diff --git a/device/bluejay/NOTES.md b/device/bluejay/NOTES.md new file mode 100644 index 000000000..a21cf751e --- /dev/null +++ b/device/bluejay/NOTES.md @@ -0,0 +1,77 @@ + +# Pixel 6a (bluejay) bring-up notes + +Findings for the Pixel 6a port, newest first. The port shares the Pixel 6 (oriole) and 6 Pro (raven) effort tracked in [issue #2](https://github.com/derekross/omarchy-mobile/issues/2). Facts are marked [V] (checked, with the source) or [I] (an inference to check on a phone or in the factory image). Nothing has been flashed or booted yet, and no 6a has been queried: every fact below comes from public GPL sources on the laptop. + +Sources: + +- `G/`: Google's GPL device tree for bluejay, [kernel/devices/google/bluejay](https://android.googlesource.com/kernel/devices/google/bluejay/+/refs/tags/android-16.0.0_r0.9/dts/) at tag `android-16.0.0_r0.9`. +- `S/`: Google's GPL gs101 SoC device tree, [kernel/devices/google/gs101](https://android.googlesource.com/kernel/devices/google/gs101/+/refs/tags/android-16.0.0_r0.9/dts/) at the same tag. +- `R/`: Google's GPL device tree for oriole and raven, [kernel/devices/google/raviole](https://android.googlesource.com/kernel/devices/google/raviole/+/refs/tags/android-16.0.0_r0.9/dts/) at the same tag, as the control: what it says about oriole is checked against [oriole's notes](../oriole/NOTES.md). +- `U/`: upstream Linux, `arch/arm64/boot/dts/exynos/google/gs101*.dts(i)` and `Documentation/devicetree/bindings/arm/google.yaml`. +- `P/`: the pmaports `linux-postmarketos-gs101` package (7.3_rc6) and the tree it builds, [tensor-exynos/linux @ e8f42849a581](https://gitlab.postmarketos.org/tensor-exynos/linux/-/tree/e8f42849a581), for a second reading of the board file. +- `O/`: [oriole's notes](../oriole/NOTES.md), for the facts a 6a shares with a Pixel 6 (the factory image layout, the retail phone's partitions). Each is [I] for bluejay until its own image or phone confirms it. + +## 2026-11-01: stage 0, part 1: facts from Google's GPL device trees (no phone, no factory image) + +### Base DTB and `soc_id` + +- **Bluejay boots the same two base DTBs as oriole and raven** [V] (G/ `Makefile`, `gs101_dtbs`): `gs101/gs101-a0.dtb` and `gs101/gs101-b0.dtb`, with the board as a dtbo overlay on top. So, as on oriole, our board DTS is the base DTB and ABL picks it by `soc_id`. +- **`soc_id` A0 is `<0x09845000 0x0>`, B0 is `<0x09845001 0x0>`** [V] (S/ `gs101-a0.dts`, `gs101-b0.dts`, line 12 of each). This settles the label that O/ left [I]: `0x...001` is B0, as cheetah's `0x9855001` is. Which of the two a retail 6a has is still [I]; `bin/omamobile images` ships both. +- **The first-light board DTS** (`gs101-bluejay.dts`, prepared on a local `omamobile-gs101-bluejay-dts` kernel branch, to be opened as a pull request against `omamobile`) is oriole's file with bluejay's model and compatible (`google,gs101-bluejay`, `google,gs101`), `soc_id` set to B0, `/chosen/bootargs` with `panic=10`, the 1080x2400 framebuffer and splash carve-out, and without the two LDOs described next. U/ `master` has no bluejay binding or DTS; the binding entry goes in the same pull request. + +### Board dtbo + +[V] (G/ `gs101-bluejay-*.dts`, read from each overlay's `board_id`/`board_rev`): + +| Overlay | board_id | board_rev | +| --- | --- | --- | +| `dev` | 0x20501 | 0x10000 | +| `proto1_0`, `proto1_1` | 0x20502 | 0x10000, 0x10100 | +| `evt1_0`, `evt1_1`, `evt1_2` | 0x20503 | 0x10000, 0x10100, 0x10200 | +| `dvt1_0`, `dvt1_1` | 0x20504 | 0x10000, 0x10100 | +| `pvt1_0` | 0x20505 | 0x10000 | +| `mp1_0` | 0x20506 | 0x10000 | + +- **The overlay's model and compatible** [V]: `GS101 BLUEJAY MP 1.0 based on GS101`, `google,GS101 BLUEJAY` (G/ `gs101-bluejay-mp1_0.dts`). +- **`device.conf` uses the MP 1.0 entry, 0x20506/0x10000.