Skip to content

device/akita: first light and the boot console (M1) - #414

Merged
derekross merged 2 commits into
derekross:mainfrom
aninmucek:device-akita-m1
Nov 5, 2026
Merged

derekross merged 2 commits into
derekross:mainfrom
aninmucek:device-akita-m1

Conversation

@aninmucek

@aninmucek aninmucek commented Nov 3, 2026 •

Copy link
Copy Markdown
Contributor

Follows #308 and #349. The Pixel 8a (akita) boots our kernel from slot B, with USB networking, SSH and the framebuffer console. The board file is in derekross/linux-omarchy-mobile#192.

  • device/akita/BRINGUP-M1.md (new): the M1 runbook, in husky's format: the build, the checks before a flash, the first boot and the console. It also records results: tests 1-5 and the image checksums.
  • device/akita/kernel.config (new): the soft and hard lockup detectors the board file's bootargs rely on, as in husky's.
  • device/akita/device.conf: SIMPLEFB_TRIG_CON="0x19470030 0x3061". As you asked on device/akita: stage 0 notes and descriptor for the Pixel 8a #308, TRIG_CON was read once from our initramfs shell with the bootloader's logo still on the panel, and recorded in the notes before the write went in. It reads 0x3070 (masked), husky's value.
  • device/akita/NOTES.md:
    • a 2026-11-03 section covering the booted tree, the kernel command line, the three display registers (DPP0 RDMA_BASEADDR_P0 0xfac00000, DECON0 GLOBAL_CON 0x133, TRIG_CON 0x3070) and the panel's command mode;
    • the provenance table row for the new key.
  • docs/DEVICES.md: the 8a's clause in the zuma Wanted row.

Two findings that may matter beyond akita:

  • This bootloader keeps VENDOR_CMDLINE. The kernel command line is our DTB bootargs, then VENDOR_CMDLINE, then ABL's own arguments. cheetah's notes say gs201's ABL ignores the vendor command line, and husky's test 9 says the same of zuma; on the 8a it doesn't. So splash is on, and the init draws Omarchy's splash, as for Plymouth, instead of its status lines. The panel showed the splash in place and in the right colours, then console text written over it.
  • Google's ak3b panel driver gives the panel as 70 × 149 mm. That is where stock's DPI comes from, and it's larger than a 6.1-inch 20:9 panel (about 64 × 141 mm), so the board file sets no size yet.

How it was tested:

  • Only the five *_b boot images were written, by bin/omamobile flash. Slot A was never touched, and slot B was never marked successful.
  • One reboot to fastboot went through our own reboot-to bootloader.
  • The images that booted are kept as known-good sets.
  • The raw captures (dmesg, the booted tree, the command line) stay private, since they hold the serial.

Checks:

  • tests/all: 375 ok, 0 not ok.
  • tools/pr-check: passes (identifiers, blobs, sign-offs, scripts).
  • The loader check, DEVICE=akita bin/omamobile initramfs-config: it packs the trigger pair and google,zuma-akita.
  • tools/phone-smoke: not run. It needs the Arch root and a session, and M1 boots only the initramfs.

🤖 Generated with Claude Code

The Pixel 8a boots our kernel from slot B, with USB networking, SSH and
the framebuffer console.

- BRINGUP-M1.md: the M1 runbook and its results (tests 1-5).
- kernel.config: the lockup detectors the board file's bootargs need,
  as in husky's.
- device.conf: SIMPLEFB_TRIG_CON="0x19470030 0x3061", after TRIG_CON
  read 0x3070 (masked) at handover on a booted 8a.
- NOTES.md: the booted tree, the kernel command line (this bootloader
  keeps VENDOR_CMDLINE), the three display registers and the panel's
  command mode.
- DEVICES.md: the 8a's clause in the zuma Wanted row.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@derekross

Copy link
Copy Markdown
Owner

Thank you, @aninmucek, this is a really clean first light. The runbook reads exactly like husky's, every step went the way it was written (the read-only TRIG_CON read before the descriptor got its write, one register per call, no width argument), and only slot B was ever written. The VENDOR_CMDLINE finding is a useful one: it's the first Tensor bootloader we've seen that keeps the vendor command line.

