest of the boot | +| `vendor/systemd/system-sleep/omamobile-tegu-nfc` | `/usr/lib/systemd/system-sleep` | Stops neard's poll loop and powers the ST54L off before a sleep (it does not answer on I2C after a wake otherwise), and brings it back afterwards | +| `vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf` | `/usr/lib/systemd/system` | Holds off Arch's old-module cleanup: our kernels come as module tarballs, not pacman packages, and the next-boot and rollback trees must stay | +| `vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf`, `sleep.conf.d/10-omamobile.conf` | `/usr/lib/systemd` | A hung shutdown resets the phone after a minute; suspend means s2idle (no SoC sleep sequence yet, and `deep` does not wake) | +| `vendor/udev/rules.d/10-omamobile-storage.rules` | `/usr/lib/udev/rules.d` | Continues the initramfs storage guard: Android partitions stay hidden and the whole-disk node unusable | +| `vendor/udev/rules.d/20-omamobile-torch.rules`, `70-omamobile-dma-heap.rules`, `72-omamobile-feedbackd.rules` | `/usr/lib/udev/rules.d` | The torch tile may set the LED's brightness; libcamera's software ISP gets the DMA-BUF heaps; feedbackd gets the haptics | +| `vendor/systemd/network/20-usb-gadget.network`, `vendor/NetworkManager/conf.d/10-omamobile-usb.conf` | `/usr/lib/systemd/network`, `/usr/lib/NetworkManager/conf.d` | USB networking to a host (172.16.42.1, the host as a fallback route behind Wi-Fi); NetworkManager stays off the link | +| `src/reboot-to.c` | `/usr/bin/reboot-to` | Reboot with a reason the bootloader understands. One source with the cheetah package | +| `device.conf`, `etc/omamobile-mirrorlist` | `/usr/share/omamobile`, `/etc/pacman.d` | The device descriptor (a copy of `device/tegu/device.conf`) and the `[omamobile]` repository | + +`./build [DIR]` compiles the C tools, `./install DESTDIR [DIR]` stages +everything and enables nothing beyond the two wants links, and `test/all` +runs the package tests in a scratch root (`OMAMOBILE_TEGU_ROOT`), with no +phone and no hardware. + +The package does not mark the boot slot successful (Derek decides how that +works on the 9a), does not start Wi-Fi, NFC, the modem, the charger driver +or the cameras, and carries no vendor files and no per-phone data. + +## What the kernel command line has to stop carrying + +The bring-up boots with a command line that masks units and gates its own +start scripts. The package's units cannot run under it. For the package to +take over, `/chosen/bootargs` has to lose: + +- `systemd.unit=multi-user.target` (graphical.target is the default again); +- every `systemd.mask=` for `sddm`, `display-manager`, `NetworkManager`, + `bluetooth`, the `omamobile-*` units and the `tegu-*` units; +- `systemd.wants=tegu-curated-root-guard.service`, + `systemd.wants=tegu-curated-autostart.service` and + `systemd.wants=bluetooth.service`; +- the flags `omamobile.tegu-curated-root`, `omamobile.tegu-curated-autostart`, + `omamobile.tegu-standard-reboot` and `omamobile.tegu-native-bluetooth`; +- `modprobe.blacklist=gs_decon,panthor,cs40l26_vibra,hci_uart,leds_lm3644` + (`omamobile-tegu-guards.conf` does this now). + +`mem_sleep_default=s2idle`, `panic=10`, `s3c2410_wdt.tmr_atboot=1` and +`cs40l26_vibra.max_buzz_ms=100` stay. Whether and when that DTB change +happens is the owner's call. + +## The AoC start ships disabled + +`omamobile-tegu-aoc-start` is the Pixel 7 Pro's AoC start with the 9a's +names (`154b0000.aoc`, `aoc-tegu.bin`, 93 services, the sensors' USF +channel through `aoc_channel_dev`). It needs a kernel whose device tree +enables the AoC, GSA and Trusty nodes; today the bring-up turns them on +with out-of-tree overlay modules, and so `omamobile-tegu-aoc.service` has +no wants link. Once a kernel enables them, `systemctl enable +omamobile-tegu-aoc.service` and reboot; a failed start names the step in +the journal and the marker, and the next boot leaves the AoC off until +`omamobile-tegu-aoc-start --retry`. + +## Firmware the phone needs + +Nothing here is in the package; `tools/extract-vendor-firmware` and +`tools/phone-provision` take it from the owner's own factory image: + +- `aoc-tegu.bin` (the AoC), `Top-tas2572RCA1.bin` and + `Bottom-tas2572RCA1.bin` (the speaker amplifiers); +- `brcm/brcmfmac4383-pcie.bin`, `.clm_blob`, `.txcap_blob` and `.txt` (Wi-Fi); +- `updates/brcm/BCM.google,zumapro-tegu.hcd` (Bluetooth); +- `updates/cs40l26.wmfw` and `cs40l26.bin` (haptics); +- `arm/mali/arch11.8/mali_csffw.bin` (the GPU's firmware, from Mesa's + packaging or the factory image). + +## Moving from the bring-up + +The bring-up on the first 9a lives in `/usr/local` and `/etc`, outside any +package. After installing 0.2.0 and before the reboot that drops the +command line above, take these off the phone by hand; the package cannot, +and a leftover copy in `/etc` wins over the package's file of the same name. + +- Units and drop-ins in `/etc/systemd/system`: `tegu-curated-autostart`, + `tegu-curated-root-guard`, `tegu-curated-session`, + `tegu-curated-bluetooth-prepare`, `bluetooth.service.d/20-tegu-curated.conf`, + `omamobile-tegu-audio` (keep it only while the AoC still needs the + out-of-tree start; it conflicts with nothing in the package), + `tegu-dtb-deadman`, `tegu-rt9470-early-read`, and the retired + `tegu-battery`, `tegu-battery-hold`, `tegu-display`, `tegu-display-confirm`, + `tegu-gpu`, `tegu-touch` (with `tegu-touch.service.d`), `tegu-normal-reboot`, + `tegu-native-root-verify`, `tegu-native-root-deadman`, with their links + in `multi-user.target.wants`, `graphical.target.wants` and + `sysinit.target.wants`. `systemctl disable` each first. +- `/etc/modprobe.d/omamobile-tegu-*.conf` and + `tegu-battery-gauge-bringup.conf` (the package's `omamobile-tegu-guards.conf`, + `-display.conf` and `-clk-acpm.conf` replace them; + `omamobile-tegu-fp-lhbm.conf` belongs to omamobile-fingerprint). +- `/etc/udev/rules.d/20-omamobile-tegu-gpu-always-on.rules` (the GPU's + runtime suspend is qualified), `20-omamobile-tegu-torch-bringup.rules` and + `70-omamobile-dma-heap.rules` (the package has them). +- `/etc/systemd/system.conf.d/10-tegu-bringup-watchdog.conf` (the package + sets the reboot watchdog; the 10 s runtime watchdog was the bring-up's). +- `/etc/systemd/network/20-usb-gadget.network` (same content as the package's). +- `/usr/lib/systemd/system-sleep/omamobile-tegu-wifi` and `-nfc` (the + package installs its own under the same names: pacman will refuse to + overwrite unowned files, so remove them first). +- `/usr/local/bin/reboot-to`, `g7-*`, `nfc-ab`; `/usr/local/sbin/tegu-*`; + `/usr/local/libexec/tegu-*`; `/usr/local/lib/omamobile-tegu` + (`tegu-tcpc-switch` is `/usr/bin/tegu-tcpc-switch` now) and + `/usr/local/lib/omamobile-tegu-curated`. +- `/var/lib/omamobile-tegu/curated` and the old per-kernel state + directories (`battery`, `display`, `gpu`, `touch`, `keys`, `pmu`, `time`, + `ui`, `wifi*`, `module-backup`); keep `audio`, `nfc` and `modem` while + `omamobile-tegu-audio-up` and `omamobile-tegu-modem-up` still load from them. + +`omamobile-tegu-audio-up`, `omamobile-tegu-modem-up` and the out-of-tree +modules they load stay until the kernel carries their drivers and device +tree nodes; the package's guards and theirs do not load the same modules. + +Licence: MIT. diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all index 885263848..f91059cc2 100755 --- a/packages/omamobile-device-tegu/test/all +++ b/packages/omamobile-device-tegu/test/all @@ -1,7 +1,9 @@ #!/bin/bash # SPDX-License-Identifier: MIT -# Tests for omamobile-device-tegu. reboot-to is compiled from -# src/reboot-to.c in this directory, so a copied package still builds. +# Tests for omamobile-device-tegu: the package stages what it says, enables +# only the hang-safe guards, and the scripts behave in a scratch root +# (test/graphics-start and friends, no phone, no hardware). reboot-to and +# tegu-tcpc-switch are compiled from src/, so a copied package still builds. # Do not execute reboot-to. set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." @@ -9,8 +11,9 @@ fail() { printf 'not ok - %s\n' "$1" >&2; exit 1; } pass() { printf 'ok - %s\n' "$1"; } tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT -bash -n build || fail "build parses" -bash -n install || fail "install parses" +for script in build install bin/* vendor/systemd/system-sleep/* test/*; do + bash -n "$script" || fail "$script parses" +done pass "scripts parse" grep -q '\.\./\.\.' build install && fail "build or install reaches outside the package directory" @@ -31,26 +34,54 @@ if [[ -f ../omamobile-device-cheetah/src/reboot-to.c ]]; then ../omamobile-device-cheetah/omamobile-device-cheetah.install \ || fail "the pacman install script drifted from the cheetah package" pass "the pacman install script matches the cheetah package" + cmp -s vendor/systemd/sleep.conf.d/10-omamobile.conf \ + ../omamobile-device-cheetah/vendor/systemd/sleep.conf.d/10-omamobile.conf \ + || fail "the sleep drop-in drifted from the cheetah package" + pass "the sleep drop-in matches the cheetah package" fi grep -q 'omamobile-keyring' PKGBUILD || fail "pacman needs omamobile-keyring" grep -q '^install=omamobile-device-tegu.install$' PKGBUILD \ || fail "the pacman install script is not named" +for dep in kmod i2c-tools bluez-utils; do + grep -qE "^depends=.*\b$dep\b" PKGBUILD || fail "the scripts need $dep (depends)" +done [[ -f etc/omamobile-mirrorlist && -f omamobile-device-tegu.install ]] \ || fail "the pacman mirrorlist or install script is missing" -pass "pacman gets the keyring, the mirrorlist, and the install script" +pass "pacman gets the keyring, the mirrorlist, the install script and the tools' dependencies" if [[ -f ../../device/tegu/device.conf ]]; then cmp -s device.conf ../../device/tegu/device.conf \ || fail "device.conf drifted from the descriptor" pass "device.conf matches the descriptor" fi -./build "$tmp/build" || fail "reboot-to compiles" -[[ -x $tmp/build/reboot-to ]] || fail "reboot-to is not executable" -pass "reboot-to compiles" +# Nothing that must come from the user's own phone or a kernel build +blobs=$(find . -type f \( -name '*.ko' -o -name '*.bin' -o -name '*.hcd' -o -name '*.img' -o -name '*.wmfw' \)) +[[ -z $blobs ]] || fail "binaries in the package: $blobs" +macs=$(grep -rEn '\b([0-9a-fA-F]{2}:){5}[0-9a-fA-F]{2}\b' --exclude-dir=build . || true) +[[ -z $macs ]] || fail "a MAC or Bluetooth address in the package: $macs" +pass "no blobs, no addresses" + +./build "$tmp/build" || fail "the C tools compile" +[[ -x $tmp/build/reboot-to && -x $tmp/build/tegu-tcpc-switch ]] || fail "the C tools are not executable" +pass "the C tools compile" ./install "$tmp/root" "$tmp/build" -for f in usr/bin/reboot-to \ - usr/lib/udev/rules.d/10-omamobile-storage.rules \ +for f in usr/bin/reboot-to usr/bin/tegu-tcpc-switch \ + usr/bin/omamobile-tegu-graphics-start usr/lib/systemd/system/omamobile-tegu-graphics.service \ + usr/lib/systemd/system/display-manager.service.d/10-omamobile-tegu.conf \ + usr/bin/omamobile-tegu-aoc-start usr/lib/systemd/system/omamobile-tegu-aoc.service \ + usr/bin/omamobile-tegu-usb-rescue usr/lib/systemd/system/omamobile-tegu-usb-rescue.service \ + usr/bin/omamobile-bt-addr usr/lib/systemd/system/omamobile-bt-addr@.service \ + usr/lib/udev/rules.d/80-omamobile-bt-addr.rules \ + usr/bin/omamobile-tegu-charge-limit usr/lib/udev/rules.d/75-omamobile-tegu-charger.rules \ + usr/lib/modprobe.d/omamobile-tegu-guards.conf usr/lib/modprobe.d/omamobile-tegu-aoc.conf \ + usr/lib/modprobe.d/omamobile-tegu-display.conf usr/lib/modprobe.d/omamobile-tegu-clk-acpm.conf \ + usr/lib/systemd/system-sleep/omamobile-tegu-wifi usr/lib/systemd/system-sleep/omamobile-tegu-nfc \ + usr/lib/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf \ + usr/lib/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf \ + usr/lib/systemd/sleep.conf.d/10-omamobile.conf \ + usr/lib/udev/rules.d/10-omamobile-storage.rules usr/lib/udev/rules.d/20-omamobile-torch.rules \ + usr/lib/udev/rules.d/70-omamobile-dma-heap.rules usr/lib/udev/rules.d/72-omamobile-feedbackd.rules \ usr/lib/systemd/network/20-usb-gadget.network \ usr/lib/NetworkManager/conf.d/10-omamobile-usb.conf \ usr/share/omamobile/device.conf \ @@ -59,18 +90,68 @@ for f in usr/bin/reboot-to \ etc/pacman.d/omamobile-mirrorlist; do [[ -f $tmp/root/$f ]] || fail "install stages $f" done +[[ -x $tmp/root/usr/lib/systemd/system-sleep/omamobile-tegu-wifi && -x $tmp/root/usr/lib/systemd/system-sleep/omamobile-tegu-nfc ]] \ + || fail "the sleep hooks are not executable" +for unit in omamobile-tegu-graphics omamobile-tegu-usb-rescue; do + [[ $(readlink "$tmp/root/usr/lib/systemd/system/multi-user.target.wants/$unit.service") == ../$unit.service ]] \ + || fail "$unit is enabled by a vendor wants link" +done +[[ ! -e $tmp/root/usr/lib/systemd/system/multi-user.target.wants/omamobile-tegu-aoc.service ]] \ + || fail "the AoC start ships disabled (no wants link)" grep -q 'KERNEL=="sda", MODE="0000"' "$tmp/root/usr/lib/udev/rules.d/10-omamobile-storage.rules" \ || fail "the whole-disk node is made inaccessible" grep -q 'ID_PART_ENTRY_NAME}!="userdata"' "$tmp/root/usr/lib/udev/rules.d/10-omamobile-storage.rules" \ || fail "userdata is the only partition desktop tools may see" grep -q 'Address=172.16.42.1/24' "$tmp/root/usr/lib/systemd/network/20-usb-gadget.network" \ || fail "usb0 keeps the initramfs address" +grep -q 'Gateway=172.16.42.2' "$tmp/root/usr/lib/systemd/network/20-usb-gadget.network" \ + || fail "the host stays the fallback gateway" grep -q '^PACKAGE=omamobile-device-tegu$' "$tmp/root/usr/share/omamobile/device.conf" \ || fail "the installed descriptor is tegu" +grep -qx 'ConditionPathExists=!