** That a retail 6a boots it is [I]: oriole, lynx and panther boot their MP entry, but raven boots its DVT entry (0x20204, `ro.boot.dtbo_idx` 6, [raven's descriptor](../raven/device.conf)), so only a stock `dtbo.img` or `ro.boot.dtbo_idx` on a retail phone settles it (O/ did that for oriole on 2026-10-16). The entry index is never the value (review lesson L-365). +- **Oriole's control** [V]: R/ `gs101-oriole.dts` is 0x20304/0x10000, what O/ records from the factory image and the phone, so reading Google's overlays gives the right retail entry for a sibling. + +### Panel and touch + +- **The panel is a Samsung S6E3FC3, bluejay variant** [V] (G/ `gs101-bluejay-display.dtsi`): `samsung,s6e3fc3_6a`, DSI channel 0, reset on `gpp24-1`, VCI from `m_ldo22_reg` at 3.025 V, `vddi-supply` deleted; 1080x2400; DSIM PLL p/m/s/k `0x2 0xDF 0x3 0xA800`, HS clock 684 MHz, escape clock 20 MHz. The second DSIM (`drmdsim1`) is disabled. Oriole's panel driver lead (the S6E3FC3 on kernel #116's gs-decon) applies, with a 6a panel entry still to write. +- **Touch is an ST FTS on SPI** [V] (G/ `gs101-bluejay-touch.dtsi` including S/ `gs101-fts-touch.dtsi`): 10 MHz, chip id `0x36 0x48`, maximum coordinates 1079x2399, the under-display fingerprint area at (540, 1717), firmware `ftm5_fw_b3.ftb`, the same controller family as oriole's. The driver is out of tree (P/ builds one); nothing in `omamobile` yet. + +### Regulators the board file leaves out + +- **Upstream's oriole DTS describes two LDOs on the main PMIC (`ldo14m`, `ldo31m`) under `acpm_ipc`; raven's does not, and bluejay's doesn't either.** The reason [V]: both come from the shared S/ `gs101-pmic.dtsi`, `regulator-always-on` (LDO14M `L14M_AVDD_TCXO`, 1.6-1.95 V; LDO31M `L31M_NFC`, 1.6-1.95 V), and S/ `gs101-whitefin2-pmic.dtsi`, which both bluejay and oriole include, renames LDO31M to `L31M_DISP_VDDI`. Bluejay's own G/ `gs101-bluejay-pmic.dtsi` touches neither; neither does R/ `gs101-oriole-pmic.dtsi`. So the upstream block is a description of the shared values, not an oriole-only fact, and a rail Linux doesn't describe stays as the bootloader left it (the regulator core only turns off rails it knows about, AGENTS.md). If a first boot shows the display's VDDI or the TCXO dropping, the block is the first thing to add, with that boot as the evidence. +- **Bluejay's include list** [V] (G/ `gs101-bluejay.dtsi`): `gs101-common`, the raviole battery, bluejay's camera-pmic, battery, charging, usecases, camera, display, thermal, pmic, sbbm, the whitefin modem (`whitefin-cp-s5200-sit-common`, `whitefin2-cp-s5200-sit`), NFC, `slider-gsa-gsc`, `slider2-gps`, `whitefin2-pmic`, `common-bcl` and bluejay's Type-C; the MP overlay adds audio, fingerprint, touch, the eSE and USB. The modem, GNSS, GSA and PMIC leads from raven's plans therefore apply. + +### What `device.conf` rests on + +The descriptor carries no comments; this table is its provenance. + +| Key | Value | Mark | Source | +| --- | --- | --- | --- | +| `CODENAME`, `NAME`, `SOC` | bluejay, Pixel 