Repository navigation
omamobile-modem 0.1.22, omamobile-modem-sit 0.3.35, device-raven 0.1.36: the modem runtime on the Pixel 6 Pro - #389
robotosaith wants to merge 16 commits into
Conversation
…p of 0.1.13 main moved under this PR: #386 made omamobile-modem.service Type=notify with the sleep inhibitor (Requires=/After= omamobile-modem-inhibit.service), the modem-boot target and the two markers in lib/omamobile-modem-start; #384, #387 and #388 brought the modem-sit keys that registered IMS on the 7 Pro on 2026-11-03. Resolved: - omamobile-modem.service keeps main's Requires=/After= and adds the overlay's Wants=omamobile-modem-nv-overlay.timer, so the timer hangs off the real unit; the stop's capture still runs in profile_stop after the CP power-off (profiles/modem.conf merged clean). - install: one loop over the five units; the .install note keeps both the boot-target line and the overlay line. - AGENTS.md section 4: main's longer "never restart" bullet (the modem-up marker), the overlay's NV bullet. - version 0.1.15: main is 0.1.13 and #389 claims 0.1.14. - MODEM-PLAN: the NV overlay section says what 2026-11-03 showed (IMS registers without NV persistence; what the overlay still buys). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d
|
Thank you for the port and for running it on raven with a smoke report. I read the props order, the watcher's notify path and the sitd except path, and they're consistent. Three things before I merge:
|
…1-03 keys (modem-sit 0.3.29, mmd 0.1.2, shell 0.2.96) main moved under this PR (#384, #386, #387, #388): modem-sit 0.3.24-0.3.27 brought attach_apn = none, ims_first, internet_cid, [profiles], carrier_manifest and ims_pcscf_from_isim, and the internet bearer's rmnet device is now found by omamobile-sit-data. Where both sides changed the same thing, main's code wins: bringup.start() is main's order with the carrier lookup in front of it (the SIM read once after the USIM is READY, what isn't typed comes from the lookup); data_network.py is main's device search with this PR's roaming check on the device found; paths.py, sit.py, modem-sit.conf and test_bringup.py merged by hand. What registered IMS and carried calls and texts on the 7 Pro on 2026-11-03 (NOTES, IMS-SHANNON 11-12) was five typed values beyond the APN, so the lookup now produces them: an override section carries attach_apn (an APN or none), ims_apn, ims_first, internet_cid and profile_<ID> lines, and "like = MCC/MNC" lets a carrier's codes share one section. ims_apn, ims_first and internet_cid default to auto; an empty [profiles] takes the override's set. Verizon's codes give apn = VZWINTERNET, attach_apn = none, ims_apn = IMS, ims_first = yes, internet_cid = 3 and stock's four profiles. The 11-02 value attach_apn = vzwims registered but never got IMS: gone, and the file says why. mmd treats none as no attach settings; the Mobile network page shows it as "Not sent: the carrier profile's own". Versions: modem-sit 0.3.29 (main 0.3.27, #389 claims 0.3.28), mmd 0.1.2 (main 0.1.1), shell 0.2.96 (main 0.2.94, #343 claims 0.2.95). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d
modem_device_check required ro.hardware=raven on every gs101 phone, so a Pixel 6 (google,gs101-oriole) or 6a (google,gs101-bluejay) with its own props.conf would have been refused. The board now comes from the first google,gs101-<board> compatible, and props.conf must name that board with platform gs101. Other SoCs return at once, as before. Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The crash notice is omamobile-modem 0.1.14, and the sitd exit lands in whichever omamobile-modem-sit release derekross#389 merges as (0.3.28 went to derekross#364). Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks for the careful read. All three are in:
Merging this first and rebasing #382 on it at 0.1.15 sounds good. One thing for #382 while you're there: raven's NV copies are ext4 (that's this PR's mount change), so the overlay's "files in the three f2fs copies" will need the same type check before it's raven-ready. |
modem_device_check required ro.hardware=raven on every gs101 phone, so a Pixel 6 (google,gs101-oriole) or 6a (google,gs101-bluejay) with its own props.conf would have been refused. The board now comes from the first google,gs101-<board> compatible, and props.conf must name that board with platform gs101. Other SoCs return at once, as before. Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The crash notice is omamobile-modem 0.1.14, and the sitd exit lands in whichever omamobile-modem-sit release derekross#389 merges as (0.3.28 went to derekross#364). Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