/usr/share/omamobile/device-tegu/version' \ + "$tmp/root/usr/lib/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf" \ + || fail "the module cleanup is held off by the package's version file" +grep -qx 'RebootWatchdogSec=1min' "$tmp/root/usr/lib/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf" \ + || fail "a hung shutdown resets the phone after a minute" +grep -q 'RuntimeWatchdogSec' "$tmp/root/usr/lib/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf" \ + && fail "no runtime watchdog from the package" +grep -q 'Before=display-manager.service' "$tmp/root/usr/lib/systemd/system/omamobile-tegu-graphics.service" \ + || fail "the graphics start runs before the display manager" +grep -q 'After=omamobile-tegu-graphics.service' "$tmp/root/usr/lib/systemd/system/display-manager.service.d/10-omamobile-tegu.conf" \ + || fail "the display manager waits for the graphics start" for absent in usr/bin/omamobile-boot-success usr/bin/omamobile-bootctl \ usr/lib/systemd/system/omamobile-boot-success.service \ usr/bin/omamobile-wifi-start usr/bin/omamobile-charge-guard \ - usr/bin/omamobile-camera-start; do + usr/bin/omamobile-camera-start usr/bin/omamobile-aoc-start \ + usr/lib/udev/rules.d/20-omamobile-gpu-always-on.rules; do [[ ! -e $tmp/root/$absent ]] || fail "install includes $absent" done -pass "install stages the udev rule, the USB link, and the mirrorlist, and nothing that marks a slot or starts hardware" +pass "install stages the package, enables only the guards, and nothing that marks a slot" + +# Every driver the graphics guard loads is kept from autoloading, and nothing +# in modprobe.d would break the guard's own modprobe. +for m in $(grep -oE '^modprobe [a-z0-9_]+' bin/omamobile-tegu-graphics-start | cut -d' ' -f2 | sort -u); do + grep -qx "blacklist $m" vendor/modprobe.d/omamobile-tegu-guards.conf \ + || fail "$m is loaded by the graphics guard but not blacklisted" +done +for m in gsa gs201_aoc_sysmmu mailbox_wc aoc_core aoc_channel_dev aoc_alsa_dev aoc_char_dev; do + grep -qx "blacklist $m" vendor/modprobe.d/omamobile-tegu-aoc.conf \ + || fail "$m is loaded by the AoC guard but not blacklisted" +done +grep -qE '^install ' vendor/modprobe.d/*.conf && fail "an install line in modprobe.d would break a guard's modprobe" +grep -qE '^options gs_dsim_cmd .*allow_tx=1' vendor/modprobe.d/omamobile-tegu-display.conf \ + || fail "the panel's DSI commands are allowed" +[[ $(grep -c 'max_refresh=' vendor/modprobe.d/omamobile-tegu-display.conf) == 1 ]] \ + || fail "exactly one refresh-rate ceiling" +grep -qE '^options clk_acpm zumapro_gpu_max_khz=[0-9]+' vendor/modprobe.d/omamobile-tegu-clk-acpm.conf \ + || fail "the GPU DVFS ceiling is set" +pass "modprobe.d keeps the guarded drivers from autoloading and sets the display and GPU options" + +grep -q 'chosen/config/bt_addr' bin/omamobile-bt-addr || fail "the Bluetooth address comes from the bootloader's DT" +grep -q 'HandlePowerKey' -r vendor && fail "no power-key policy from this package (omamobile-shell owns it on tegu)" +pass "the Bluetooth address is read at run time" + +for t in test/graphics-start test/aoc-start test/usb-rescue test/sleep-hooks test/charge-limit; do + "$t" || fail "$t" +done +pass "the scripts behave in a scratch root" diff --git a/packages/omamobile-device-tegu/test/aoc-start b/packages/omamobile-device-tegu/test/aoc-start new file mode 100755 index 000000000..d83cd0ff2 --- /dev/null +++ b/packages/omamobile-device-tegu/test/aoc-start @@ -0,0 +1,83 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# omamobile-tegu-aoc-start in a scratch root: the marker first, Trusty before +# the GSA before the AoC drivers, the firmware request, and the ways out +# (no node, no firmware, a leftover marker, a second start on the same boot). +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +script=$PWD/bin/omamobile-tegu-aoc-start +fail() { printf 'not ok - aoc-start: %s\n' "$1" >&2; exit 1; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT +export OMAMOBILE_TEGU_ROOT=$tmp/root +export TEST_LOG=$tmp/calls +export TEST_ROOT=$OMAMOBILE_TEGU_ROOT +marker=$OMAMOBILE_TEGU_ROOT/var/lib/omamobile/tegu-aoc-started +export TEST_MARKER=$marker +node=$OMAMOBILE_TEGU_ROOT/sys/bus/platform/devices/154b0000.aoc +mkdir -p "$tmp/bin" "$node" "$OMAMOBILE_TEGU_ROOT/proc/device-tree/trusty" "$OMAMOBILE_TEGU_ROOT/usr/lib/firmware" "$OMAMOBILE_TEGU_ROOT/dev" +printf 'okay\0' > "$OMAMOBILE_TEGU_ROOT/proc/device-tree/trusty/status" +: > "$OMAMOBILE_TEGU_ROOT/usr/lib/firmware/aoc-tegu.bin" +export PATH=$tmp/bin:$PATH +cat > "$tmp/bin/modprobe" <<'EOF' +#!/bin/bash +set -euo pipefail +test -s "$TEST_MARKER" +grep -qx synced "$TEST_LOG" +echo "$1" >> "$TEST_LOG" +[[ ${FAIL_MODULE:-} != "$1" ]] || exit 1 +case $1 in + trusty_virtio) : > "$TEST_ROOT/dev/trusty-ipc-dev0" ;; + aoc_core) : > "$TEST_ROOT/sys/bus/platform/devices/154b0000.aoc/firmware" ;; + aoc_char_dev) echo 'Services : 93' > "$TEST_ROOT/sys/bus/platform/devices/154b0000.aoc/services"; : > "$TEST_ROOT/dev/acd-com.google.usf" ;; +esac +EOF +printf '#!/bin/bash\necho synced >> "$TEST_LOG"\n' > "$tmp/bin/sync" +printf '#!/bin/bash\nexit 0\n' > "$tmp/bin/sleep" +chmod +x "$tmp/bin/"* +reset() { + : > "$TEST_LOG" + rm -rf "$marker" "$OMAMOBILE_TEGU_ROOT/run" "$node/firmware" "$OMAMOBILE_TEGU_ROOT/dev/"* + echo 'Services : 0' > "$node/services" +} +loads() { grep -v '^synced$' "$TEST_LOG" | tr '\n' ' ' | sed 's/ $//'; } + +reset +out=$("$script") +[[ $(loads) == "trusty_core trusty_log trusty_ipc trusty_virtio gsa gs201_aoc_sysmmu mailbox_wc aoc_core aoc_channel_dev aoc_alsa_dev aoc_char_dev" ]] \ + || fail "load order: $(loads)" +[[ $(cat "$node/firmware") == aoc-tegu.bin ]] || fail "the firmware request names aoc-tegu.bin" +[[ ! -e $marker ]] || fail "a finished start leaves the marker" +[[ -e $OMAMOBILE_TEGU_ROOT/run/omamobile/tegu-aoc-firmware-loaded ]] || fail "no firmware-loaded note for this boot" +[[ $out == *"up: Services : 93"* ]] || fail "summary: $out" + +# Started again on the same boot while not online: refuse, no loads +echo 'Services : 0' > "$node/services"; : > "$TEST_LOG" +"$script" >/dev/null && fail "a second start on the same boot was accepted" +[[ -z $(loads) ]] || fail "a second start loaded drivers" + +# A leftover marker keeps the AoC off +reset +echo stale > "$marker" +out=$("$script") +[[ -z $(loads) && $out == *"--retry"* ]] || fail "a leftover marker did not keep the AoC off" + +# A failed GSA load: exit 1, the marker is cleared (the load returned, no hang) +reset +FAIL_MODULE=gsa "$script" >/dev/null && fail "a failed load exits 0" +[[ $(loads) == "trusty_core trusty_log trusty_ipc trusty_virtio gsa" && ! -e $marker ]] \ + || fail "a failed load went on or kept the marker: $(loads)" + +# No firmware: refuse before anything loads +reset +mv "$OMAMOBILE_TEGU_ROOT/usr/lib/firmware/aoc-tegu.bin" "$tmp/fw" +"$script" >/dev/null && fail "a missing firmware was accepted" +[[ -z $(loads) && ! -e $marker ]] || fail "a missing firmware loaded drivers" +mv "$tmp/fw" "$OMAMOBILE_TEGU_ROOT/usr/lib/firmware/aoc-tegu.bin" + +# No AoC node: nothing to do, exit 0 +reset +mv "$node" "$tmp/node" +out=$("$script") +[[ -z $(loads) && $out == *"not enabled"* ]] || fail "a missing node was not skipped" +mv "$tmp/node" "$node" +echo 'ok - aoc-start: marker first, Trusty then GSA then the AoC, firmware named, refusals in place' diff --git a/packages/omamobile-device-tegu/test/charge-limit b/packages/omamobile-device-tegu/test/charge-limit new file mode 100755 index 000000000..6e9c3d16e --- /dev/null +++ b/packages/omamobile-device-tegu/test/charge-limit @@ -0,0 +1,32 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# omamobile-tegu-charge-limit in a scratch root: a percentage is written to +# the charger, anything else is refused and leaves the charger alone. +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +script=$PWD/bin/omamobile-tegu-charge-limit +fail() { printf 'not ok - charge-limit: %s\n' "$1" >&2; exit 1; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT +export OMAMOBILE_TEGU_ROOT=$tmp/root +file=$OMAMOBILE_TEGU_ROOT/etc/omamobile/charge-limit +threshold=$OMAMOBILE_TEGU_ROOT/sys/class/power_supply/max77779-charger/charge_control_end_threshold +mkdir -p "${file%/*}" "${threshold%/*}" +echo 100 > "$threshold" + +"$script" || fail "no file is not an error" +[[ $(cat "$threshold") == 100 ]] || fail "no file changed the charger" +echo 80 > "$file" +out=$("$script") || fail "80 is refused" +[[ $(cat "$threshold") == 80 && $out == *"charge limit 80 %"* ]] || fail "80 was not written: $out" +printf ' 090\n' > "$file" +"$script" >/dev/null || fail "a padded number is refused" +[[ $(cat "$threshold") == 90 ]] || fail "090 was not written as 90" +for bad in abc 0 101 -5 8.5; do + echo "$bad" > "$file" + "$script" >/dev/null 2>&1 && fail "$bad was accepted" + [[ $(cat "$threshold") == 90 ]] || fail "$bad changed the charger" +done +rm -f "$threshold" +echo 80 > "$file" +"$script" >/dev/null || fail "a missing charger is an error" +echo 'ok - charge-limit: percentages only, the charger otherwise untouched' diff --git a/packages/omamobile-device-tegu/test/graphics-start b/packages/omamobile-device-tegu/test/graphics-start new file mode 100755 index 000000000..981c1d15f --- /dev/null +++ b/packages/omamobile-device-tegu/test/graphics-start @@ -0,0 +1,111 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# omamobile-tegu-graphics-start in a scratch root: the marker goes down and +# is synced before the first driver loads, the drivers load in order, a +# failure leaves the marker naming the step, and a marker from a previous +# boot keeps every driver off until --retry. No hardware is touched: modprobe, +# sync and sleep are fakes on PATH. +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +script=$PWD/bin/omamobile-tegu-graphics-start +fail() { printf 'not ok - graphics-start: %s\n' "$1" >&2; exit 1; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT +export OMAMOBILE_TEGU_ROOT=$tmp/root +export TEST_LOG=$tmp/calls +export TEST_ROOT=$OMAMOBILE_TEGU_ROOT +marker=$OMAMOBILE_TEGU_ROOT/var/lib/omamobile/tegu-graphics-started +export TEST_MARKER=$marker +dt=$OMAMOBILE_TEGU_ROOT/proc/device-tree +devices=$OMAMOBILE_TEGU_ROOT/sys/bus/platform/devices +mkdir -p "$tmp/bin" "$devices/1f000000.gpu" "$devices/19470000.display-controller" \ + "$OMAMOBILE_TEGU_ROOT/sys/bus/platform/drivers/panthor" "$OMAMOBILE_TEGU_ROOT/sys/bus/platform/drivers/gs-decon" \ + "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43" \ + "$dt/soc/usi@108b00c0/i2c@108b0000/led-controller@63" \ + "$dt/soc/usi@155d00c0/serial@155d0000/bluetooth" +printf 'okay\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" +export PATH=$tmp/bin:$PATH +# modprobe: the marker must already be on disk; a module "binds" by creating +# what the script waits for; FAIL_MODULE fails that one load. +cat > "$tmp/bin/modprobe" <<'EOF' +#!/bin/bash +set -euo pipefail +test -s "$TEST_MARKER" +grep -qx synced "$TEST_LOG" +echo "$*" >> "$TEST_LOG" +[[ ${FAIL_MODULE:-} != "$1" ]] || exit 1 +case $1 in + panthor) ln -sfn "$TEST_ROOT/sys/bus/platform/drivers/panthor" "$TEST_ROOT/sys/bus/platform/devices/1f000000.gpu/driver" ;; + gs_decon) ln -sfn "$TEST_ROOT/sys/bus/platform/drivers/gs-decon" "$TEST_ROOT/sys/bus/platform/devices/19470000.display-controller/driver" ;; + leds_lm3644) mkdir -p "$TEST_ROOT/sys/class/leds/white:torch"; echo "${TORCH_LIT:-0}" > "$TEST_ROOT/sys/class/leds/white:torch/brightness" ;; + max77779_battery) mkdir -p "$TEST_ROOT/sys/class/power_supply/max77779-battery"; echo 77 > "$TEST_ROOT/sys/class/power_supply/max77779-battery/capacity" ;; +esac +EOF +printf '#!/bin/bash\necho synced >> "$TEST_LOG"\n' > "$tmp/bin/sync" +printf '#!/bin/bash\nexit 0\n' > "$tmp/bin/sleep" +chmod +x "$tmp/bin/"* +reset() { + : > "$TEST_LOG" + rm -rf "$marker" "$devices/1f000000.gpu/driver" "$devices/19470000.display-controller/driver" \ + "$OMAMOBILE_TEGU_ROOT/sys/class" "$OMAMOBILE_TEGU_ROOT/sys/module" +} +loads() { grep -v '^synced$' "$TEST_LOG" | cut -d' ' -f1 | tr '\n' ' ' | sed 's/ $//'; } + +# A full start: the order, the gs_decon options, the marker gone, the summary +reset +out=$("$script") +[[ ! -e $marker ]] || fail "a finished start leaves the marker" +[[ $(loads) == "panthor gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] \ + || fail "load order: $(loads)" +grep -q '^gs_decon stage=2 fbdev=0$' "$TEST_LOG" || fail "gs_decon gets the full driver stage" +[[ $out == *"up: GPU, display, haptics, torch, gauge 77 %, Bluetooth"* ]] || fail "summary: $out" + +# A failed display load: exit 1, the marker names the step, nothing after it +reset +FAIL_MODULE=gs_decon "$script" >/dev/null && fail "a failed load exits 0" +[[ -s $marker ]] || fail "a failed load removes the marker" +grep -q ' display$' "$marker" || fail "the marker names the failed step: $(cat "$marker")" +[[ $(loads) == "panthor gs_decon" ]] || fail "a failed load went on: $(loads)" + +# The next boot with that marker: nothing loads, the hint is printed +: > "$TEST_LOG" +rm -rf "$devices/1f000000.gpu/driver" "$devices/19470000.display-controller/driver" +out=$("$script") +[[ -s $marker && -z $(loads) ]] || fail "a leftover marker did not keep the drivers off" +[[ $out == *"--retry"* ]] || fail "no retry hint" + +# --retry clears it and starts again +out=$("$script" --retry) +[[ ! -e $marker && $(loads) == "panthor gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] \ + || fail "--retry did not start again: $(loads)" + +# A torch that comes up lit is a failure that keeps the marker +reset +TORCH_LIT=1 "$script" >/dev/null && fail "a lit torch passed" +grep -q ' torch$' "$marker" || fail "the marker names the torch step" + +# A disabled haptics node is skipped, a missing Bluetooth node too +reset +printf 'disabled\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" +mv "$dt/soc/usi@155d00c0" "$tmp/bt-node" +out=$("$script") +[[ $(loads) == "panthor gs_decon leds_lm3644 max77779_battery" ]] || fail "disabled nodes were loaded: $(loads)" +[[ $out == *"no haptics in this DTB"* && $out == *"no Bluetooth in this DTB"* ]] || fail "skips are not reported" +printf 'okay\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" +mv "$tmp/bt-node" "$dt/soc/usi@155d00c0" + +# A driver already loaded this boot is never retried +reset +mkdir -p "$OMAMOBILE_TEGU_ROOT/sys/module/panthor" +out=$("$script") +[[ -z $(loads) && ! -e $marker && $out == *"already started"* ]] || fail "a loaded driver was retried" + +# No GPU in the DTB: nothing to do +reset +mv "$devices/1f000000.gpu" "$tmp/gpu" +out=$("$script") +[[ -z $(loads) && ! -e $marker && $out == *"no GPU"* ]] || fail "a missing GPU node was not skipped" +mv "$tmp/gpu" "$devices/1f000000.gpu" + +# A bad argument +"$script" --bogus >/dev/null 2>&1 && fail "a bad argument is accepted" +echo 'ok - graphics-start: marker first, order kept, failures named, --retry works' diff --git a/packages/omamobile-device-tegu/test/sleep-hooks b/packages/omamobile-device-tegu/test/sleep-hooks new file mode 100755 index 000000000..e129cba80 --- /dev/null +++ b/packages/omamobile-device-tegu/test/sleep-hooks @@ -0,0 +1,90 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# The sleep hooks in a scratch root with fake lsmod, modprobe, rmmod, +# systemctl, logger and busctl: Wi-Fi is left alone when kept and present, +# unloaded and reloaded otherwise, and falls back after a lost sleep; NFC is +# stopped and powered off before the sleep and brought back after it. +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +wifi=$PWD/vendor/systemd/system-sleep/omamobile-tegu-wifi +nfc=$PWD/vendor/systemd/system-sleep/omamobile-tegu-nfc +fail() { printf 'not ok - sleep-hooks: %s\n' "$1" >&2; exit 1; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT +export OMAMOBILE_TEGU_ROOT=$tmp/root +export TEST_LOG=$tmp/calls +export TEST_LOADED=$tmp/loaded +export TEST_NFC=$tmp/nfc +mkdir -p "$tmp/bin" "$OMAMOBILE_TEGU_ROOT/etc/omamobile" "$TEST_NFC" +export PATH=$tmp/bin:$PATH +printf '#!