6a, gs101 | [V] | G/, the kernel binding | +| `COMPATIBLE` | `google,gs101-bluejay` | [V] | the first-light DTS and its binding entry | +| `PACKAGE` | `omamobile-device-bluejay` | the package doesn't exist yet; `bin/omamobile initramfs` falls back when its `omamobile-bootctl.c` is missing, as for oriole | | +| `DTB` | `exynos/google/gs101-bluejay.dtb` | [V] | the Makefile line on the local `omamobile-gs101-bluejay-dts` kernel branch, to be opened as a pull request against `omamobile`; in no published kernel yet, so `DEVICE=bluejay bin/omamobile images` waits for that pull request | +| `SOC_IDS` | `b0:0x9845001 a0:0x9845000` | [V] | S/ `gs101-b0.dts`, `gs101-a0.dts` | +| `DTBO_SOURCE`, `DTBO_BOARD_ID`, `DTBO_BOARD_REV` | the empty overlay, 0x20506, 0x10000 | [V] for the MP 1.0 entry, [I] that retail boots it | G/ `gs101-bluejay-mp1_0.dts` | +| `VENDOR_CMDLINE` | oriole's line: `earlycon` at `0x10A00000`, the gs101 UART | [I] | O/ (stock oriole's bootargs); the 6a's stock `vendor_boot` confirms | +| `VENDOR_PAGESIZE` | 0x800 | [I] | O/ (oriole's v4 header); the 6a's stock `vendor_boot` header confirms | +| `OS_VERSION`, `OS_PATCH_LEVEL` | 17.0.0, 2026-10 | [I] | O/ (oriole's factory build); the current bluejay factory build id replaces them. Stock's GKI `boot` header leaves `os_version` at 0 and the AVB footer carries the version (O/), so these header fields are informational; the rollback question is the AVB one below | +| `BOOT_IMAGES`, `UFS_RW` | `boot vendor_boot dtbo`; `boot_b vendor_boot_b dtbo_b userdata` | [I] | O/ (no `init_boot`, no `vendor_kernel_boot` on a Pixel 6); the 6a's `fastboot-info.txt` or `getvar all` confirms | +| `INITRAMFS_INIT` | the shared init | [V] | as oriole | +| `PANEL_SIZE` | 1080x2400 | [V] | G/ `gs101-bluejay-display.dtsi` | +| `AVB_FOOTERS`, `AVB_ROLLBACK_INDEX`, `AVB_SECURITY_PATCH` | unset | open | whether the 6a's `vbmeta_b` chains `boot` and `vendor_boot` as stock's does is not known: the factory `vbmeta.img` (`avbtool info_image`, chain partition descriptors) and the stock images' footers (rollback index, `com.android.build.*.security_patch`) decide it ([RELEASES](../../docs/RELEASES.md) §3.1). Oriole's native boot records a `vbmeta_b` change made without approval (O/), so footer-less images on slot B are not known to boot on a gs101 Pixel; no flash until this is settled | +| `SIMPLEFB_TRIG_CON`, `BRINGUP_PROBE` | unset | | gs101's DECON trigger register (`0x1c300030`, [V] on raven) is read on a booted 6a before it is written anywhere (review lessons L-4, L-365) | + +### Still needs the factory image or a phone + +From the current `bluejay-*-factory-*.zip` (HTTP range reads, facts only, as O/ did for oriole): `android-info.txt` (`board=`, the required bootloader), `fastboot-info.txt` (the partitions flashed), the `vendor_boot` header (version, page size, the two DTBs' `soc_id`, stock `/chosen/bootargs`), the `boot` header, the `dtbo.img` table (expect the six board ids above; which entry is 0x20506), the build id for `OS_PATCH_LEVEL`, and `vbmeta.img`'s chain partition descriptors with the `boot` and `vendor_boot` footers' rollback index and security patch, which decide whether slot B needs `AVB_FOOTERS`. + +From a retail 6a, read-only, before anything is written: `getvar all` (product, slot, partition list, `hw-revision`), `ro.boot.dtbo_idx`, the build id and security patch; with root or `bin/omamobile capture`, the