Skip to content

omamobile-modem 0.1.19: one unclean end no longer costs the next boot its modem - #634

Merged
derekross merged 1 commit into
mainfrom
modem-forced-reset-policy
Nov 10, 2026
Merged

derekross merged 1 commit into
mainfrom
modem-forced-reset-policy

Conversation

@derekross

Copy link
Copy Markdown
Owner

What changes

Since 0.1.13, a boot that ended with the modem up and no clean stop made the next boot skip the modem once. For an owner that is the wrong way round: the phone freezes, they hold the keys, and it comes back unable to call or text, with nothing on the screen saying why, until a second restart.

Now the first unclean end is counted and the modem starts on that boot all the same. Unchanged:

  • a second unclean end in a row holds the modem off until omamobile-modem-start --retry;
  • /etc/omamobile/modem-hold gives the hard stop from the first;
  • the bring-up latch: a hang in the bring-up itself still stops at the first.

The cost

A phone that hangs because the modem is up hangs twice before the hold, not once.

What it doesn't use

On the Pixel 7 Pro (2026-11-10) a reset forced with Volume Down + Power boots with androidboot.bootreason = "reboot", "longkey", "master_dc". The guard doesn't read it: an owner holds the keys precisely when the phone has hung, so it can't clear the modem of having caused the hang.

Tested

packages/omamobile-modem/test/all: one unclean end goes on with no refusal for the tile, the second holds, a clean stop starts the count over. The new check fails with the refusal put back. tests/packages passes.

Not yet run on a phone. The runbook steps in MODEM-PLAN that expect a refusal after the first unclean end describe 0.1.13 to 0.1.18 and need a re-run.

Not in this PR

A notification when the modem is held off. Today the Mobile data tile shows "Off after a hang or reset (held)". The version number collides with other open PRs (#403, #424) and is set at merge.

🤖 Generated with Claude Code

… its modem

Since 0.1.13 a boot that ended with the modem up and no clean stop made
the next boot skip the modem once. For an owner that is the wrong way
round: the phone freezes, they hold the keys, and it comes back unable
to call or text, with nothing on the screen saying why, until a second
restart.

The first stale modem-up marker is now counted (moved aside, as before)
and the modem starts on that boot all the same. A second unclean end in
a row still holds the modem off until a human has looked (--retry), and
/etc/omamobile/modem-hold still gives the hard stop from the first. The
bring-up latch is unchanged: a hang in the bring-up itself stops at the
first.

The cost: a phone that hangs because the modem is up hangs twice before
the hold, not once.

Seen on the Pixel 7 Pro on 2026-11-10: a reset forced with Volume Down +
Power boots with androidboot.bootreason "reboot", "longkey",
"master_dc". The guard doesn't use it: an owner holds the keys
precisely when the phone has hung, so it can't clear the modem of having
caused it.

Tests: the start decision in packages/omamobile-modem/test/all (one
unclean end goes on with no refusal for the tile; the second holds; a
clean stop starts the count over); the new check fails with the refusal
put back. Not yet run on a phone.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
@derekross

Copy link
Copy Markdown
Owner Author

Run on my Pixel 7 Pro, 2026-11-10, 7.3.0-rc6-omamobile-push1-gc7fb37460f84, this branch's omamobile-modem 0.1.19 installed as a package.

With the modem up, I restarted the phone with reboot-to omamobile and no stop first (the unclean end, on purpose). On the next boot:

  • the start logged "ended without a clean stop: a hang or a reset", the two journalctl pointers, and "one unclean end, so the modem starts on this boot all the same";
  • the stale marker was moved to modem-up.refused-<time>-<pid> and a new modem-up written at ONLINE;
  • the CP came ONLINE, no failed units, no line in /run/omamobile/modem-refused, no UFS error-handler pass and no I/O error.

Not run: the second unclean end in a row (the hold), which is unchanged code and covered by the package test, and the clean stop that removes the counted copy.

@derekross
derekross merged commit cd0d857 into main Nov 10, 2026
2 checks passed
@derekross
derekross deleted the modem-forced-reset-policy branch November 10, 2026 19:37
derekross added a commit to robotosaith/omarchy-mobile that referenced this pull request Nov 10, 2026
Main moved to omamobile-modem 0.1.19 (derekross#634) and this branch bumped no
version, so the squash would have shipped new contents under a published
number. Tested on a Pixel 7 Pro on 2026-11-10 as this tree plus derekross#389.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Nov 10, 2026
…unclean end

Main's 0.1.19 (derekross#634) starts the modem after one unclean end and holds it off after two in a row. The note still matters there: a reboot-to with the modem up would count as one of the two. The test now checks that a note naming another boot leaves the first unclean end counted and the second held, and that a note naming the marker's boot clears a counted copy too. AGENTS.md's stop-the-modem rule says what the note changes.
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
… its modem (derekross#634)

Since 0.1.13 a boot that ended with the modem up and no clean stop made the next boot skip the modem once. For an owner that is the wrong way round: the phone freezes, they hold the keys, and it comes back unable to call or text, with nothing on the screen saying why, until a second restart.

The first unclean end is now counted and the modem starts on that boot all the same. A second unclean end in a row still holds the modem off until omamobile-modem-start --retry, /etc/omamobile/modem-hold still gives the hard stop from the first, and the bring-up latch is unchanged. The cost: a phone that hangs because the modem is up hangs twice before the hold, not once.

A reset forced with Volume Down + Power boots with androidboot.bootreason "reboot", "longkey", "master_dc" on the 7 Pro. The guard doesn't use it: an owner holds the keys precisely when the phone has hung.

Tested on a Pixel 7 Pro on 2026-11-10 (7.3.0-rc6-omamobile-push1-gc7fb37460f84): after a reboot-to with the modem up and no stop, the next boot logged the unclean end, started the modem and reached ONLINE; a clean stop afterwards removed the counted copy.

Co-authored-by: Claude Fable 5.2 <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.

1 participant