/bin/bash\n[[ -e $TEST_LOADED ]] && echo "brcmfmac 1 0"\nexit 0\n' > "$tmp/bin/lsmod" +printf '#!/bin/bash\necho "modprobe $*" >> "$TEST_LOG"; : > "$TEST_LOADED"\n' > "$tmp/bin/modprobe" +printf '#!/bin/bash\necho "rmmod $*" >> "$TEST_LOG"; rm -f "$TEST_LOADED"\n' > "$tmp/bin/rmmod" +printf '#!/bin/bash\necho "systemctl $*" >> "$TEST_LOG"\n' > "$tmp/bin/systemctl" +printf '#!/bin/bash\necho "logger $*" >> "$TEST_LOG"\n' > "$tmp/bin/logger" +printf '#!/bin/bash\nshift; exec "$@"\n' > "$tmp/bin/timeout" +printf '#!/bin/bash\nexit 0\n' > "$tmp/bin/sleep" +printf '#!/bin/bash\nexit 0\n' > "$tmp/bin/sync" +# busctl: Polling and Powered are files in TEST_NFC; calls and sets are logged +cat > "$tmp/bin/busctl" <<'EOF' +#!/bin/bash +case $1 in + get-property) [[ -e $TEST_NFC/$5 ]] && echo 'b true' || echo 'b false' ;; + call) echo "busctl call $5" >> "$TEST_LOG" + [[ $5 == StopPollLoop ]] && rm -f "$TEST_NFC/Polling" + [[ $5 == StartPollLoop ]] && : > "$TEST_NFC/Polling" ;; + set-property) echo "busctl set $5 $7" >> "$TEST_LOG" + [[ $7 == true ]] && : > "$TEST_NFC/Powered" || rm -f "$TEST_NFC/Powered" ;; +esac +exit 0 +EOF +chmod +x "$tmp/bin/"* +calls() { tr '\n' ';' < "$TEST_LOG" | sed 's/;$//'; } +run=$OMAMOBILE_TEGU_ROOT/run/omamobile +keep=$OMAMOBILE_TEGU_ROOT/etc/omamobile/wifi-keep-across-sleep +iface() { mkdir -p "$OMAMOBILE_TEGU_ROOT/sys/bus/pci/drivers/brcmfmac/0000:01:00.0" "$OMAMOBILE_TEGU_ROOT/sys/class/ieee80211/phy0"; } +no_iface() { rm -rf "$OMAMOBILE_TEGU_ROOT/sys"; } + +# Wi-Fi never loaded this boot: the hook does nothing +: > "$TEST_LOG"; rm -f "$TEST_LOADED" "$keep" +"$wifi" pre; "$wifi" post +[[ -z $(calls) ]] || fail "an unloaded Wi-Fi was touched: $(calls)" + +# Loaded, no keep flag: unload before, reload after +: > "$TEST_LOG"; : > "$TEST_LOADED" +"$wifi" pre +[[ $(calls) == "systemctl stop iwd;rmmod brcmfmac_wcc brcmfmac" && -e $run/wifi-unloaded-for-sleep ]] \ + || fail "no keep flag: not unloaded before the sleep: $(calls)" +"$wifi" post +[[ $(calls) == "systemctl stop iwd;rmmod brcmfmac_wcc brcmfmac;modprobe brcmfmac;systemctl start iwd" && ! -e $run/wifi-unloaded-for-sleep ]] \ + || fail "no keep flag: not reloaded after the sleep: $(calls)" + +# Kept and back after the sleep: nothing happens +: > "$TEST_LOG"; : > "$TEST_LOADED"; : > "$keep"; iface +"$wifi" pre; "$wifi" post +[[ -z $(calls) ]] || fail "kept Wi-Fi was touched: $(calls)" + +# Kept but lost in the sleep: fall back to unloading for the rest of the boot +: > "$TEST_LOG"; no_iface +"$wifi" post +[[ $(calls) == "logger -t omamobile-tegu-wifi Wi-Fi lost across sleep; unloading it around sleeps until reboot;systemctl stop iwd;rmmod brcmfmac_wcc brcmfmac;modprobe brcmfmac;systemctl start iwd" ]] \ + || fail "a lost Wi-Fi was not reloaded: $(calls)" +[[ -e $run/wifi-unload-across-sleep ]] || fail "no fallback flag after a lost sleep" +: > "$TEST_LOG" +"$wifi" pre +[[ $(calls) == "systemctl stop iwd;rmmod brcmfmac_wcc brcmfmac" ]] || fail "the fallback flag is not honoured: $(calls)" + +# NFC powered and polling: stopped and powered off before, back after +: > "$TEST_LOG"; : > "$TEST_NFC/Polling"; : > "$TEST_NFC/Powered" +"$nfc" pre +[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false" ]] || fail "NFC pre: $(calls)" +[[ -e $run/nfc-was-polling && -e $run/nfc-was-powered ]] || fail "NFC pre left no notes" +"$nfc" post +[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false;busctl set Powered true;busctl call StartPollLoop" ]] || fail "NFC post: $(calls)" +[[ ! -e $run/nfc-was-polling && ! -e $run/nfc-was-powered ]] || fail "NFC post left its notes" +grep -q '^.* pre: polling=b false powered=b false$' "$OMAMOBILE_TEGU_ROOT/var/log/omamobile/nfc-sleep.log" || fail "NFC pre is not logged" + +# NFC off (no neard): nothing is called +: > "$TEST_LOG"; rm -f "$TEST_NFC"/* +"$nfc" pre; "$nfc" post +[[ -z $(calls) ]] || fail "an off NFC adapter was touched: $(calls)" +echo 'ok - sleep-hooks: Wi-Fi kept or reloaded as flagged, NFC stopped and restored' diff --git a/packages/omamobile-device-tegu/test/usb-rescue b/packages/omamobile-device-tegu/test/usb-rescue new file mode 100755 index 000000000..4070a79d0 --- /dev/null +++ b/packages/omamobile-device-tegu/test/usb-rescue @@ -0,0 +1,45 @@ +#!/bin/bash +# SPDX-License-Identifier: MIT +# omamobile-tegu-usb-rescue in a scratch root, one pass at a time: a +# configured gadget is left alone, an open switch on an unconfigured gadget +# is closed once, a closed switch is not written, and a Type-C driver that +# owns the chip ends the watch before the chip is read. +set -euo pipefail +cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." +script=$PWD/bin/omamobile-tegu-usb-rescue +fail() { printf 'not ok - usb-rescue: %s\n' "$1" >&2; exit 1; } +tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT +export OMAMOBILE_TEGU_ROOT=$tmp/root +export OMAMOBILE_TEGU_RESCUE_GRACE=0 +export TEST_LOG=$tmp/calls +export TEGU_TCPC_SWITCH=$tmp/bin/tegu-tcpc-switch +udc=$OMAMOBILE_TEGU_ROOT/sys/class/udc/11110000.dwc3 +mkdir -p "$tmp/bin" "$udc" "$OMAMOBILE_TEGU_ROOT/sys/bus/i2c/devices/21-0025" +cat > "$tmp/bin/tegu-tcpc-switch" <<'EOF' +#!/bin/bash +echo "$1" >> "$TEST_LOG" +case $1 in switch) echo "${SW:-0x09}" ;; connect) echo 0x09 ;; esac +EOF +chmod +x "$tmp/bin/tegu-tcpc-switch" +calls() { tr '\n' ' ' < "$TEST_LOG" | sed 's/ $//'; } + +: > "$TEST_LOG"; echo configured > "$udc/state" +SW=0x00 "$script" --once >/dev/null +[[ -z $(calls) ]] || fail "a configured gadget was touched: $(calls)" + +: > "$TEST_LOG"; echo 'not attached' > "$udc/state" +out=$(SW=0x00 "$script" --once) +[[ $(calls) == "switch connect" ]] || fail "an open switch was not closed once: $(calls)" +[[ $out == *"closing the data switch"* ]] || fail "the close is not logged" + +: > "$TEST_LOG" +SW=0x09 "$script" --once >/dev/null +[[ $(calls) == "switch" ]] || fail "a closed switch was written: $(calls)" + +: > "$TEST_LOG"; ln -s /dev/null "$OMAMOBILE_TEGU_ROOT/sys/bus/i2c/devices/21-0025/driver" +out=$(SW=0x00 "$script" --once) +[[ -z $(calls) && $out == *"owns the chip"* ]] || fail "a Type-C driver did not end the watch: $(calls)" + +rm -f "$TEGU_TCPC_SWITCH" +"$script" --once >/dev/null || fail "a missing helper is not fatal" +echo 'ok - usb-rescue: closes an open switch once, defers to a Type-C driver' diff --git a/packages/omamobile-device-tegu/version b/packages/omamobile-device-tegu/version index b1e80bb24..0ea3a944b 100644 --- a/packages/omamobile-device-tegu/version +++ b/packages/omamobile-device-tegu/version @@ -1 +1 @@ -0.1.3 +0.2.0 From 596fce30ecbe0f26f4c858ea2652e7a73a76f8d3 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:25:00 -0647 Subject: [PATCH 10/20] omamobile-device-tegu: the blob scan skips makepkg's pkg and src trees check() runs with pkg/ beside the sources, and the scan died on it. --- packages/omamobile-device-tegu/test/all | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all index f91059cc2..12a8c3f13 100755 --- a/packages/omamobile-device-tegu/test/all +++ b/packages/omamobile-device-tegu/test/all @@ -55,7 +55,8 @@ if [[ -f ../../device/tegu/device.conf ]]; then fi # Nothing that must come from the user's own phone or a kernel build -blobs=$(find . -type f \( -name '*.ko' -o -name '*.bin' -o -name '*.hcd' -o -name '*.img' -o -name '*.wmfw' \)) +# makepkg's pkg/ and src/build/ are not the package (check() runs beside them) +blobs=$(find . \( -path ./pkg -o -path ./src/build \) -prune -o -type f \( -name '*.ko' -o -name '*.bin' -o -name '*.hcd' -o -name '*.img' -o -name '*.wmfw' \) -print) [[ -z $blobs ]] || fail "binaries in the package: $blobs" macs=$(grep -rEn '\b([0-9a-fA-F]{2}:){5}[0-9a-fA-F]{2}\b' --exclude-dir=build . || true) [[ -z $macs ]] || fail "a MAC or Bluetooth address in the package: $macs" From a7d728f499cd56fee3962a6ff8b8c5b3a3552a56 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:31:39 -0647 Subject: [PATCH 11/20] omamobile-device-tegu: --retry goes on within the same boot The first package boot on the palmtop stopped at the gauge (a stale bring-up modprobe.d file still had install max77779_battery /bin/false), and --retry then refused to run because the GPU was already loaded. A retry now continues the sequence; modprobe leaves the loaded drivers alone. --- packages/omamobile-device-tegu/README.md | 2 +- .../omamobile-device-tegu/bin/omamobile-tegu-graphics-start | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index 914530a90..c9f90667a 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -10,7 +10,7 @@ packaged yet. What is here is what runs on in-tree drivers. | Path | Installs to | What it does | | --- | --- | --- | -| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` tries again. The display manager waits for it | +| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). The display manager waits for it | | `vendor/modprobe.d/omamobile-tegu-guards.conf` | `/usr/lib/modprobe.d` | Keeps those drivers (and the Wi-Fi, PCIe and NFC ones, for the day they are in tree) from autoloading ahead of the guard. No `install` lines: they would break the guard's own modprobe | | `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | DSI commands for brightness and sleep, 120 Hz, the panel really off on DPMS; the GPU DVFS ceiling at 890 MHz | | `bin/omamobile-tegu-aoc-start` + `vendor/systemd/system/omamobile-tegu-aoc.service` + `vendor/modprobe.d/omamobile-tegu-aoc.conf` | `/usr/bin`, `/usr/lib/systemd/system` (not enabled), `/usr/lib/modprobe.d` | Boots the AoC (audio and sensor DSP) hang-safe: Trusty, the GSA, the AoC drivers, `aoc-tegu.bin`, then waits for its 93 services. Ships disabled: see below | diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start index 05155041c..f9bef5ad9 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start @@ -64,7 +64,9 @@ if [[ ! -e $gpu ]]; then exit 0 fi # Once per boot: a loaded driver is never unloaded or retried (AGENTS.md). -if loaded panthor || loaded gs_decon; then +# --retry goes on from where the last start stopped: modprobe leaves a +# loaded driver as it is, so the steps already done cost nothing. +if [[ ${1:-} != --retry ]] && { loaded panthor || loaded gs_decon; }; then log "already started this boot" exit 0 fi From 422ef555a8ffcc336588dca8df4c9256c6c5ed8d Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:34:00 -0647 Subject: [PATCH 12/20] omamobile-device-tegu: say which brcmfmac the Wi-Fi sleep hook reloads --- packages/omamobile-device-tegu/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index c9f90667a..98b881392 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -17,7 +17,7 @@ packaged yet. What is here is what runs on in-tree drivers. | `bin/omamobile-bt-addr` + `vendor/systemd/system/omamobile-bt-addr@.service` + `vendor/udev/rules.d/80-omamobile-bt-addr.rules` | `/usr/bin`, `/usr/lib/systemd/system`, `/usr/lib/udev/rules.d` | Gives the BCM4383 its real Bluetooth address from the bootloader's device tree when the controller appears, powering it on once first (set on a never-powered controller, the kernel withdraws the index for good). The Pixel 7 Pro's script, with the extra waits the 9a needed | | `bin/omamobile-tegu-usb-rescue` + `src/tegu-tcpc-switch.c` + `vendor/systemd/system/omamobile-tegu-usb-rescue.service` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Closes the MAX77779's USB data switch after a cable replug while no Type-C driver owns the chip; with one loaded it exits. The helper reads the chip's identity and VBUS before its one write and never forces an address a driver holds | | `bin/omamobile-tegu-charge-limit` + `vendor/udev/rules.d/75-omamobile-tegu-charger.rules` | `/usr/bin`, `/usr/lib/udev/rules.d` | Applies `/etc/omamobile/charge-limit` (a percentage, say `80`) to the charger when it appears; charging stops there and starts again 5 below. No file: charge to 100 % | -| `vendor/systemd/system-sleep/omamobile-tegu-wifi` | `/usr/lib/systemd/system-sleep` | Keeps Wi-Fi loaded across s2idle when `/etc/omamobile/wifi-keep-across-sleep` exists (the chip reloads its firmware on resume and iwd reconnects), otherwise unloads and reloads brcmfmac around the sleep; after a sleep that lost Wi-Fi it unloads for the rest of the boot | +| `vendor/systemd/system-sleep/omamobile-tegu-wifi` | `/usr/lib/systemd/system-sleep` | Keeps Wi-Fi loaded across s2idle when `/etc/omamobile/wifi-keep-across-sleep` exists (the chip reloads its firmware on resume and iwd reconnects), otherwise unloads and reloads brcmfmac around the sleep; after a sleep that lost Wi-Fi it unloads for the rest of the boot. It reloads the in-tree `brcmfmac`; a phone still running an out-of-tree build of it keeps its own hook until the Wi-Fi series lands | | `vendor/systemd/system-sleep/omamobile-tegu-nfc` | `/usr/lib/systemd/system-sleep` | Stops neard's poll loop and powers the ST54L off before a sleep (it does not answer on I2C after a wake otherwise), and brings it back afterwards | | `vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf` | `/usr/lib/systemd/system` | Holds off Arch's old-module cleanup: our kernels come as module tarballs, not pacman packages, and the next-boot and rollback trees must stay | | `vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf`, `sleep.conf.d/10-omamobile.conf` | `/usr/lib/systemd` | A hung shutdown resets the phone after a minute; suspend means s2idle (no SoC sleep sequence yet, and `deep` does not wake) | From 9084b92f24be4b307fe775453c3f0f92bffc7130 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 09:47:22 -0647 Subject: [PATCH 13/20] omamobile-device-tegu: tell the session which DRM node is the display The session's uwsm/env takes the display controller's card from /run/omamobile-tegu-display-card and hardware GL from /run/omamobile-tegu-gpu-ready.json, both written by the curated start. Without them the first packaged boot gave Hyprland /dev/dri/card0, the GPU on that boot, and it aborted with no KMS device; sddm then showed its greeter, which takes no touch. The graphics start writes both now. --- packages/omamobile-device-tegu/README.md | 2 +- .../omamobile-device-tegu/bin/omamobile-tegu-graphics-start | 6 ++++++ 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index 98b881392..d23ee21b6 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -10,7 +10,7 @@ packaged yet. What is here is what runs on in-tree drivers. | Path | Installs to | What it does | | --- | --- | --- | -| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). The display manager waits for it | +| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). The display manager