live DT (`soc_id`, `/chosen/bootargs`, every `reserved-memory` region against `/proc/iomem`). Post facts, never the files. A retail 6a has a locked bootloader, and unlocking wipes it; the stock capture (CONTRIBUTING §5.1) comes first. + +Before the first boot of slot B: the AVB answer above, the board DTS's DTB in a kernel build (`DEVICE=bluejay bin/omamobile images`), the oriole initramfs path (`UFS_RW` packed into `/etc/omamobile-ufs-rw`, `tests/storage-guard`), and a human at the phone with the way back to slot A. diff --git a/device/bluejay/bluejay-m1-overlay.dts b/device/bluejay/bluejay-m1-overlay.dts new file mode 100644 index 000000000..6df6913d7 --- /dev/null +++ b/device/bluejay/bluejay-m1-overlay.dts @@ -0,0 +1,14 @@ +// SPDX-License-Identifier: GPL-2.0-only +/dts-v1/; +/plugin/; + +/ { + board_id = <0x20506>; + board_rev = <0x10000>; + + fragment@0 { + target-path = "/"; + __overlay__ { + }; + }; +}; diff --git a/device/bluejay/device.conf b/device/bluejay/device.conf new file mode 100644 index 000000000..aa582c287 --- /dev/null +++ b/device/bluejay/device.conf @@ -0,0 +1,19 @@ +# SPDX-License-Identifier: MIT +CODENAME=bluejay +NAME="Pixel 6a" +SOC=gs101 +COMPATIBLE="google,gs101-bluejay" +PACKAGE=omamobile-device-bluejay +DTB=exynos/google/gs101-bluejay.dtb +SOC_IDS="b0:0x9845001 a0:0x9845000" +DTBO_SOURCE=bluejay-m1-overlay.dts +DTBO_BOARD_ID=0x20506 +DTBO_BOARD_REV=0x10000 +VENDOR_CMDLINE="console=tty0 earlycon=exynos4210,0x10A00000 clk_ignore_unused pd_ignore_unused printk.devkmsg=on bootconfig" +VENDOR_PAGESIZE=0x800 +OS_VERSION=17.0.0 +OS_PATCH_LEVEL=2026-10 +BOOT_IMAGES="boot vendor_boot dtbo" +UFS_RW="boot_b vendor_boot_b dtbo_b userdata" +INITRAMFS_INIT=/initramfs/init +PANEL_SIZE=1080x2400 diff --git a/docs/DEVICES.md b/docs/DEVICES.md index 16cc933ef..a70426c7c 100644 --- a/docs/DEVICES.md +++ b/docs/DEVICES.md @@ -69,7 +69,7 @@ Devices we'd like to support, with the closest starting point. Open an issue to | Pixel Tablet | tangorpro | gs201 | as panther; no modem | | Pixel 6 | oriole | gs101 | **WIP, community port: [#2](https://github.com/derekross/omarchy-mobile/issues/2)**; our preferred second device. Upstream mainline `gs101-oriole.dts` | | Pixel 6 Pro | raven | gs101 | upstream mainline `gs101-raven.dts`; shares oriole's port ([#2](https://github.com/derekross/omarchy-mobile/issues/2)). Boots our kernel from slot B to Omarchy's Hyprland on the Mali-G78 ([kernel #2](https://github.com/derekross/linux-omarchy-mobile/pull/2), [#3](https://github.com/derekross/linux-omarchy-mobile/pull/3)); its device package and plans ([#41](https://github.com/derekross/omarchy-mobile/pull/41)) and provisioning tools ([#22](https://github.com/derekross/omarchy-mobile/pull/22)) are merged | -| Pixel 6a | bluejay | gs101 | upstream `gs101.dtsi`; a new board DTS; shares oriole's port ([#2](https://github.com/derekross/omarchy-mobile/issues/2)) | +| Pixel 6a | bluejay | gs101 | **WIP, community port: [#2](https://github.com/derekross/omarchy-mobile/issues/2)**; stage 0 facts from Google's GPL bluejay tree, no phone yet ([notes](../device/bluejay/NOTES.md)); its first-light board DTS (`gs101-bluejay.dts`) is prepared on a local `omamobile-gs101-bluejay-dts` kernel branch, to be opened as a pull request against `omamobile`, and is in no published kernel; shares oriole's port | | Pixel 8, 8 Pro, 8a | shiba, husky, akita | zuma (Tensor G3) | **WIP, community port: [#1](https://github.com/derekross/omarchy-mobile/issues/1)**. 