Skip to content

device-dodge: follow-ups from #111 (fastboot refuse, install bind trap, park deliberate panic) - #218

Merged
derekross merged 4 commits into
derekross:mainfrom
mrlund:dodge/followup-1
Oct 31, 2026
Merged

derekross merged 4 commits into
derekross:mainfrom
mrlund:dodge/followup-1

Conversation

@mrlund

@mrlund mrlund commented Oct 31, 2026

Copy link
Copy Markdown
Contributor

dodge follow-ups from the #111 merge note

Small follow-up for the three non-blocking items from the #111 merge. Branch dodge/followup-1 on top of today's main (8efed1a).

  1. omamobile fastboot refuses on dodge. Same early refuse as flash / push for BOOT_METHOD=qcom-abl, before any phone I/O. The sysrq-c fallthrough stays uncharacterized on this ABL; prefer the buttons. Covered in tests/boot-method (including an android-v4 cheetah path that still gets past the refuse). BRINGUP §2/§5 updated.

  2. Install bind of / undone in a trap. device/dodge/audio/install and device/dodge/modem/install privately bind / to write through the kmod-overlay. They now trap umount/rmdir on EXIT so a mid-install failure can't leave the bind up (same idea as bluetooth/install). tests/boot-method checks the trap is present and that a forced failure after mount --bind / still umounts.

  3. Deliberate panic parked until gated. Moved the deliberate panic test out of Still open into Parked until gated: don't run it until there's an explicit gate, a stop condition, and a way back.

tools/pr-check, tests/boot-method, and tests/pr-check are green on this tip.

Commits:

  • 75d9033 omamobile: refuse fastboot on qcom-abl
  • 8202df3 dodge: undo private bind of / in install traps
  • 26cc15d dodge: park deliberate panic until gated
  • 709d553 tests: assert android-v4 gets past fastboot qcom-abl refuse

Derek's derekross#111 merge follow-up: omamobile fastboot falls through to sysrq c
when reboot-to bootloader is ignored, and panic on this ABL is still
uncharacterized. Refuse BOOT_METHOD=qcom-abl before any phone I/O (same
class as flash/push). Prefer the buttons. BRINGUP §2/§5 updated; test
added in tests/boot-method.

Signed-off-by: Martin Lund <mrlund@hotmail.com>
Assisted-by: Grok Bot
Derek's derekross#111 merge follow-up: audio and modem install privately bind /
to write through the kmod-overlay, but a mid-install failure left that
bind up. Trap umount/rmdir on EXIT (same idea as bluetooth/install).
tests/boot-method checks the trap is present and that a forced failure
after mount --bind / still umounts.

Signed-off-by: Martin Lund <mrlund@hotmail.com>
Assisted-by: Grok Bot
Derek's derekross#111 merge follow-up: the deliberate sysrq-c / panic step listed
under BRINGUP §3.1 was actionable without a gate, stop condition, or way
back. Move it out of Still open into Parked until gated (off by default;
do not run). Soften the §5 invitation that pointed at running it. Prefer
the minimal honest option over inventing a gate in this follow-up.

Signed-off-by: Martin Lund <mrlund@hotmail.com>
Assisted-by: Grok Bot
Claude round7: cmd_fastboot now load_device's before the refuse; cover a
non-qcom descriptor so Pixel paths stay past the gate.

Signed-off-by: Martin Lund <mrlund@hotmail.com>
Assisted-by: Grok Bot
@derekross
derekross merged commit 10daa3d into derekross:main Oct 31, 2026
2 checks passed
@derekross

Copy link
Copy Markdown
Owner

Thanks, all three are clean, and the test covers both boot methods. Merged. One question for the plan: the shared init panics on any boot where no USB host shows up within 45 s, so has dodge ever booted unplugged? If so, that boot is your panic data point; if not, keep the cable in for now.

@mrlund

mrlund commented Oct 31, 2026

Copy link
Copy Markdown
Contributor Author

Yes, and I just re-checked it on the merged shared init. RAM boot of an image packed from current main, cable pulled as soon as fastboot finished. It reached the Omarchy lock screen with no panic across 90+ s unplugged: switch_root at 1.86 s, which kills the 45 s job long before it fires, the same boot id the whole time and no "no USB host" message. Carrier came back on replug.

For context: before Oct 24 our own init waited on USB before switch_root, so cable-out boots raced the panic (one warned at ~47 s and only just switched at ~62 s). Removing that wait fixed it, and the shared init doesn't have the wait either.

robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…l bind on failure, park the deliberate panic (derekross#218)

Follow-ups from derekross#111: bin/omamobile fastboot refuses a qcom-abl device after reading its descriptor and before any phone I/O (Pixels only gain the descriptor read; tests/boot-method covers both boot methods), the install scripts undo their bind of / in an EXIT trap when a step fails, and BRINGUP parks the deliberate sysrq panic until it has a gated runbook.
quatrixone added a commit to quatrixone/omarchy-mobile that referenced this pull request Nov 18, 2026
Record that the manual EA8074 test used an unpublished kernel while panel driver derekross#218 remains in review.

AI assistance: Codex.

Signed-off-by: quatrixone <quatrix1@gmail.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