waits for it. It also writes `/run/omamobile-tegu-display-card` (the display controller's DRM node by path, since the GPU may take card0) and `/run/omamobile-tegu-gpu-ready.json`, which the session's `uwsm/env` reads to pick the KMS device and hardware GL | | `vendor/modprobe.d/omamobile-tegu-guards.conf` | `/usr/lib/modprobe.d` | Keeps those drivers (and the Wi-Fi, PCIe and NFC ones, for the day they are in tree) from autoloading ahead of the guard. No `install` lines: they would break the guard's own modprobe | | `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | DSI commands for brightness and sleep, 120 Hz, the panel really off on DPMS; the GPU DVFS ceiling at 890 MHz | | `bin/omamobile-tegu-aoc-start` + `vendor/systemd/system/omamobile-tegu-aoc.service` + `vendor/modprobe.d/omamobile-tegu-aoc.conf` | `/usr/bin`, `/usr/lib/systemd/system` (not enabled), `/usr/lib/modprobe.d` | Boots the AoC (audio and sensor DSP) hang-safe: Trusty, the GSA, the AoC drivers, `aoc-tegu.bin`, then waits for its 93 services. Ships disabled: see below | diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start index f9bef5ad9..0835e6793 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start @@ -86,6 +86,12 @@ bound "$gpu" panthor 60 || fail "the GPU did not bind to panthor" phase display modprobe gs_decon stage=2 fbdev=0 || fail "gs_decon failed" bound "$decon" gs-decon 60 || fail "the display controller did not bind to gs-decon" +# The session's environment (uwsm/env) reads these: the display controller's +# DRM node, whichever card number it got, and that the GPU is up. +mkdir -p "$root/run" +card=$(readlink -f "$root/dev/dri/by-path/platform-19470000.display-controller-card" 2>/dev/null || true) +[[ -n $card ]] && printf '%s\n' "${card#"$root"}" > "$root/run/omamobile-tegu-display-card" +printf '{"gpu": "panthor", "by": "omamobile-tegu-graphics-start"}\n' > "$root/run/omamobile-tegu-gpu-ready.json" up="GPU, display" if enabled "$haptics_node"; then From fcf07de0d578a84afe00f103f40c991ea5a2adfe Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:18 -0647 Subject: [PATCH 14/20] omamobile-device-tegu: a GPU that fails must not take the display with it; allow_tx for google-tg4c only The graphics start keyed everything on the GPU node, and a panthor load or bind that returned and failed kept the marker and skipped gs_decon. Since gs_decon is blacklisted, nothing else loads it: a phone without mali_csffw.bin yet got no display, and the marker kept it off at every boot until --retry over SSH; a DTB with the display but no GPU went dark. Now each step keys on its own device node (1f000000.gpu for panthor, 19470000.display-controller for gs_decon, the device-tree nodes for the rest), the unit's two ConditionPathExists=| lines run it when either exists, and a load or bind that returns and fails is logged and skipped while the rest still loads; the marker is cleared at the end as after a good start, and the unit ends failed. Only a step that never comes back keeps the marker, which is what it is for. A torch that comes up lit is turned off and counted as failed instead of stopping the start. The gpu-ready file is written only when the GPU came up, so the session falls back to software GL. The panel's DSI commands (gs_dsim_cmd.allow_tx) leave modprobe.d, where they applied to every panel: the start turns the parameter on, before gs_decon loads (gs-decon decides on 120 Hz at probe), only when the command line says exynos_drm.panel_name=google-tg4c, the panel the palmtop's bring-up tested. On another panel the display stays read-only at 60 Hz. gs_dsim_cmd is loaded first so the parameter file exists; it is not guarded, since it autoloads from its node ahead of the start and sends nothing until allow_tx is on. Tests: google-tg4c reads allow_tx 1 when gs_decon loads, another panel and no panel name read 0, a failed gs_decon or panthor still loads the rest and clears the marker, no GPU node still loads gs_decon, no display node still loads panthor, a hang marker still keeps everything off, and the unit's two conditions and the modprobe.d file are checked. --- packages/omamobile-device-tegu/README.md | 4 +- .../bin/omamobile-tegu-graphics-start | 143 +++++++++++++----- packages/omamobile-device-tegu/test/all | 22 ++- .../omamobile-device-tegu/test/graphics-start | 132 +++++++++++----- .../modprobe.d/omamobile-tegu-display.conf | 6 +- .../system/omamobile-tegu-graphics.service | 7 +- 6 files changed, 230 insertions(+), 84 deletions(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index d23ee21b6..e6c36f73a 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -10,9 +10,9 @@ packaged yet. What is here is what runs on in-tree drivers. | Path | Installs to | What it does | | --- | --- | --- | -| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. A boot that hangs in one of them leaves the marker, and the next boot keeps all six off and says so; `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). The display manager waits for it. It also writes `/run/omamobile-tegu-display-card` (the display controller's DRM node by path, since the GPU may take card0) and `/run/omamobile-tegu-gpu-ready.json`, which the session's `uwsm/env` reads to pick the KMS device and hardware GL | +| `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. Each step keys on its own device node (`1f000000.gpu`, `19470000.display-controller`, the device-tree nodes of the rest), and the unit runs when the GPU or the display controller exists, so a DTB with the display but no GPU still gets its display. A load or bind that returns but fails (panthor without `mali_csffw.bin`, say) is logged and skipped, the rest still loads, and the marker is cleared; the unit then ends failed, and `--retry` tries the skipped steps again. Only a boot that hangs in a step leaves the marker: the next boot keeps all six off and says so, and `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). Before `gs_decon`, and only when the command line says `exynos_drm.panel_name=google-tg4c`, it turns `gs_dsim_cmd.allow_tx` on, so the panel's DSI commands (brightness, sleep, 120 Hz; kernel #90) go to the panel the bring-up tested and no other: on another panel the display stays read-only at 60 Hz. The display manager waits for it. It also writes `/run/omamobile-tegu-display-card` (the display controller's DRM node by path, since the GPU may take card0) and, when the GPU came up, `/run/omamobile-tegu-gpu-ready.json` | | `vendor/modprobe.d/omamobile-tegu-guards.conf` | `/usr/lib/modprobe.d` | Keeps those drivers (and the Wi-Fi, PCIe and NFC ones, for the day they are in tree) from autoloading ahead of the guard. No `install` lines: they would break the guard's own modprobe | -| `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | DSI commands for brightness and sleep, 120 Hz, the panel really off on DPMS; the GPU DVFS ceiling at 890 MHz | +| `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | 120 Hz as the ceiling and the panel really off on DPMS (the DSI commands themselves are allowed by the graphics start, for google-tg4c only); the GPU DVFS ceiling at 890 MHz | | `bin/omamobile-tegu-aoc-start` + `vendor/systemd/system/omamobile-tegu-aoc.service` + `vendor/modprobe.d/omamobile-tegu-aoc.conf` | `/usr/bin`, `/usr/lib/systemd/system` (not enabled), `/usr/lib/modprobe.d` | Boots the AoC (audio and sensor DSP) hang-safe: Trusty, the GSA, the AoC drivers, `aoc-tegu.bin`, then waits for its 93 services. Ships disabled: see below | | `bin/omamobile-bt-addr` + `vendor/systemd/system/omamobile-bt-addr@.service` + `vendor/udev/rules.d/80-omamobile-bt-addr.rules` | `/usr/bin`, `/usr/lib/systemd/system`, `/usr/lib/udev/rules.d` | Gives the BCM4383 its real Bluetooth address from the bootloader's device tree when the controller appears, powering it on once first (set on a never-powered controller, the kernel withdraws the index for good). The Pixel 7 Pro's script, with the extra waits the 9a needed | | `bin/omamobile-tegu-usb-rescue` + `src/tegu-tcpc-switch.c` + `vendor/systemd/system/omamobile-tegu-usb-rescue.service` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Closes the MAX77779's USB data switch after a cable replug while no Type-C driver owns the chip; with one loaded it exits. The helper reads the chip's identity and VBUS before its one write and never forces an address a driver holds | diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start index 0835e6793..2f31a08f8 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start @@ -4,10 +4,20 @@ # Bluetooth controller at boot, hang-safe. All six drivers are in the kernel # tree and blacklisted for autoload (modprobe.d/omamobile-tegu-guards.conf), # so this script is the only thing that loads them, in this order, after a -# marker has been written and synced. If the phone hangs in one of the steps, -# the next boot finds the marker, leaves all of it alone and says so; after +# marker has been written and synced. Each step keys on its own device node +# (the GPU on 1f000000.gpu, the display on 19470000.display-controller, the +# rest on their device-tree nodes), so a DTB without the GPU still gets its +# display. A load or bind that returns but fails is logged and skipped, and +# the start goes on to the next step: panthor without mali_csffw.bin must not +# take the display with it. Only a step that never comes back keeps the +# marker: the next boot finds it, leaves all of it alone and says so; after # reading the previous boot's log, run "omamobile-tegu-graphics-start --retry". # +# The panel's DSI commands (gs_dsim_cmd.allow_tx, for brightness, sleep and +# 120 Hz) are allowed only when the bootloader names the google-tg4c panel +# on the command line (exynos_drm.panel_name=google-tg4c, kernel #90), and +# before gs_decon loads, since gs-decon decides on 120 Hz at probe. +# # This is the curated start of the bring-up (device/tegu/bringup) without its # file-hash manifest, EGL smoke test, mask checks, sddm start and bootloader # dead-man: the marker is the safety net, as on the Pixel 7 Pro. @@ -22,8 +32,10 @@ fi state=$root/var/lib/omamobile marker=$state/tegu-graphics-started dt=$root/proc/device-tree +cmdline=$root/proc/cmdline gpu=$root/sys/bus/platform/devices/1f000000.gpu decon=$root/sys/bus/platform/devices/19470000.display-controller +allow_tx=$root/sys/module/gs_dsim_cmd/parameters/allow_tx torch=$root/sys/class/leds/white:torch gauge=$root/sys/class/power_supply/max77779-battery haptics_node=$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43 @@ -47,6 +59,14 @@ bound() { done return 1 } +# The panel the bootloader named on the kernel command line, or nothing. +panel_name() { + local word + for word in $(cat "$cmdline" 2>/dev/null); do + [[ $word == exynos_drm.panel_name=* ]] && { echo "${word#*=}"; return 0; } + done + echo "" +} case ${1:-} in '') ;; @@ -59,8 +79,8 @@ if [[ -e $marker ]]; then log "after checking the previous boot's log, retry with: omamobile-tegu-graphics-start --retry" exit 0 fi -if [[ ! -e $gpu ]]; then - log "no GPU in this DTB; nothing to do" +if [[ ! -e $gpu && ! -e $decon ]]; then + log "no GPU and no display controller in this DTB; nothing to do" exit 0 fi # Once per boot: a loaded driver is never unloaded or retried (AGENTS.md). @@ -71,70 +91,117 @@ if [[ ${1:-} != --retry ]] && { loaded panthor || loaded gs_decon; }; then exit 0 fi -mkdir -p "$state" +mkdir -p "$state" "$root/run" # The marker names the step in progress, so a hang is easy to place. phase() { printf '%s %s\n' "$(date -Is)" "$1" > "$marker"; sync -f "$marker" 2>/dev/null || sync; } finish() { rm -f "$marker"; sync -f "$state" 2>/dev/null || sync; } -fail() { log "$*"; log "the marker stays: the next boot skips this start until --retry"; exit 1; } +# A step that returned and failed: logged, skipped, the start goes on to the +# next one. The marker is cleared at the end as after a good start: it is for +# a step that never comes back, not for one that said no. +up= +failed= +skip() { log "$2; skipped"; failed+="${failed:+, }$1"; } -phase gpu -modprobe panthor || fail "panthor failed" -bound "$gpu" panthor 60 || fail "the GPU did not bind to panthor" +if [[ -e $gpu ]]; then + phase gpu + if ! modprobe panthor; then + skip GPU "panthor failed (no mali_csffw.bin yet? see dmesg)" + elif ! bound "$gpu" panthor 60; then + skip GPU "the GPU did not bind to panthor (see dmesg)" + else + # The session's environment (uwsm/env.d/omamobile-tegu-graphics) reads + # this and picks hardware GL. + printf '{"gpu": "panthor", "by": "omamobile-tegu-graphics-start"}\n' > "$root/run/omamobile-tegu-gpu-ready.json" + up+="GPU" + fi +else + log "no GPU in this DTB; skipped" +fi -# stage=2 is the full driver (the lower stages are bring-up probes); the -# refresh rate and the panel-off behaviour come from modprobe.d. -phase display -modprobe gs_decon stage=2 fbdev=0 || fail "gs_decon failed" -bound "$decon" gs-decon 60 || fail "the display controller did not bind to gs-decon" -# The session's environment (uwsm/env) reads these: the display controller's -# DRM node, whichever card number it got, and that the GPU is up. -mkdir -p "$root/run" -card=$(readlink -f "$root/dev/dri/by-path/platform-19470000.display-controller-card" 2>/dev/null || true) -[[ -n $card ]] && printf '%s\n' "${card#"$root"}" > "$root/run/omamobile-tegu-display-card" -printf '{"gpu": "panthor", "by": "omamobile-tegu-graphics-start"}\n' > "$root/run/omamobile-tegu-gpu-ready.json" -up="GPU, display" +if [[ -e $decon ]]; then + phase display + # gs_dsim_cmd is not guarded: it autoloads from its node ahead of this + # start and sends nothing until allow_tx is on. Loading it here (a no-op + # when it already is) makes the parameter file exist before the write. + modprobe gs_dsim_cmd 2>/dev/null || true + panel=$(panel_name) + if [[ $panel == google-tg4c ]]; then + if echo 1 > "$allow_tx" 2>/dev/null; then + log "panel $panel: DSI commands allowed (gs_dsim_cmd.allow_tx=1)" + else + log "panel $panel: could not write $allow_tx; the display stays read-only at 60 Hz" + fi + else + log "panel ${panel:-unnamed}: DSI commands stay off; the display stays read-only at 60 Hz" + fi + # stage=2 is the full driver (the lower stages are bring-up probes); the + # refresh rate and the panel-off behaviour come from modprobe.d. + if ! modprobe gs_decon stage=2 fbdev=0; then + skip display "gs_decon failed (see dmesg)" + elif ! bound "$decon" gs-decon 60; then + skip display "the display controller did not bind to gs-decon (see dmesg)" + else + # The session's environment reads the display controller's DRM node, + # whichever card number it got (the GPU may take card0). + card=$(readlink -f "$root/dev/dri/by-path/platform-19470000.display-controller-card" 2>/dev/null || true) + [[ -n $card ]] && printf '%s\n' "${card#"$root"}" > "$root/run/omamobile-tegu-display-card" + up+="${up:+, }display" + fi +else + log "no display controller in this DTB; skipped" +fi if enabled "$haptics_node"; then phase haptics - modprobe cs40l26_vibra || fail "cs40l26_vibra failed" - up+=", haptics" + if modprobe cs40l26_vibra; then up+="${up:+, }haptics"; else skip haptics "cs40l26_vibra failed (see dmesg)"; fi else log "no haptics in this DTB; skipped" fi if enabled "$torch_node"; then phase torch - modprobe leds_lm3644 || fail "leds_lm3644 failed" - for _ in $(seq 40); do [[ -e $torch/brightness ]] && break; sleep 0.25; done - if [[ ! -e $torch/brightness ]]; then - log "no white:torch LED after the driver loaded (see dmesg)" - elif [[ $(cat "$torch/brightness") != 0 ]]; then - fail "the torch came up lit; it must start off" + if ! modprobe leds_lm3644; then + skip torch "leds_lm3644 failed (see dmesg)" else - up+=", torch" + for _ in $(seq 40); do [[ -e $torch/brightness ]] && break; sleep 0.25; done + if [[ ! -e $torch/brightness ]]; then + log "no white:torch LED after the driver loaded (see dmesg)" + elif [[ $(cat "$torch/brightness") != 0 ]]; then + # It must start off: turn it off here and say so. + echo 0 > "$torch/brightness" 2>/dev/null + skip torch "the torch came up lit; turned off (brightness now $(cat "$torch/brightness" 2>/dev/null))" + else + up+="${up:+, }torch" + fi fi else log "no torch in this DTB; skipped" fi phase gauge -modprobe max77779_battery || fail "max77779_battery failed" -for _ in $(seq 40); do [[ -e $gauge/capacity ]] && break; sleep 0.25; done -if [[ -e $gauge/capacity ]]; then - up+=", gauge $(cat "$gauge/capacity") %" +if ! modprobe max77779_battery; then + skip gauge "max77779_battery failed (see dmesg)" else - log "no max77779-battery supply after the driver loaded (see dmesg)" + for _ in $(seq 40); do [[ -e $gauge/capacity ]] && break; sleep 0.25; done + if [[ -e $gauge/capacity ]]; then + up+="${up:+, }gauge $(cat "$gauge/capacity") %" + else + log "no max77779-battery supply after the driver loaded (see dmesg)" + fi fi # The address and the power-on are omamobile-bt-addr@'s, started by udev. if enabled "$bluetooth_node"; then phase bluetooth - modprobe hci_uart || fail "hci_uart failed" - up+=", Bluetooth" + if modprobe hci_uart; then up+="${up:+, }Bluetooth"; else skip Bluetooth "hci_uart failed (see dmesg)"; fi else log "no Bluetooth in this DTB; skipped" fi finish -log "up: $up" +log "up: ${up:-nothing}" +if [[ -n $failed ]]; then + log "failed: $failed (these loads returned, so the marker is clear; --retry tries them again)" + exit 1 +fi exit 0 diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all index 12a8c3f13..923499f4f 100755 --- a/packages/omamobile-device-tegu/test/all +++ b/packages/omamobile-device-tegu/test/all @@ -118,6 +118,10 @@ grep -q 'RuntimeWatchdogSec' "$tmp/root/usr/lib/systemd/system.conf.d/10-omamobi && fail "no runtime watchdog from the package" grep -q 'Before=display-manager.service' "$tmp/root/usr/lib/systemd/system/omamobile-tegu-graphics.service" \ || fail "the graphics start runs before the display manager" +for node in 1f000000.gpu 19470000.display-controller; do + grep -qx "ConditionPathExists=|/sys/bus/platform/devices/$node" "$tmp/root/usr/lib/systemd/system/omamobile-tegu-graphics.service" \ + || fail "the graphics start runs when $node exists (ORed condition)" +done grep -q 'After=omamobile-tegu-graphics.service' "$tmp/root/usr/lib/systemd/system/display-manager.service.d/10-omamobile-tegu.conf" \ || fail "the display manager waits for the graphics start" for absent in usr/bin/omamobile-boot-success usr/bin/omamobile-bootctl \ @@ -129,19 +133,27 @@ for absent in usr/bin/omamobile-boot-success usr/bin/omamobile-bootctl \ done pass "install stages the package, enables only the guards, and nothing that marks a slot" -# Every driver the graphics guard loads is kept from autoloading, and nothing -# in modprobe.d would break the guard's own modprobe. -for m in $(grep -oE '^modprobe [a-z0-9_]+' bin/omamobile-tegu-graphics-start | cut -d' ' -f2 | sort -u); do +# Every driver the graphics guard loads is kept from autoloading (gs_dsim_cmd +# aside: it autoloads from its node and sends nothing until allow_tx is on, +# which only the guard sets), and nothing in modprobe.d would break the +# guard's own modprobe. +for m in $(grep -v '^ *#' bin/omamobile-tegu-graphics-start | grep -oE '\bmodprobe [a-z0-9_]+' | cut -d' ' -f2 | sort -u); do + [[ $m == gs_dsim_cmd ]] && continue grep -qx "blacklist $m" vendor/modprobe.d/omamobile-tegu-guards.conf \ || fail "$m is loaded by the graphics guard but not blacklisted" done +for m in panthor gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart; do + grep -qE "modprobe $m\b" bin/omamobile-tegu-graphics-start || fail "the graphics guard no longer loads $m" +done for m in gsa gs201_aoc_sysmmu mailbox_wc aoc_core aoc_channel_dev aoc_alsa_dev aoc_char_dev; do grep -qx "blacklist $m" vendor/modprobe.d/omamobile-tegu-aoc.conf \ || fail "$m is loaded by the AoC guard but not blacklisted" done grep -qE '^install ' vendor/modprobe.d/*.conf && fail "an install line in modprobe.d would break a guard's modprobe" -grep -qE '^options gs_dsim_cmd .*allow_tx=1' vendor/modprobe.d/omamobile-tegu-display.conf \ - || fail "the panel's DSI commands are allowed" +grep -qE '^options gs_dsim_cmd' vendor/modprobe.d/omamobile-tegu-display.conf \ + && fail "allow_tx belongs to the graphics start (google-tg4c only), not modprobe.d" +grep -q 'exynos_drm.panel_name=google-tg4c' bin/omamobile-tegu-graphics-start \ + || fail "the graphics start keys allow_tx on the bootloader's panel name" [[ $(grep -c 'max_refresh=' vendor/modprobe.d/omamobile-tegu-display.conf) == 1 ]] \ || fail "exactly one refresh-rate ceiling" grep -qE '^options clk_acpm zumapro_gpu_max_khz=[0-9]+' vendor/modprobe.d/omamobile-tegu-clk-acpm.conf \ diff --git a/packages/omamobile-device-tegu/test/graphics-start b/packages/omamobile-device-tegu/test/graphics-start index 981c1d15f..a7ff62b58 100755 --- a/packages/omamobile-device-tegu/test/graphics-start +++ b/packages/omamobile-device-tegu/test/graphics-start @@ -1,10 +1,12 @@ #!/bin/bash # SPDX-License-Identifier: MIT # omamobile-tegu-graphics-start in a scratch root: the marker goes down and -# is synced before the first driver loads, the drivers load in order, a -# failure leaves the marker naming the step, and a marker from a previous -# boot keeps every driver off until --retry. No hardware is touched: modprobe, -# sync and sleep are fakes on PATH. +# is synced before the first driver loads, the drivers load in order, a load +# that returns and fails is logged, skipped and clears the marker (the rest +# still loads), each step keys on its own device node, the panel's DSI +# commands are allowed for google-tg4c only and before gs_decon, and a +# marker from a previous boot keeps every driver off until --retry. No +# hardware is touched: modprobe, sync and sleep are fakes on PATH. set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." script=$PWD/bin/omamobile-tegu-graphics-start @@ -17,6 +19,7 @@ marker=$OMAMOBILE_TEGU_ROOT/var/lib/omamobile/tegu-graphics-started export TEST_MARKER=$marker dt=$OMAMOBILE_TEGU_ROOT/proc/device-tree devices=$OMAMOBILE_TEGU_ROOT/sys/bus/platform/devices +allow_tx=$OMAMOBILE_TEGU_ROOT/sys/module/gs_dsim_cmd/parameters/allow_tx mkdir -p "$tmp/bin" "$devices/1f000000.gpu" "$devices/19470000.display-controller" \ "$OMAMOBILE_TEGU_ROOT/sys/bus/platform/drivers/panthor" "$OMAMOBILE_TEGU_ROOT/sys/bus/platform/drivers/gs-decon" \ "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43" \ @@ -25,70 +28,136 @@ mkdir -p "$tmp/bin" "$devices/1f000000.gpu" "$devices/19470000.display-controlle printf 'okay\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" export PATH=$tmp/bin:$PATH # modprobe: the marker must already be on disk; a module "binds" by creating -# what the script waits for; FAIL_MODULE fails that one load. +# what the script waits for; FAIL_MODULE fails that one load. gs_dsim_cmd +# brings its allow_tx parameter (0, the kernel's default), and gs_decon's +# line records what allow_tx read when it loaded. cat > "$tmp/bin/modprobe" <<'EOF' #!/bin/bash set -euo pipefail test -s "$TEST_MARKER" grep -qx synced "$TEST_LOG" -echo "$*" >> "$TEST_LOG" -[[ ${FAIL_MODULE:-} != "$1" ]] || exit 1 +[[ ${FAIL_MODULE:-} != "$1" ]] || { echo "$*" >> "$TEST_LOG"; exit 1; } case $1 in panthor) ln -sfn "$TEST_ROOT/sys/bus/platform/drivers/panthor" "$TEST_ROOT/sys/bus/platform/devices/1f000000.gpu/driver" ;; - gs_decon) ln -sfn "$TEST_ROOT/sys/bus/platform/drivers/gs-decon" "$TEST_ROOT/sys/bus/platform/devices/19470000.display-controller/driver" ;; + gs_dsim_cmd) p=$TEST_ROOT/sys/module/gs_dsim_cmd/parameters; [[ -e $p/allow_tx ]] || { mkdir -p "$p"; echo 0 > "$p/allow_tx"; } ;; + gs_decon) ln -sfn "$TEST_ROOT/sys/bus/platform/drivers/gs-decon" "$TEST_ROOT/sys/bus/platform/devices/19470000.display-controller/driver" + set -- "$@" "allow_tx=$(cat "$TEST_ROOT/sys/module/gs_dsim_cmd/parameters/allow_tx" 2>/dev/null || echo none)" ;; leds_lm3644) mkdir -p "$TEST_ROOT/sys/class/leds/white:torch"; echo "${TORCH_LIT:-0}" > "$TEST_ROOT/sys/class/leds/white:torch/brightness" ;; max77779_battery) mkdir -p "$TEST_ROOT/sys/class/power_supply/max77779-battery"; echo 77 > "$TEST_ROOT/sys/class/power_supply/max77779-battery/capacity" ;; esac +echo "$*" >> "$TEST_LOG" EOF printf '#!/bin/bash\necho synced >> "$TEST_LOG"\n' > "$tmp/bin/sync" printf '#!/bin/bash\nexit 0\n' > "$tmp/bin/sleep" chmod +x "$tmp/bin/"* +# The command line the bootloader built, with its panel name +cmdline() { mkdir -p "$OMAMOBILE_TEGU_ROOT/proc"; echo "console=null panic=10 ${1:+exynos_drm.panel_name=$1 }mem_sleep_default=s2idle" > "$OMAMOBILE_TEGU_ROOT/proc/cmdline"; } reset() { : > "$TEST_LOG" rm -rf "$marker" "$devices/1f000000.gpu/driver" "$devices/19470000.display-controller/driver" \ - "$OMAMOBILE_TEGU_ROOT/sys/class" "$OMAMOBILE_TEGU_ROOT/sys/module" + "$OMAMOBILE_TEGU_ROOT/sys/class" "$OMAMOBILE_TEGU_ROOT/sys/module" "$OMAMOBILE_TEGU_ROOT/run" + cmdline google-tg4c } loads() { grep -v '^synced$' "$TEST_LOG" | cut -d' ' -f1 | tr '\n' ' ' | sed 's/ $//'; } +all="panthor gs_dsim_cmd gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" -# A full start: the order, the gs_decon options, the marker gone, the summary +# A full start: the order, the gs_decon options, the marker gone, the summary, +# the session's two files reset out=$("$script") [[ ! -e $marker ]] || fail "a finished start leaves the marker" -[[ $(loads) == "panthor gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] \ - || fail "load order: $(loads)" -grep -q '^gs_decon stage=2 fbdev=0$' "$TEST_LOG" || fail "gs_decon gets the full driver stage" +[[ $(loads) == "$all" ]] || fail "load order: $(loads)" +grep -q '^gs_decon stage=2 fbdev=0 ' "$TEST_LOG" || fail "gs_decon gets the full driver stage" [[ $out == *"up: GPU, display, haptics, torch, gauge 77 %, Bluetooth"* ]] || fail "summary: $out" +[[ -s $OMAMOBILE_TEGU_ROOT/run/omamobile-tegu-gpu-ready.json ]] || fail "no gpu-ready file for the session" + +# google-tg4c on the command line: allow_tx reads 1 when gs_decon loads +grep -q '^gs_decon .*allow_tx=1$' "$TEST_LOG" || fail "allow_tx was not 1 before gs_decon loaded: $(grep '^gs_decon' "$TEST_LOG")" +[[ $(cat "$allow_tx") == 1 ]] || fail "allow_tx was not written" +[[ $out == *"panel google-tg4c: DSI commands allowed"* ]] || fail "the allow is not logged: $out" + +# Another panel: allow_tx stays 0, gs_decon still loads +reset; cmdline google-xyz9 +out=$("$script") +grep -q '^gs_decon .*allow_tx=0$' "$TEST_LOG" || fail "another panel got allow_tx: $(grep '^gs_decon' "$TEST_LOG")" +[[ $(cat "$allow_tx") == 0 ]] || fail "allow_tx was touched for another panel" +[[ $out == *"panel google-xyz9: DSI commands stay off"* ]] || fail "the other panel is not named: $out" +[[ $(loads) == "$all" ]] || fail "another panel changed the loads: $(loads)" + +# No panel name at all: the same, and said so +reset; cmdline "" +out=$("$script") +grep -q '^gs_decon .*allow_tx=0$' "$TEST_LOG" || fail "an unnamed panel got allow_tx" +[[ $out == *"panel unnamed: DSI commands stay off"* ]] || fail "an unnamed panel is not reported: $out" + +# A failed display load: exit 1, logged and skipped, the rest still loads, +# the marker is cleared (the load returned, no hang) +reset +out=$(FAIL_MODULE=gs_decon "$script") && fail "a failed load exits 0" +[[ ! -e $marker ]] || fail "a failed load kept the marker: $(cat "$marker")" +[[ $(loads) == "$all" ]] || fail "a failed load stopped the rest: $(loads)" +[[ $out == *"gs_decon failed"*"skipped"* && $out == *"failed: display"* ]] || fail "the failed step is not named: $out" +[[ $out == *"up: GPU, haptics, torch, gauge 77 %, Bluetooth"* ]] || fail "summary after a failed display: $out" +[[ ! -e $OMAMOBILE_TEGU_ROOT/run/omamobile-tegu-display-card ]] || fail "a display that did not come up left a card file" + +# panthor fails (no mali_csffw.bin yet): the display still comes up, and the +# session gets no gpu-ready file +reset +out=$(FAIL_MODULE=panthor "$script") && fail "a failed GPU exits 0" +[[ ! -e $marker ]] || fail "a failed GPU kept the marker" +[[ $(loads) == "$all" ]] || fail "a failed GPU took the display with it: $(loads)" +[[ $out == *"panthor failed"* && $out == *"up: display, haptics, torch, gauge 77 %, Bluetooth"* ]] || fail "panthor fails: $out" +[[ ! -e $OMAMOBILE_TEGU_ROOT/run/omamobile-tegu-gpu-ready.json ]] || fail "a GPU that did not come up was reported ready" + +# No GPU in the DTB: the display and the rest still come up, the unit's own +# condition aside +reset +mv "$devices/1f000000.gpu" "$tmp/gpu" +out=$("$script") || fail "no GPU node is not a failure" +[[ $(loads) == "gs_dsim_cmd gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] || fail "no GPU node: $(loads)" +[[ ! -e $marker && $out == *"no GPU in this DTB"* && $out == *"up: display, haptics"* ]] || fail "no GPU node: $out" +mv "$tmp/gpu" "$devices/1f000000.gpu" -# A failed display load: exit 1, the marker names the step, nothing after it +# No display controller in the DTB: the GPU and the rest still come up reset -FAIL_MODULE=gs_decon "$script" >/dev/null && fail "a failed load exits 0" -[[ -s $marker ]] || fail "a failed load removes the marker" -grep -q ' display$' "$marker" || fail "the marker names the failed step: $(cat "$marker")" -[[ $(loads) == "panthor gs_decon" ]] || fail "a failed load went on: $(loads)" +mv "$devices/19470000.display-controller" "$tmp/decon" +out=$("$script") || fail "no display node is not a failure" +[[ $(loads) == "panthor cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] || fail "no