I checked the descriptor line against your notes: 0x19470030 is akita's own DECON0 plus TRIG_CON, the 0x3070 you read is husky's handover value, and 0x3061 is the same bit change husky makes. tools/pr-check passes, and there's nothing identifying in the diff or the thread.

Two small things before I merge:

  1. The status text predates the kernel PR. BRINGUP-M1.md's status line still says the kernel branch is "not yet pushed" and that the framebuffer change is "not yet committed", and the M1.4 results say "still uncommitted". Both are now Pixel 8a (akita): the first-light board file, with USB, SSH and the boot framebuffer linux-omarchy-mobile#192. Could you name that PR and its two DTS commits there? Please keep the -dirty release in the results, since it's the record of what ran.
  2. Tying the M1.4 run to tools: verify signed root images and isolate device builds #192. That run's kernel was -dirty, so let's record that it's the same source. tools: verify signed root images and isolate device builds #192's zuma-akita.dts blob is 381ad1e0e7cbf00a74a8d59f1cf48af97046adcb, and my build of tools: verify signed root images and isolate device builds #192 on today's omamobile gives a zuma-akita.dtb with sha256 9a5d92a580a8275d39a67b3a13571532374c85d37c8030111d026257d91a123d (no zuma file changed since 4cf2c8be1). If git hash-object on the DTS you kept with akita-m1.4-gce160abd9ffc-dirty, and the sha256 of that set's DTB, give the same values, one line in the results saying so is all it needs.

Smaller suggestions, none blocking:

  • In Risks: each unmarked boot uses one of slot B's tries, and at zero ABL boots stock on slot A. If that happens, decline any factory reset (Volume Down + Power gets you to fastboot), and take a fresh read-only backup afterwards, because stock rewrites efs, persist and modem_userdata when it boots.
  • M1.2 step 4 uses ~/android/akita/bin/akita-state, which nobody else has. Could you list the fastboot getvar lines it runs? And please add -s <serial> to the two fastboot --set-active=a rollback lines.
  • tools/phone-smoke needs the Arch root and a session, so it can't run on an initramfs-only boot. A sentence in the PR saying so covers the template.

On the panel size: leaving it out for M1 is fine. You're right that 70 x 149 mm is too big for a 6.1-inch active area (stock's DPI reproduces exactly those numbers). We'll need some value before the session runs, because simpledrm otherwise assumes 96 dpi and the keyboard comes out half-height on husky. We can settle that then.

I'll merge #192 first, then this one. Thanks again!

From the review of derekross#414:

- The status line names derekross/linux-omarchy-mobile#192 and its
  commits. M1.3's kernel commit, ce160abd9, is derekross#192's 1253914a9 with
  the same files. M1.4's -dirty run is derekross#192's 6a2999804: the kept DTS
  is blob 381ad1e0 and the flashed DTB has sha256 9a5d92a5, as derekross#192's.
- Risks: every unmarked boot uses one of slot B's tries. If stock boots,
  decline any factory reset and take a fresh backup.
- M1.2 lists the fastboot getvar values instead of a local helper, and
  the two rollbacks pin the serial.
- NOTES.md: the COMPATIBLE and DTB rows point at derekross#192.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@aninmucek

Copy link
Copy Markdown
Contributor Author

Thanks for the thorough review! Both are done in the new commit:

  1. The status line now names Pixel 8a (akita): the first-light board file, with USB, SSH and the boot framebuffer linux-omarchy-mobile#192 and its commits: the binding 7385cab4b, the first-light DTS 1253914a9 and the framebuffer 6a2999804. The M1.4 results keep the -dirty release, and say the change is now 6a2999804.
  2. The source check matches. git hash-object on the DTS kept with the akita-m1.4-gce160abd9ffc-dirty set gives 381ad1e0e7cbf00a74a8d59f1cf48af97046adcb, tools: verify signed root images and isolate device builds #192's blob. That set's DTB has sha256 9a5d92a580a8275d39a67b3a13571532374c85d37c8030111d026257d91a123d, as does the DTB inside the vendor_kernel_boot image that was flashed. Both match your build. That line is now in the M1.4 results.

The same goes for M1.3's boot. Its kernel commit, ce160abd9, is #192's 1253914a9 before the message gained the test result and the move onto fe9627c00. The DTS, the Makefile, the binding and zuma.dtsi are identical, and the M1.1 results now say so.