8 Pro (husky): WIP, in the matrix above; its device package and UFS plan are merged ([#42](https://github.com/derekross/omarchy-mobile/pull/42)), and its bootloader picks the base DTB by `/soc_compatible`, not `/soc_id`. 8a (akita): wanted; the earlier hardware offer was withdrawn, and husky's zuma work carries over | | Pixel 9 family | tokay, caiman, komodo, comet, tegu | zumapro (Tensor G4) | 9a (tegu): WIP, in the matrix above; SoC patches are merging into linux-omarchy-mobile, the board DTS isn't public yet; tracking [#28](https://github.com/derekross/omarchy-mobile/issues/28) | | Pixel 10 family | mustang (10 Pro XL) | laguna (Tensor G5) | 10 Pro XL (mustang): stage 0 stock facts ([notes](../device/mustang/NOTES.md)), contributors welcome ([#18](https://github.com/derekross/omarchy-mobile/issues/18)) | diff --git a/tests/storage-guard b/tests/storage-guard index 695d09cf7..409f5ec3d 100755 --- a/tests/storage-guard +++ b/tests/storage-guard @@ -14,9 +14,10 @@ fail() { printf 'not ok - %s\n' "$1" >&2; exit 1; } pass() { printf 'ok - %s\n' "$1"; } t=$(mktemp -d); trap '[[ -n ${KEEP:-} ]] || rm -rf "$t"' EXIT cheetah_rw="boot_b init_boot_b vendor_boot_b vendor_kernel_boot_b dtbo_b userdata" +gs101_rw="boot_b vendor_boot_b dtbo_b userdata" # --- Pack time: the real descriptors --- -for d in cheetah:"$cheetah_rw" felix:"$cheetah_rw" lynx:"$cheetah_rw" oriole:"boot_b vendor_boot_b dtbo_b userdata"; do +for d in cheetah:"$cheetah_rw" felix:"$cheetah_rw" lynx:"$cheetah_rw" oriole:"$gs101_rw" bluejay:"$gs101_rw"; do dev=${d%%:*} want=${d#*:} DEVICE=$dev bin/omamobile initramfs-config "$t/etc-$dev" > "$t/log" 2>&1 || { cat "$t/log"; fail "$dev: initramfs-config"; } [[ $(cat "$t/etc-$dev/omamobile-ufs-rw") == "$want" ]] || fail "$dev packs its UFS_RW (got $(cat "$t/etc-$dev/omamobile-ufs-rw"))" @@ -30,7 +31,7 @@ if [[ $(grep -c '^ *UFS_RW=' initramfs/init) != 2 ]] || ! grep -q '^UFS_RW=""$' fail "initramfs/init has no allowlist of its own, only the packed one" fi grep -q '^ initramfs_config "\$stage/etc"$' bin/omamobile || fail "initramfs packs the files into the image's /etc" -pass "cheetah and felix pack the six-name list, oriole its three boot images and userdata" +pass "cheetah and felix pack the six-name list, oriole and bluejay their three boot images and userdata" # --- Pack time: refused descriptors, in a repository-shaped tree --- R=$t/repo diff --git a/tests/usb-product b/tests/usb-product index 03bdb9fd1..6f0f4ea9d 100755 --- a/tests/usb-product +++ b/tests/usb-product @@ -10,12 +10,12 @@ pass() { printf 'ok - %s\n' "$1"; } t=$(mktemp -d); trap '[[ -n ${KEEP:-} ]] || rm -rf "$t"' EXIT # --- Pack time --- -for d in "cheetah:Pixel 7 Pro (M1)" "felix:Pixel Fold" "oriole:Pixel 6"; do +for d in "cheetah:Pixel 7 Pro (M1)" "felix:Pixel Fold" "oriole:Pixel 6" "bluejay:Pixel 6a"; do dev=${d%%:*} want=${d#*:} DEVICE=$dev bin/omamobile initramfs-config "$t/etc-$dev" > "$t/log" 2>&1 || { cat "$t/log"; fail "$dev: initramfs-config"; } [[ $(cat "$t/etc-$dev/omamobile-usb-product") == "$want" ]] || fail "$dev packs '$want' (got $(cat "$t/etc-$dev/omamobile-usb-product"))" done -pass "cheetah keeps 'Pixel 7 Pro (M1)'; felix and oriole use their NAME" +pass "cheetah keeps 'Pixel 7 Pro (M1)'; felix, oriole and bluejay use their NAME" R=$t/repo mkdir -p "$R/bin" "$R/device/six"