display node: $(loads)" +[[ $out == *"no display controller in this DTB"* ]] || fail "no display node: $out" +mv "$tmp/decon" "$devices/19470000.display-controller" + +# Neither: nothing to do +reset +mv "$devices/1f000000.gpu" "$tmp/gpu"; mv "$devices/19470000.display-controller" "$tmp/decon" +out=$("$script") +[[ -z $(loads) && ! -e $marker && $out == *"nothing to do"* ]] || fail "no GPU and no display were not skipped" +mv "$tmp/gpu" "$devices/1f000000.gpu"; mv "$tmp/decon" "$devices/19470000.display-controller" -# The next boot with that marker: nothing loads, the hint is printed -: > "$TEST_LOG" -rm -rf "$devices/1f000000.gpu/driver" "$devices/19470000.display-controller/driver" +# A marker left by a boot that hung: nothing loads, the hint is printed +reset +printf '%s display\n' "$(date -Is)" > "$marker" out=$("$script") [[ -s $marker && -z $(loads) ]] || fail "a leftover marker did not keep the drivers off" [[ $out == *"--retry"* ]] || fail "no retry hint" # --retry clears it and starts again out=$("$script" --retry) -[[ ! -e $marker && $(loads) == "panthor gs_decon cs40l26_vibra leds_lm3644 max77779_battery hci_uart" ]] \ - || fail "--retry did not start again: $(loads)" +[[ ! -e $marker && $(loads) == "$all" ]] || fail "--retry did not start again: $(loads)" -# A torch that comes up lit is a failure that keeps the marker +# A torch that comes up lit is turned off and counted as failed; the rest +# still loads reset -TORCH_LIT=1 "$script" >/dev/null && fail "a lit torch passed" -grep -q ' torch$' "$marker" || fail "the marker names the torch step" +out=$(TORCH_LIT=1 "$script") && fail "a lit torch passed" +[[ $(cat "$OMAMOBILE_TEGU_ROOT/sys/class/leds/white:torch/brightness") == 0 ]] || fail "a lit torch was left lit" +[[ $out == *"came up lit"* && $out == *"failed: torch"* && $(loads) == "$all" ]] || fail "lit torch: $out" +[[ ! -e $marker ]] || fail "a lit torch kept the marker" # A disabled haptics node is skipped, a missing Bluetooth node too reset printf 'disabled\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" mv "$dt/soc/usi@155d00c0" "$tmp/bt-node" out=$("$script") -[[ $(loads) == "panthor gs_decon leds_lm3644 max77779_battery" ]] || fail "disabled nodes were loaded: $(loads)" +[[ $(loads) == "panthor gs_dsim_cmd gs_decon leds_lm3644 max77779_battery" ]] || fail "disabled nodes were loaded: $(loads)" [[ $out == *"no haptics in this DTB"* && $out == *"no Bluetooth in this DTB"* ]] || fail "skips are not reported" printf 'okay\0' > "$dt/soc/usi@10c900c0/i2c@10c90000/haptic-driver@43/status" mv "$tmp/bt-node" "$dt/soc/usi@155d00c0" @@ -99,13 +168,6 @@ mkdir -p "$OMAMOBILE_TEGU_ROOT/sys/module/panthor" out=$("$script") [[ -z $(loads) && ! -e $marker && $out == *"already started"* ]] || fail "a loaded driver was retried" -# No GPU in the DTB: nothing to do -reset -mv "$devices/1f000000.gpu" "$tmp/gpu" -out=$("$script") -[[ -z $(loads) && ! -e $marker && $out == *"no GPU"* ]] || fail "a missing GPU node was not skipped" -mv "$tmp/gpu" "$devices/1f000000.gpu" - # A bad argument "$script" --bogus >/dev/null 2>&1 && fail "a bad argument is accepted" -echo 'ok - graphics-start: marker first, order kept, failures named, --retry works' +echo 'ok - graphics-start: marker first, order kept, failures skipped and named, allow_tx for google-tg4c only, --retry works' diff --git a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf index 0ad529694..083e7232a 100644 --- a/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf +++ b/packages/omamobile-device-tegu/vendor/modprobe.d/omamobile-tegu-display.conf @@ -1,6 +1,8 @@ # SPDX-License-Identifier: MIT -# The panel's brightness, sleep and refresh rate go over DSI commands. -options gs_dsim_cmd allow_tx=1 +# The panel's DSI commands (brightness, sleep, the refresh rate) are not +# allowed here: omamobile-tegu-graphics-start turns gs_dsim_cmd.allow_tx on +# only when the bootloader names the google-tg4c panel (kernel #90), before +# gs_decon loads. On any other panel the 120 Hz ceiling falls back to 60 Hz. # 120 Hz (kernel #88) and the panel really off on DPMS, as the owner's phone # runs it. omamobile-tegu-graphics-start adds stage=2 (the full driver) and # fbdev=0 on its modprobe line. diff --git a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service index a8a3ada55..b14a5d188 100644 --- a/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service +++ b/packages/omamobile-device-tegu/vendor/systemd/system/omamobile-tegu-graphics.service @@ -4,12 +4,15 @@ # blacklisted for autoload (modprobe.d/omamobile-tegu-guards.conf), so this # hang-safe script is the only thing that loads them; if a start hangs, the # next boot leaves them off (see the script). Runs before the display -# manager (display-manager.service.d/10-omamobile-tegu.conf). +# manager (display-manager.service.d/10-omamobile-tegu.conf). The two +# ConditionPathExists=| lines are ORed by systemd: a DTB with the display +# but no GPU still gets its display. [Unit] Description=Bring up the Pixel 9a GPU, display, haptics, torch, gauge and Bluetooth (hang-safe) After=local-fs.target systemd-modules-load.service Before=display-manager.service -ConditionPathExists=/sys/bus/platform/devices/1f000000.gpu +ConditionPathExists=|/sys/bus/platform/devices/1f000000.gpu +ConditionPathExists=|/sys/bus/platform/devices/19470000.display-controller [Service] Type=oneshot From 0ce11847bb88d91362697a1da254cdc4d678f7ab Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:18 -0647 Subject: [PATCH 15/20] omamobile-device-tegu: the NFC sleep hook stops polling; the power-off step behind a flag file The hook stopped neard's poll loop and then powered the adapter off before every sleep. The variant with the record is the stop-poll one: polling through s2idle reset the palmtop (2026-10-27), and with the poll loop stopped first 35 of 35 cycles woke that day, while the power-off variant has one wake that needed the power key and one that worked. So the hook now stops the poll loop, reads Polling back and logs it, and starts the loop again after the wake if it was on. Powering the adapter off, the workaround for a controller that did not answer on I2C after a wake, runs only while /etc/omamobile/nfc-power-off-across-sleep exists (off by default), and then the adapter is powered on again after the wake. The header points to issue #132 for the symptom and linux-omarchy-mobile #131 for the fix. Tests: the default stops and restarts polling with the adapter left powered and both readings logged, an idle adapter is logged and not touched, and the flag file brings the power-off and the power-on back. --- packages/omamobile-device-tegu/README.md | 2 +- .../omamobile-device-tegu/test/sleep-hooks | 43 ++++++++++++++----- .../systemd/system-sleep/omamobile-tegu-nfc | 42 ++++++++++++------ 3 files changed, 62 insertions(+), 25 deletions(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index e6c36f73a..7978f9577 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -18,7 +18,7 @@ packaged yet. What is here is what runs on in-tree drivers. | `bin/omamobile-tegu-usb-rescue` + `src/tegu-tcpc-switch.c` + `vendor/systemd/system/omamobile-tegu-usb-rescue.service` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Closes the MAX77779's USB data switch after a cable replug while no Type-C driver owns the chip; with one loaded it exits. The helper reads the chip's identity and VBUS before its one write and never forces an address a driver holds | | `bin/omamobile-tegu-charge-limit` + `vendor/udev/rules.d/75-omamobile-tegu-charger.rules` | `/usr/bin`, `/usr/lib/udev/rules.d` | Applies `/etc/omamobile/charge-limit` (a percentage, say `80`) to the charger when it appears; charging stops there and starts again 5 below. No file: charge to 100 % | | `vendor/systemd/system-sleep/omamobile-tegu-wifi` | `/usr/lib/systemd/system-sleep` | Keeps Wi-Fi loaded across s2idle when `/etc/omamobile/wifi-keep-across-sleep` exists (the chip reloads its firmware on resume and iwd reconnects), otherwise unloads and reloads brcmfmac around the sleep; after a sleep that lost Wi-Fi it unloads for the rest of the boot. It reloads the in-tree `brcmfmac`; a phone still running an out-of-tree build of it keeps its own hook until the Wi-Fi series lands | -| `vendor/systemd/system-sleep/omamobile-tegu-nfc` | `/usr/lib/systemd/system-sleep` | Stops neard's poll loop and powers the ST54L off before a sleep (it does not answer on I2C after a wake otherwise), and brings it back afterwards | +| `vendor/systemd/system-sleep/omamobile-tegu-nfc` | `/usr/lib/systemd/system-sleep` | Stops neard's poll loop before a sleep (polling through s2idle reset the palmtop; with it stopped, 35 of 35 cycles woke), reads `Polling` back and logs it, and starts the loop again after the wake if it was on. Powering the ST54L off as well, the workaround for a controller that did not answer on I2C after a wake (issue #132; fixed by linux-omarchy-mobile #131), runs only while `/etc/omamobile/nfc-power-off-across-sleep` exists: off by default, since its record is one wake that needed the power key and one that worked. Log: `/var/log/omamobile/nfc-sleep.log` | | `vendor/systemd/system/linux-modules-cleanup.service.d/10-omamobile-tegu.conf` | `/usr/lib/systemd/system` | Holds off Arch's old-module cleanup: our kernels come as module tarballs, not pacman packages, and the next-boot and rollback trees must stay | | `vendor/systemd/system.conf.d/10-omamobile-reboot-watchdog.conf`, `sleep.conf.d/10-omamobile.conf` | `/usr/lib/systemd` | A hung shutdown resets the phone after a minute; suspend means s2idle (no SoC sleep sequence yet, and `deep` does not wake) | | `vendor/udev/rules.d/10-omamobile-storage.rules` | `/usr/lib/udev/rules.d` | Continues the initramfs storage guard: Android partitions stay hidden and the whole-disk node unusable | diff --git a/packages/omamobile-device-tegu/test/sleep-hooks b/packages/omamobile-device-tegu/test/sleep-hooks index e129cba80..11dd15a38 100755 --- a/packages/omamobile-device-tegu/test/sleep-hooks +++ b/packages/omamobile-device-tegu/test/sleep-hooks @@ -2,8 +2,9 @@ # SPDX-License-Identifier: MIT # The sleep hooks in a scratch root with fake lsmod, modprobe, rmmod, # systemctl, logger and busctl: Wi-Fi is left alone when kept and present, -# unloaded and reloaded otherwise, and falls back after a lost sleep; NFC is -# stopped and powered off before the sleep and brought back after it. +# unloaded and reloaded otherwise, and falls back after a lost sleep; NFC's +# poll loop is stopped before the sleep and started again after it, and the +# adapter is powered off only behind the flag file. set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." wifi=$PWD/vendor/systemd/system-sleep/omamobile-tegu-wifi @@ -33,7 +34,8 @@ case $1 in [[ $5 == StopPollLoop ]] && rm -f "$TEST_NFC/Polling" [[ $5 == StartPollLoop ]] && : > "$TEST_NFC/Polling" ;; set-property) echo "busctl set $5 $7" >> "$TEST_LOG" - [[ $7 == true ]] && : > "$TEST_NFC/Powered" || rm -f "$TEST_NFC/Powered" ;; + [[ $7 == true ]] && : > "$TEST_NFC/Powered" || rm -f "$TEST_NFC/Powered" + [[ $7 == true ]] || rm -f "$TEST_NFC/Polling" ;; esac exit 0 EOF @@ -41,6 +43,8 @@ chmod +x "$tmp/bin/"* calls() { tr '\n' ';' < "$TEST_LOG" | sed 's/;$//'; } run=$OMAMOBILE_TEGU_ROOT/run/omamobile keep=$OMAMOBILE_TEGU_ROOT/etc/omamobile/wifi-keep-across-sleep +power_off=$OMAMOBILE_TEGU_ROOT/etc/omamobile/nfc-power-off-across-sleep +nfc_log=$OMAMOBILE_TEGU_ROOT/var/log/omamobile/nfc-sleep.log iface() { mkdir -p "$OMAMOBILE_TEGU_ROOT/sys/bus/pci/drivers/brcmfmac/0000:01:00.0" "$OMAMOBILE_TEGU_ROOT/sys/class/ieee80211/phy0"; } no_iface() { rm -rf "$OMAMOBILE_TEGU_ROOT/sys"; } @@ -73,18 +77,37 @@ no_iface() { rm -rf "$OMAMOBILE_TEGU_ROOT/sys"; } "$wifi" pre [[ $(calls) == "systemctl stop iwd;rmmod brcmfmac_wcc brcmfmac" ]] || fail "the fallback flag is not honoured: $(calls)" -# NFC powered and polling: stopped and powered off before, back after -: > "$TEST_LOG"; : > "$TEST_NFC/Polling"; : > "$TEST_NFC/Powered" +# NFC powered and polling, no flag file: the poll loop is stopped before and +# started again after; the adapter stays powered; both readings are logged +: > "$TEST_LOG"; rm -f "$power_off" "$nfc_log"; : > "$TEST_NFC/Polling"; : > "$TEST_NFC/Powered" "$nfc" pre -[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false" ]] || fail "NFC pre: $(calls)" -[[ -e $run/nfc-was-polling && -e $run/nfc-was-powered ]] || fail "NFC pre left no notes" +[[ $(calls) == "busctl call StopPollLoop" ]] || fail "NFC pre: $(calls)" +[[ -e $run/nfc-was-polling && ! -e $run/nfc-was-powered ]] || fail "NFC pre notes: polling only" +[[ -e $TEST_NFC/Powered ]] || fail "NFC pre powered the adapter off without the flag" "$nfc" post -[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false;busctl set Powered true;busctl call StartPollLoop" ]] || fail "NFC post: $(calls)" +[[ $(calls) == "busctl call StopPollLoop;busctl call StartPollLoop" ]] || fail "NFC post: $(calls)" [[ ! -e $run/nfc-was-polling && ! -e $run/nfc-was-powered ]] || fail "NFC post left its notes" -grep -q '^.* pre: polling=b false powered=b false$' "$OMAMOBILE_TEGU_ROOT/var/log/omamobile/nfc-sleep.log" || fail "NFC pre is not logged" +grep -q ' pre: polling=b false powered=b true$' "$nfc_log" || fail "NFC pre is not logged: $(cat "$nfc_log")" +grep -q ' post: polling=b true powered=b true$' "$nfc_log" || fail "NFC post is not logged: $(cat "$nfc_log")" + +# Powered but not polling: nothing to stop, nothing to restart, still logged +: > "$TEST_LOG"; rm -f "$nfc_log" "$TEST_NFC/Polling" +"$nfc" pre; "$nfc" post +[[ -z $(calls) ]] || fail "an idle NFC adapter was touched: $(calls)" +grep -q ' pre: polling=b false powered=b true$' "$nfc_log" || fail "an idle adapter is not logged" + +# With the flag file: stopped and powered off before, powered on and polling after +: > "$TEST_LOG"; : > "$power_off"; : > "$TEST_NFC/Polling" +"$nfc" pre +[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false" ]] || fail "NFC pre with the flag: $(calls)" +[[ -e $run/nfc-was-polling && -e $run/nfc-was-powered ]] || fail "NFC pre with the flag left no notes" +"$nfc" post +[[ $(calls) == "busctl call StopPollLoop;busctl set Powered false;busctl set Powered true;busctl call StartPollLoop" ]] || fail "NFC post with the flag: $(calls)" +[[ -e $TEST_NFC/Powered && -e $TEST_NFC/Polling ]] || fail "NFC post with the flag did not bring the adapter back" +rm -f "$power_off" # NFC off (no neard): nothing is called : > "$TEST_LOG"; rm -f "$TEST_NFC"/* "$nfc" pre; "$nfc" post [[ -z $(calls) ]] || fail "an off NFC adapter was touched: $(calls)" -echo 'ok - sleep-hooks: Wi-Fi kept or reloaded as flagged, NFC stopped and restored' +echo 'ok - sleep-hooks: Wi-Fi kept or reloaded as flagged, NFC polling stopped and restored, power-off behind the flag' diff --git a/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc index f05ac8b83..7f8705cd9 100755 --- a/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc +++ b/packages/omamobile-device-tegu/vendor/systemd/system-sleep/omamobile-tegu-nfc @@ -1,17 +1,27 @@ #!