I took the smaller suggestions too:

  • Risks: every boot that doesn't mark slot B uses one of its tries, a working one included. Our first boot left 2 of 3. If stock comes up on slot A, decline any factory reset (Volume Down + Power for fastboot), then take a fresh read-only backup.
  • M1.2 step 4 now lists the fastboot getvar values (current-slot, slot-successful:a/:b, unlocked, version-bootloader) instead of my local helper. Both rollback lines use fastboot -s <serial> --set-active=a.
  • tools/phone-smoke: the PR description now says it wasn't run, since it needs the Arch root and a session and M1 boots only the initramfs.

I also pointed NOTES.md's COMPATIBLE and DTB rows at #192. The DTB row still said the board DTS wasn't written yet.

On the panel size: agreed, it can wait until the session needs it.

@derekross

Copy link
Copy Markdown
Owner

Thank you, @aninmucek, everything I asked is in, and the source check matches. Your DTS blob 381ad1e0… is #192's, and I rebuilt #192 on today's omamobile (which has moved by one lynx commit since) and got the same zuma-akita.dtb, sha256 9a5d92a5…. So the -dirty M1.4 run is #192's source on today's base as well. The new Risks lines, the getvar list and the -s <serial> rollbacks all read well.

I'll merge derekross/linux-omarchy-mobile#192 first, once its CI has run, then this. At merge I'll update the DEVICES.md clause from "a board DTS, not yet merged" to a link to #192. Thanks again for such a clean first light!

@derekross
derekross merged commit 905aa91 into derekross:main Nov 5, 2026
2 checks passed
@aninmucek
aninmucek deleted the device-akita-m1 branch November 6, 2026 05:14
derekross added a commit that referenced this pull request Nov 7, 2026
The matrix re-read against what merged and ran on hardware since 2026-11-02:
- Pixel 7: Bluetooth and the torch move to testing (kernel #196, #197, #426); fingerprint moves to WIP (kernel #207; the package side, #458, is in review); AVB footers and reboot=warm noted.
- Pixel 8a: a new row, from first light and UFS behind the guard (kernel #192, #206; #414).
- Pixel 6 Pro: GPU scaling and cooling (kernel #189) and UFS suspend (kernel #188) merged.
- Pixel 9a: GPU levels merged through kernel #96; GNSS's moving-car session (#114) and the link driver (kernel #185).
- Pixel 7a: Bluetooth's device tree merged (kernel #98).
The Pixel 8 Pro's row is left to #472.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
The Pixel 8a (akita) boots our kernel from slot B with USB networking,
SSH and the framebuffer console; the board file is
linux-omarchy-mobile#192. BRINGUP-M1.md holds the runbook in husky's
format with tests 1-5 and the image checksums; kernel.config adds the
lockup detectors the board file's bootargs rely on; device.conf gets
SIMPLEFB_TRIG_CON="0x19470030 0x3061", from TRIG_CON read once (0x3070,
husky's value) before the write went in. NOTES records the booted tree,
the command line and the display registers, and two findings: this
bootloader keeps VENDOR_CMDLINE (so `splash` takes effect), and Google's
panel driver gives a size too large for the panel, so the board sets
none yet. Only slot B's five images were written, slot B was never
marked, and the -dirty M1.4 kernel's DTS and DTB match derekross#192's.
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
The matrix re-read against what merged and ran on hardware since 2026-11-02:
- Pixel 7: Bluetooth and the torch move to testing (kernel derekross#196, derekross#197, derekross#426); fingerprint moves to WIP (kernel derekross#207; the package side, derekross#458, is in review); AVB footers and reboot=warm noted.
- Pixel 8a: a new row, from first light and UFS behind the guard (kernel derekross#192, derekross#206; derekross#414).
- Pixel 6 Pro: GPU scaling and cooling (kernel derekross#189) and UFS suspend (kernel derekross#188) merged.
- Pixel 9a: GPU levels merged through kernel derekross#96; GNSS's moving-car session (derekross#114) and the link driver (kernel derekross#185).
- Pixel 7a: Bluetooth's device tree merged (kernel derekross#98).
The Pixel 8 Pro's row is left to derekross#472.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants