[h4rkl]

What and why

Follow-up #1 from #74 (comment), now that #152 carries the package.

• device/tegu/NOTES.md: the second 9a's results as one dated entry (2026-10-28), in newest-first order. It states slot A's current state once, at the top: stock CP3A.261005.005, all five live boot images match the official October images by hash, and stock recovery boots (recovery ADB and its shell answered). That is not a full Android boot. This replaces the draft's "not a qualified rollback" line, which was written before the recovery check. The entry also notes that reboot-to bootloader came back to a normal boot on this phone, next to the button combination.
• device/tegu/kernel.config: the fragment these boots used. kernel.sources and prepare-kernel are left out, as asked.

The panel name and the answer on the touch marker (follow-up #2) will follow in a comment.

Device and build

• Phone: Pixel 9a (tegu), the second one; docs and kernel config only
• Kernel release (uname -r): the releases each result rests on are in the entry (…-g9ea0e19f8dc2, …-g9386dfaa1d5d, …-gde2e0689c517)
• Packages changed, with versions: none

Checks (no phone needed)

• tools/pr-check origin/main passes
• tests/all passes (CI; no scripts change)
• tests/packages passes (no package changes)
• Each changed package has its version bumped (none changed)
• Commits have area prefixes and bodies that say why; AI help is a Co-Authored-By: trailer

On the phone

• The results in the entry are from my own phone
• What I did not test: a full stock Android boot on slot A, a clean install, 0.2.0 on this phone
• Nothing in this change writes outside slot B and userdata, weakens the storage guard, or reads registers from userspace (AGENTS.md §1)

🤖 Generated with Claude Code

[derekross]

Thank you, @h4rkl. This is just what I hoped for: the second 9a's record, with slot A's state stated plainly. Stock CP3A.261005.005 with the live images matching Google's by hash, and a verified recovery boot while being clear that a full Android boot wasn't tested, is exactly the answer §3.1 needs. Thanks for correcting the old "not a qualified rollback" line, and for the reboot-to bootloader note.

I merged it (a squash on main, authored by you) with two small changes to kernel.config: a header saying it's the fragment of the 2026-10-28 boots and bring-up only (the release kernel is one build from the shared fragment; the eUSB2 PHY, mailbox, ACPM and S2MPG14 lines go once #227 puts them there; TG4C is already shared; touch as a module is a test choice), and the MIT SPDX line the other fragments use. Two wording nits for a later pass: "Omarchy on camera" reads oddly (filmed?), and "(the cap this branch's package set; 0.2.0 drops it)" is missing a word. I'm looking forward to the panel name and the answer on the touch marker.