/bin/bash # SPDX-License-Identifier: MIT -# systemd-sleep hook for the Pixel 9a's NFC (ST54L on nxp-nci): after an -# s2idle wake the controller doesn't answer on I2C until the adapter is -# powered off and on (kernel #132). Before the sleep: stop the poll loop, -# power the adapter off and check that Powered reads false (the kernel -# refuses a device-down while polling, EBUSY). If it doesn't, leave it with -# polling stopped. After the wake: power it on, check, and restart the poll -# loop if it was polling. Logs every step. Nothing happens without neard. +# systemd-sleep hook for the Pixel 9a's NFC (ST54L on nxp-nci). Polling +# through s2idle reset the palmtop (2026-10-27); with the poll loop stopped +# first, every cycle since has woken (35 of 35 that day). So before the +# sleep: stop neard's poll loop, read Polling back and log it. After the +# wake: start the poll loop again if it was polling. Nothing happens +# without neard, or with the adapter off. +# +# Powering the adapter off as well (the kernel refuses a device-down while +# polling, EBUSY, so the poll loop stops first) was the workaround for a +# controller that didn't answer on I2C after a wake: the symptom is +# https://github.com/derekross/omarchy-mobile/issues/132, the fix +# https://github.com/derekross/linux-omarchy-mobile/pull/131. That step is +# off by default: it runs only while /etc/omamobile/nfc-power-off-across-sleep +# exists (its record is one wake that needed the power key and one that +# worked), and then the adapter is powered on again after the wake. +# Logs every step to /var/log/omamobile/nfc-sleep.log. # OMAMOBILE_TEGU_ROOT: a scratch root for the package tests. root=${OMAMOBILE_TEGU_ROOT:-} A=/org/neard/nfc0 D=$root/run/omamobile L=$root/var/log/omamobile/nfc-sleep.log +POWER_OFF=$root/etc/omamobile/nfc-power-off-across-sleep mkdir -p "$D" "${L%/*}" log() { echo "$(date -Is) $1: $2" >> "$L"; } prop() { busctl get-property org.neard $A org.neard.Adapter "$1" 2>/dev/null; } @@ -19,18 +29,22 @@ is_true() { prop "$1" | grep -c true >/dev/null; } case "$1" in pre) rm -f "$D/nfc-was-polling" "$D/nfc-was-powered" - is_true Polling && : > "$D/nfc-was-polling" - if is_true Powered; then - : > "$D/nfc-was-powered" + if ! is_true Powered; then + log pre "adapter off or no neard; nothing to do" + exit 0 + fi + if is_true Polling; then + : > "$D/nfc-was-polling" out=$(busctl call org.neard $A org.neard.Adapter StopPollLoop 2>&1) || log pre "StopPollLoop: $out" for i in 1 2 3 4 5; do is_true Polling || break; sleep 0.2; done + fi + if [[ -e $POWER_OFF ]]; then + : > "$D/nfc-was-powered" out=$(busctl set-property org.neard $A org.neard.Adapter Powered b false 2>&1) || log pre "Powered false: $out" for i in 1 2 3 4 5; do is_true Powered || break; sleep 0.2; done - log pre "polling=$(prop Polling) powered=$(prop Powered)" - sync - else - log pre "adapter already off" fi + log pre "polling=$(prop Polling) powered=$(prop Powered)" + sync ;; post) if [[ -e $D/nfc-was-powered ]]; then From 6be4150751440fa1598a7061892e063bcc213fdb Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:18 -0647 Subject: [PATCH 16/20] omamobile-device-tegu: the USB rescue reads the power register first and logs only when it acts With the cable out the gadget is not configured and the switch reads 0x00, as the MAX77779 leaves it on unplug, so the loop called connect every grace period: the helper refused without VBUS, and both lines went to the journal, about 20,000 a day on battery. The loop now reads the power register first and reads the switch and writes only while VBUS is present, so a day on battery costs one register read per grace period and no journal lines. Tests: the cable out costs one power read and no line beyond the watch line; a closed switch is read and not logged. --- .../bin/omamobile-tegu-usb-rescue | 27 ++++++++++++------- .../omamobile-device-tegu/test/usb-rescue | 24 ++++++++++++----- 2 files changed, 34 insertions(+), 17 deletions(-) diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue b/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue index 71a8c29a9..ad5f3aea8 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-usb-rescue @@ -3,14 +3,17 @@ # Pixel 9a: close the MAX77779's USB data switch after a cable replug. The # chip opens USBSW_CTRL (0x93) on unplug and leaves it open when VBUS # returns, so the phone never re-enumerates (device/tegu/NOTES.md 2026-10-26). -# While the gadget is not configured and the switch is open, tegu-tcpc-switch -# connect (identity and VBUS checked, I2C_SLAVE never forced) writes 0x09 -# back. A Type-C driver that owns the chip (tcpci_maxim, since kernel #124) -# closes the switch itself: then this exits and leaves the chip to it. +# While the gadget is not configured, VBUS is present and the switch is open, +# tegu-tcpc-switch connect (identity and VBUS checked, I2C_SLAVE never +# forced) writes 0x09 back. With the cable out (no VBUS) the switch is open +# too, and the helper would refuse: the power register is read first, so a +# day on battery costs I2C reads, not journal lines. A Type-C driver that +# owns the chip (tcpci_maxim, since kernel #124) closes the switch itself: +# then this exits and leaves the chip to it. # omamobile-tegu-usb-rescue [--once] (--once: one pass, for the tests) # OMAMOBILE_TEGU_ROOT: a scratch root for the package tests; # OMAMOBILE_TEGU_RESCUE_GRACE: seconds the gadget may stay unconfigured -# before the switch is read (default 4). +# before the chip is read (default 4; then once per grace period). set -uo pipefail root=${OMAMOBILE_TEGU_ROOT:-} @@ -26,6 +29,8 @@ log() { echo "omamobile-tegu-usb-rescue: $*"; } [[ -x $tool ]] || { log "no $tool"; exit 0; } tcpm_owns() { local d; for d in "$root"/sys/bus/i2c/devices/*-0025; do [[ -e $d/driver ]] && return 0; done; return 1; } gadget_state() { cat "$root"/sys/class/udc/*/state 2>/dev/null | head -1; } +# TCPC_POWER_STATUS (0x1e) bit 2: VBUS present +vbus_present() { local p; p=$("$tool" power 2>/dev/null | tail -1); [[ -n $p ]] && (( (p & 4) != 0 )); } open_since=0 log "watching the gadget state and the TCPC data switch" @@ -41,11 +46,13 @@ while :; do now=$(date +%s) (( open_since == 0 )) && open_since=$now if (( now - open_since >= grace )); then - sw=$("$tool" switch 2>/dev/null | tail -1) - if [[ -n $sw && $sw != 0x09 ]]; then - log "gadget ${state:-absent}, USBSW_CTRL $sw: closing the data switch" - "$tool" connect 2>&1 | tail -1 - open_since=$now + open_since=$now + if vbus_present; then + sw=$("$tool" switch 2>/dev/null | tail -1) + if [[ -n $sw && $sw != 0x09 ]]; then + log "gadget ${state:-absent}, VBUS present, USBSW_CTRL $sw: closing the data switch" + "$tool" connect 2>&1 | tail -1 + fi fi fi fi diff --git a/packages/omamobile-device-tegu/test/usb-rescue b/packages/omamobile-device-tegu/test/usb-rescue index 4070a79d0..377d7eaa3 100755 --- a/packages/omamobile-device-tegu/test/usb-rescue +++ b/packages/omamobile-device-tegu/test/usb-rescue @@ -2,8 +2,9 @@ # SPDX-License-Identifier: MIT # omamobile-tegu-usb-rescue in a scratch root, one pass at a time: a # configured gadget is left alone, an open switch on an unconfigured gadget -# is closed once, a closed switch is not written, and a Type-C driver that -# owns the chip ends the watch before the chip is read. +# with VBUS is closed once, a closed switch is not written, the cable out +# (no VBUS) costs one register read and no log line, and a Type-C driver +# that owns the chip ends the watch before the chip is read. set -euo pipefail cd -- "$(dirname -- "${BASH_SOURCE[0]}")/.." script=$PWD/bin/omamobile-tegu-usb-rescue @@ -15,10 +16,11 @@ export TEST_LOG=$tmp/calls export TEGU_TCPC_SWITCH=$tmp/bin/tegu-tcpc-switch udc=$OMAMOBILE_TEGU_ROOT/sys/class/udc/11110000.dwc3 mkdir -p "$tmp/bin" "$udc" "$OMAMOBILE_TEGU_ROOT/sys/bus/i2c/devices/21-0025" +# power: TCPC_POWER_STATUS, bit 2 VBUS present (0x04 unless PW says otherwise) cat > "$tmp/bin/tegu-tcpc-switch" <<'EOF' #!/bin/bash echo "$1" >> "$TEST_LOG" -case $1 in switch) echo "${SW:-0x09}" ;; connect) echo 0x09 ;; esac +case $1 in power) echo "${PW:-0x04}" ;; switch) echo "${SW:-0x09}" ;; connect) echo 0x09 ;; esac EOF chmod +x "$tmp/bin/tegu-tcpc-switch" calls() { tr '\n' ' ' < "$TEST_LOG" | sed 's/ $//'; } @@ -29,12 +31,20 @@ SW=0x00 "$script" --once >/dev/null : > "$TEST_LOG"; echo 'not attached' > "$udc/state" out=$(SW=0x00 "$script" --once) -[[ $(calls) == "switch connect" ]] || fail "an open switch was not closed once: $(calls)" +[[ $(calls) == "power switch connect" ]] || fail "an open switch was not closed once: $(calls)" [[ $out == *"closing the data switch"* ]] || fail "the close is not logged" : > "$TEST_LOG" -SW=0x09 "$script" --once >/dev/null -[[ $(calls) == "switch" ]] || fail "a closed switch was written: $(calls)" +out=$(SW=0x09 "$script" --once) +[[ $(calls) == "power switch" ]] || fail "a closed switch was written: $(calls)" +[[ $out != *"closing"* ]] || fail "a closed switch was logged as closed again" + +# The cable out: the switch reads open (as the chip leaves it), but with no +# VBUS nothing is written and nothing is logged beyond the watch line +: > "$TEST_LOG" +out=$(PW=0x00 SW=0x00 "$script" --once) +[[ $(calls) == "power" ]] || fail "the cable out still read or wrote the switch: $(calls)" +[[ $(printf '%s\n' "$out" | grep -c .) == 1 && $out == *"watching"* ]] || fail "the cable out wrote to the journal: $out" : > "$TEST_LOG"; ln -s /dev/null "$OMAMOBILE_TEGU_ROOT/sys/bus/i2c/devices/21-0025/driver" out=$(SW=0x00 "$script" --once) @@ -42,4 +52,4 @@ out=$(SW=0x00 "$script" --once) rm -f "$TEGU_TCPC_SWITCH" "$script" --once >/dev/null || fail "a missing helper is not fatal" -echo 'ok - usb-rescue: closes an open switch once, defers to a Type-C driver' +echo 'ok - usb-rescue: closes an open switch once with VBUS, quiet with the cable out, defers to a Type-C driver' From a54cdf2467a0907a43b7f92edcec0b2effffb40f Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:19 -0647 Subject: [PATCH 17/20] omamobile-device-tegu: ship the session's reader of the display card and gpu-ready files The graphics start writes /run/omamobile-tegu-display-card and /run/omamobile-tegu-gpu-ready.json for the session, but their reader lived in the palmtop's own ~/.config/uwsm/env, so a second owner would have hit the card0 failure. The package now ships the reading lines as /usr/share/uwsm/env.d/omamobile-tegu-graphics, which uwsm sources (its env loader walks XDG_DATA_DIRS too, with a POSIX shell; a user's ~/.config/uwsm/env still runs after it): AQ_DRM_DEVICES from the card file when it names a character device, hardware GL when the gpu-ready file exists and the render node is there, llvmpipe otherwise, so a phone whose GPU did not come up still gets a desktop. Tests: the file parses as POSIX shell, is staged, and reads both files. --- packages/omamobile-device-tegu/README.md | 1 + packages/omamobile-device-tegu/install | 3 ++ packages/omamobile-device-tegu/test/all | 6 ++++ .../uwsm/env.d/omamobile-tegu-graphics | 29 +++++++++++++++++++ 4 files changed, 39 insertions(+) create mode 100644 packages/omamobile-device-tegu/uwsm/env.d/omamobile-tegu-graphics diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index 7978f9577..9b0df2701 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -11,6 +11,7 @@ packaged yet. What is here is what runs on in-tree drivers. | Path | Installs to | What it does | | --- | --- | --- | | `bin/omamobile-tegu-graphics-start` + `vendor/systemd/system/omamobile-tegu-graphics.service` + `display-manager.service.d/10-omamobile-tegu.conf` | `/usr/bin`, `/usr/lib/systemd/system` (enabled by a vendor wants link) | Loads the GPU (panthor), the display controller (gs_decon), haptics (cs40l26_vibra), the torch (leds_lm3644), the battery gauge (max77779_battery) and the Bluetooth controller (hci_uart) at boot, in that order, after writing and syncing a marker. Each step keys on its own device node (`1f000000.gpu`, `19470000.display-controller`, the device-tree nodes of the rest), and the unit runs when the GPU or the display controller exists, so a DTB with the display but no GPU still gets its display. A load or bind that returns but fails (panthor without `mali_csffw.bin`, say) is logged and skipped, the rest still loads, and the marker is cleared; the unit then ends failed, and `--retry` tries the skipped steps again. Only a boot that hangs in a step leaves the marker: the next boot keeps all six off and says so, and `omamobile-tegu-graphics-start --retry` goes on from the step that failed (on the same boot too: the drivers already loaded stay as they are). Before `gs_decon`, and only when the command line says `exynos_drm.panel_name=google-tg4c`, it turns `gs_dsim_cmd.allow_tx` on, so the panel's DSI commands (brightness, sleep, 120 Hz; kernel #90) go to the panel the bring-up tested and no other: on another panel the display stays read-only at 60 Hz. The display manager waits for it. It also writes `/run/omamobile-tegu-display-card` (the display controller's DRM node by path, since the GPU may take card0) and, when the GPU came up, `/run/omamobile-tegu-gpu-ready.json` | +| `uwsm/env.d/omamobile-tegu-graphics` | `/usr/share/uwsm/env.d` | The session's side of the graphics start, sourced by uwsm (it walks `XDG_DATA_DIRS` too; a user's `~/.config/uwsm/env` runs after it): the KMS device from `/run/omamobile-tegu-display-card`, hardware GL when `/run/omamobile-tegu-gpu-ready.json` exists and llvmpipe otherwise, so a phone whose GPU did not come up still gets a desktop | | `vendor/modprobe.d/omamobile-tegu-guards.conf` | `/usr/lib/modprobe.d` | Keeps those drivers (and the Wi-Fi, PCIe and NFC ones, for the day they are in tree) from autoloading ahead of the guard. No `install` lines: they would break the guard's own modprobe | | `vendor/modprobe.d/omamobile-tegu-display.conf`, `omamobile-tegu-clk-acpm.conf` | `/usr/lib/modprobe.d` | 120 Hz as the ceiling and the panel really off on DPMS (the DSI commands themselves are allowed by the graphics start, for google-tg4c only); the GPU DVFS ceiling at 890 MHz | | `bin/omamobile-tegu-aoc-start` + `vendor/systemd/system/omamobile-tegu-aoc.service` + `vendor/modprobe.d/omamobile-tegu-aoc.conf` | `/usr/bin`, `/usr/lib/systemd/system` (not enabled), `/usr/lib/modprobe.d` | Boots the AoC (audio and sensor DSP) hang-safe: Trusty, the GSA, the AoC drivers, `aoc-tegu.bin`, then waits for its 93 services. Ships disabled: see below | diff --git a/packages/omamobile-device-tegu/install b/packages/omamobile-device-tegu/install index ed1d3f8ec..0d6921be2 100755 --- a/packages/omamobile-device-tegu/install +++ b/packages/omamobile-device-tegu/install @@ -20,6 +20,9 @@ while IFS= read -r -d '' file; do install -d "$destination/usr/lib/$(dirname "${file#vendor/}")" ln -sfn "$(readlink "$file")" "$destination/usr/lib/${file#vendor/}" done < <(find vendor -type l -print0) +# The session's side of the graphics start, sourced by uwsm from the XDG +# data dirs (uwsm-libexec/prepare-env.sh walks XDG_DATA_DIRS too). +install -Dm644 uwsm/env.d/omamobile-tegu-graphics "$destination/usr/share/uwsm/env.d/omamobile-tegu-graphics" install -Dm644 device.conf "$destination/usr/share/omamobile/device.conf" install -Dm644 etc/omamobile-mirrorlist "$destination/etc/pacman.d/omamobile-mirrorlist" install -Dm644 LICENSE "$destination/usr/share/licenses/omamobile-device-tegu/LICENSE" diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all index 923499f4f..d9ff2323b 100755 --- a/packages/omamobile-device-tegu/test/all +++ b/packages/omamobile-device-tegu/test/all @@ -14,6 +14,8 @@ tmp=$(mktemp -d); trap 'rm -rf "$tmp"' EXIT for script in build install bin/* vendor/systemd/system-sleep/* test/*; do bash -n "$script" || fail "$script parses" done +# uwsm sources the env.d file with a POSIX shell +sh -n uwsm/env.d/omamobile-tegu-graphics || fail "the uwsm env.d file parses as POSIX shell" pass "scripts parse" grep -q '\.\./\.\.' build install && fail "build or install reaches outside the package directory" @@ -85,6 +87,7 @@ for f in usr/bin/reboot-to usr/bin/tegu-tcpc-switch \ usr/lib/udev/rules.d/70-omamobile-dma-heap.rules usr/lib/udev/rules.d/72-omamobile-feedbackd.rules \ usr/lib/systemd/network/20-usb-gadget.network \ usr/lib/NetworkManager/conf.d/10-omamobile-usb.conf \ + usr/share/uwsm/env.d/omamobile-tegu-graphics \ usr/share/omamobile/device.conf \ usr/share/omamobile/device-tegu/version \ usr/share/licenses/omamobile-device-tegu/LICENSE \ @@ -122,6 +125,9 @@ for node in 1f000000.gpu 19470000.display-controller; do grep -qx "ConditionPathExists=|/sys/bus/platform/devices/$node" "$tmp/root/usr/lib/systemd/system/omamobile-tegu-graphics.service" \ || fail "the graphics start runs when $node exists (ORed condition)" done +grep -q 'omamobile-tegu-display-card' "$tmp/root/usr/share/uwsm/env.d/omamobile-tegu-graphics" \ + && grep -q 'omamobile-tegu-gpu-ready.json' "$tmp/root/usr/share/uwsm/env.d/omamobile-tegu-graphics" \ + || fail "the session reads the graphics start's two files" grep -q 'After=omamobile-tegu-graphics.service' "$tmp/root/usr/lib/systemd/system/display-manager.service.d/10-omamobile-tegu.conf" \ || fail "the display manager waits for the graphics start" for absent in usr/bin/omamobile-boot-success usr/bin/omamobile-bootctl \ diff --git a/packages/omamobile-device-tegu/uwsm/env.d/omamobile-tegu-graphics b/packages/omamobile-device-tegu/uwsm/env.d/omamobile-tegu-graphics new file mode 100644 index 000000000..7b953365a --- /dev/null +++ b/packages/omamobile-device-tegu/uwsm/env.d/omamobile-tegu-graphics @@ -0,0 +1,29 @@ +# SPDX-License-Identifier: MIT +# The session's side of omamobile-tegu-graphics-start, sourced by uwsm from +# /usr/share/uwsm/env.d (POSIX shell; a user's ~/.config/uwsm/env runs after +# it and may override). The display controller's DRM node is the one the +# start wrote to /run/omamobile-tegu-display-card, whichever card number it +# got (the GPU may take card0); hardware GL only when the start reported the +# GPU up in /run/omamobile-tegu-gpu-ready.json, otherwise llvmpipe, so a +# phone whose GPU did not come up still gets a desktop. +if [ -r /run/omamobile-tegu-display-card ]; then + read -r tegu_display_card < /run/omamobile-tegu-display-card + case "$tegu_display_card" in + /dev/dri/card[0-9]*) + if [ -c "$tegu_display_card" ]; then + export AQ_DRM_DEVICES="$tegu_display_card" + fi + ;; + esac + unset tegu_display_card +fi +if [ -c /dev/dri/renderD128 ] && [ -s /run/omamobile-tegu-gpu-ready.json ]; then + export LIBGL_ALWAYS_SOFTWARE=0 + unset GALLIUM_DRIVER QT_QUICK_BACKEND + export QSG_RHI_BACKEND=opengl +else + export LIBGL_ALWAYS_SOFTWARE=1 + export GALLIUM_DRIVER=llvmpipe + export QT_QUICK_BACKEND=software + unset QSG_RHI_BACKEND +fi From 77851552891b47d4b2e9d7573bf3ab34784d1c4a Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:19 -0647 Subject: [PATCH 18/20] omamobile-device-tegu: README on the AoC marker, the watchdog, Bluetooth and the smoke report The README said a failed AoC start keeps the marker; the code and its test clear it, since a load that returned is not a hang. It now says so. The AoC start's header cited docs/HANDOFF.md, which is not in the repository: it now points to the palmtop's bring-up script and the 2026-10-26 entry in device/tegu/NOTES.md. Two lines added: s3c2410_wdt.tmr_atboot=1 in the bootargs is what resets the palmtop after a hang, while the public tegu DTS has panic=10 only, so there a hang that does not panic ends at the power key; and Bluetooth comes up powered at every boot, as on the Pixel 7 Pro. One more says why a smoke report from the palmtop lists Wi-Fi and NFC as loaded beside this package: the bring-up's omamobile-tegu-audio-up, still running there, loads them, not the package. --- packages/omamobile-device-tegu/README.md | 19 +++++++++++++++++-- .../bin/omamobile-tegu-aoc-start | 10 ++++++---- 2 files changed, 23 insertions(+), 6 deletions(-) diff --git a/packages/omamobile-device-tegu/README.md b/packages/omamobile-device-tegu/README.md index 9b0df2701..97591b17a 100644 --- a/packages/omamobile-device-tegu/README.md +++ b/packages/omamobile-device-tegu/README.md @@ -28,6 +28,10 @@ packaged yet. What is here is what runs on in-tree drivers. | `src/reboot-to.c` | `/usr/bin/reboot-to` | Reboot with a reason the bootloader understands. One source with the cheetah package | | `device.conf`, `etc/omamobile-mirrorlist` | `/usr/share/omamobile`, `/etc/pacman.d` | The device descriptor (a copy of `device/tegu/device.conf`) and the `[omamobile]` repository | +Bluetooth comes up powered at every boot, as on the Pixel 7 Pro +(`omamobile-bt-addr@` powers the controller on to set its address, and +leaves it on). + `./build [DIR]` compiles the C tools, `./install DESTDIR [DIR]` stages everything and enables nothing beyond the two wants links, and `test/all` runs the package tests in a scratch root (`OMAMOBILE_TEGU_ROOT`), with no @@ -58,6 +62,11 @@ take over, `/chosen/bootargs` has to lose: `cs40l26_vibra.max_buzz_ms=100` stay. Whether and when that DTB change happens is the owner's call. +`s3c2410_wdt.tmr_atboot=1` is what resets the palmtop after a hang: the +watchdog runs from boot, so a hung graphics start ends in a reset and the +marker. The public tegu DTS carries `panic=10` but not that, so on it a +hang that does not panic ends at the power key (held about 30 s). + ## The AoC start ships disabled `omamobile-tegu-aoc-start` is the Pixel 7 Pro's AoC start with the 9a's @@ -66,8 +75,10 @@ channel through `aoc_channel_dev`). It needs a kernel whose device tree enables the AoC, GSA and Trusty nodes; today the bring-up turns them on with out-of-tree overlay modules, and so `omamobile-tegu-aoc.service` has no wants link. Once a kernel enables them, `systemctl enable -omamobile-tegu-aoc.service` and reboot; a failed start names the step in -the journal and the marker, and the next boot leaves the AoC off until +omamobile-tegu-aoc.service` and reboot. A step that returns and fails is +named in the journal and clears the marker (a load that came back is not a +hang; `test/aoc-start` checks this). Only a boot that hangs in the start +leaves the marker, and then the next boot leaves the AoC off until `omamobile-tegu-aoc-start --retry`. ## Firmware the phone needs @@ -126,5 +137,9 @@ and a leftover copy in `/etc` wins over the package's file of the same name. `omamobile-tegu-audio-up`, `omamobile-tegu-modem-up` and the out-of-tree modules they load stay until the kernel carries their drivers and device tree nodes; the package's guards and theirs do not load the same modules. +So when a smoke report from the palmtop (`tools/phone-smoke`) lists Wi-Fi +and NFC as loaded beside this package, it was `omamobile-tegu-audio-up` +that loaded them, from its own staged copies, not the package: the package +blacklists both and starts neither. Licence: MIT. diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start index 9bdb54821..b81e28fd5 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-aoc-start @@ -1,8 +1,9 @@ #!/bin/bash # SPDX-License-Identifier: MIT # Boot the Pixel 9a's AoC (audio and sensor DSP), hang-safe, in the order -# the bring-up proved (omamobile-tegu-audio-up, docs/HANDOFF.md S5) and the -# Pixel 7 Pro's omamobile-aoc-start uses: +# the bring-up proved on the palmtop (its omamobile-tegu-audio-up, still +# outside the package; device/tegu/NOTES.md 2026-10-26) and the Pixel 7 +# Pro's omamobile-aoc-start uses: # 1. Trusty (trusty_core, trusty_log, trusty_ipc, then trusty_virtio): # the AoC is started by Trusty's hwmgr.aoc service. # 2. The GSA mailbox (gsa), which authenticates the image. @@ -17,8 +18,9 @@ # This waits for a kernel whose device tree enables the AoC, GSA and Trusty # nodes: today they are turned on by out-of-tree overlay modules that no # package can ship, so the unit has no wants link (README.md). -# A marker is written and synced before any of it, and removed at the end. -# If a boot hangs here, the next one finds the marker and leaves the AoC off; +# A marker is written and synced before any of it, and removed at the end, +# after a step that returned and failed too (that is not a hang). If a boot +# hangs here, the next one finds the marker and leaves the AoC off; # "omamobile-tegu-aoc-start --retry" tries again. # omamobile-tegu-aoc-start [--retry] set -uo pipefail From 727575b6aab10d20bd465d59c523a3d8dc7eb4ac Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:35:19 -0647 Subject: [PATCH 19/20] omamobile-device-tegu: drop i2c-tools from the dependencies Nothing in the package runs i2cget, i2cset or i2cdetect any more: the USB rescue goes through tegu-tcpc-switch, which uses /dev/i2c-* directly, and the unit loads i2c-dev with kmod. --- packages/omamobile-device-tegu/PKGBUILD | 2 +- packages/omamobile-device-tegu/test/all | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/packages/omamobile-device-tegu/PKGBUILD b/packages/omamobile-device-tegu/PKGBUILD index 3520c3d44..a35c8d62b 100644 --- a/packages/omamobile-device-tegu/PKGBUILD +++ b/packages/omamobile-device-tegu/PKGBUILD @@ -7,7 +7,7 @@ pkgdesc="Omarchy Mobile support for the Google Pixel 9a (tegu): storage guard, U arch=(aarch64) url="https://github.com/derekross/omarchy-mobile" license=(MIT) -depends=(bash systemd kmod i2c-tools bluez-utils omamobile-keyring) +depends=(bash systemd kmod bluez-utils omamobile-keyring) provides=(omamobile-device) conflicts=(omamobile-device) backup=(etc/pacman.d/omamobile-mirrorlist) diff --git a/packages/omamobile-device-tegu/test/all b/packages/omamobile-device-tegu/test/all index d9ff2323b..62bab9a55 100755 --- a/packages/omamobile-device-tegu/test/all +++ b/packages/omamobile-device-tegu/test/all @@ -44,7 +44,7 @@ fi grep -q 'omamobile-keyring' PKGBUILD || fail "pacman needs omamobile-keyring" grep -q '^install=omamobile-device-tegu.install$' PKGBUILD \ || fail "the pacman install script is not named" -for dep in kmod i2c-tools bluez-utils; do +for dep in kmod bluez-utils; do grep -qE "^depends=.*\b$dep\b" PKGBUILD || fail "the scripts need $dep (depends)" done [[ -f etc/omamobile-mirrorlist && -f omamobile-device-tegu.install ]] \ From ffbf9587aab3f201bde91d2a0bc0a60218f9d614 Mon Sep 17 00:00:00 2001 From: W4HO <87742609+W4HO-ham@users.noreply.github.com> Date: Thu, 29 Oct 2026 12:40:27 -0647 Subject: [PATCH 20/20] omamobile-device-tegu: the panel name on the command line carries the panel's id The bootloader names the panel as exynos_drm.panel_name=google-tg4c.2d008004 on the palmtop, the id appended, so the exact match never allowed the DSI commands and the first boot of this round came up at 60 Hz with the warning from gs-decon. The match now takes google-tg4c with or without a suffix, and the test has the bootloader's form. --- .../bin/omamobile-tegu-graphics-start | 6 ++++-- packages/omamobile-device-tegu/test/graphics-start | 7 +++++++ 2 files changed, 11 insertions(+), 2 deletions(-) diff --git a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start index 2f31a08f8..4550601ce 100755 --- a/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start +++ b/packages/omamobile-device-tegu/bin/omamobile-tegu-graphics-start @@ -15,7 +15,8 @@ # # The panel's DSI commands (gs_dsim_cmd.allow_tx, for brightness, sleep and # 120 Hz) are allowed only when the bootloader names the google-tg4c panel -# on the command line (exynos_drm.panel_name=google-tg4c, kernel #90), and +# on the command line (exynos_drm.panel_name=google-tg4c, with the panel's +# id appended by the bootloader, kernel #90), and # before gs_decon loads, since gs-decon decides on 120 Hz at probe. # # This is the curated start of the bring-up (device/tegu/bringup) without its @@ -125,7 +126,8 @@ if [[ -e $decon ]]; then # when it already is) makes the parameter file exist before the write. modprobe gs_dsim_cmd 2>/dev/null || true panel=$(panel_name) - if [[ $panel == google-tg4c ]]; then + # the bootloader appends the panel's id (google-tg4c.2d008004 on the palmtop) + if [[ $panel == google-tg4c || $panel == google-tg4c.* ]]; then if echo 1 > "$allow_tx" 2>/dev/null; then log "panel $panel: DSI commands allowed (gs_dsim_cmd.allow_tx=1)" else diff --git a/packages/omamobile-device-tegu/test/graphics-start b/packages/omamobile-device-tegu/test/graphics-start index a7ff62b58..e214d690e 100755 --- a/packages/omamobile-device-tegu/test/graphics-start +++ b/packages/omamobile-device-tegu/test/graphics-start @@ -76,6 +76,13 @@ grep -q '^gs_decon .*allow_tx=1$' "$TEST_LOG" || fail "allow_tx was not 1 before [[ $(cat "$allow_tx") == 1 ]] || fail "allow_tx was not written" [[ $out == *"panel google-tg4c: DSI commands allowed"* ]] || fail "the allow is not logged: $out" +# The bootloader's form, the name with the panel's id appended (the palmtop's +# command line): allowed too +reset; cmdline google-tg4c.2d008004 +out=$("$script") +grep -q '^gs_decon .*allow_tx=1$' "$TEST_LOG" || fail "the suffixed panel name was not allowed: $(grep '^gs_decon' "$TEST_LOG")" +[[ $out == *"panel google-tg4c.2d008004: DSI commands allowed"* ]] || fail "the suffixed allow is not logged: $out" + # Another panel: allow_tx stays 0, gs_decon still loads reset; cmdline google-xyz9 out=$("$script")