74b674e to
bfec539
Compare
…0: the Pixel 7 Pro's modem stack on by default (#398) * omamobile-modem 0.1.16, device-cheetah 0.1.43, omamobile-shell 0.2.98: the Pixel 7 Pro's modem stack on by default The maintainer's decision of 2026-11-03: the next release ships the modem working out of the box (calls, texts and data on the tested network, M4.6 through three reboots). The switch of M4.6, omamobile-modem-boot.target, is now on by default on cheetah, and a boot that can't or shouldn't start the modem costs as little as possible: - omamobile-device-cheetah 0.1.43 enables the target with a vendor wants link under multi-user.target.wants, as it does for the AoC and Wi-Fi units; the four units the target wants get no link of their own, and omamobile-modem itself still enables nothing (its test holds it to that; tegu and raven are untouched). Off: systemctl mask the target. The stack's packages (omamobile-modem-sit, omamobile-mmd) join cheetah's ROOTFS_PACKAGES so the release image carries them. - A phone that isn't provisioned costs nothing: the modem and inhibit units carry ConditionPathExists= for the chroot, the NV checksums and the board id (what tools/phone-provision puts there), so both are skipped, not failed, and the inhibitor isn't held. The inhibitor needs its own copy because the modem unit's Requires= pulls it in even when that unit's conditions skip it. sitd skips itself without /dev/umts_ipc0; sit-data now does too, through a device drop-in: without it a condition-skipped sitd counted as started for sit-data's Requires=, and the fallback would have failed after 60 s and restarted every 10 s all boot. The start script checks the same three paths first, for a start by hand. - The modem-up guard clears itself after one refused boot: a stale marker still refuses that boot and logs where to look (journalctl -b -1), then moves to modem-up.refused-<time>-<pid> (the last three kept) so the next boot starts the modem again. A phone that hangs with the modem up every time loses the modem on every other boot at most, never for good. /etc/omamobile/modem-hold restores the old hard stop; --retry still clears both markers; the bring-up latch keeps its hard stop. - Every refusal leaves one short line in /run/omamobile/modem-refused, and the shell's Mobile data tile shows it under its name instead of "No modem" (omamobile-shell 0.2.98, tile-state.sh): no new notification path, the journal has the rest. Versions: #372 and #380 claim device-cheetah 0.1.42, #389 omamobile-modem 0.1.14 and #382 0.1.15, main has omamobile-shell 0.2.97 (#394); hence 0.1.43, 0.1.16 and 0.2.98. Docs: RELEASES (ROOTFS_PACKAGES, decision 12 amended, checklist step 6), the release notes draft (the modem known issue becomes what works and what still needs a hand), INSTALL, DEVICES, and MODEM-PLAN M4.6b with the runbook: install, two reboots, a refused-boot drill by planting a stale marker, the hold flag. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * omamobile-modem 0.1.16, device-cheetah 0.1.43, omamobile-shell 0.2.100: review round 2, the modem-up guard bounded The review of #398 asked for three changes and a few corrections. The self-clearing modem-up guard was unbounded: a phone that hangs because the modem is up went up, hang, refused, up, hang, refused, for ever, with only a tile line on the boots in between. Now the refused copy the script already keeps is the counter: one automatic retry, then a hold. refuse_unclean keeps the marker and refuses "(held)" when a modem-up.refused-* copy exists (no clean stop since the last refusal); --stopped on this boot's marker removes the copies, so a clean stop resets the count; --retry clears both markers and the copies. The test's "four unclean boots in a row" block becomes "the second refusal holds, a clean stop resets", and M4.6b gets drill B9b for it. systemctl is-enabled reads "disabled" for a unit started by a vendor wants link alone (it reports /etc and /run links only), so B6's pass and stop lines expected the wrong answer and would have stopped a correct install; B6 now expects "enabled" before and "disabled" after the disable, and proves the link with ls -l of the /usr/lib link and systemctl list-dependencies --plain multi-user.target. The README, the target's header and the install message say the same, so nobody reads "disabled" and skips the mask. The read-only NV checksum check (M3.5) opens B6 as well as closing the runbook (AGENTS.md section 1). B8's is-active expectation is "failed inactive inactive active" (mmd runs). omamobile-shell takes 0.2.100: main is 0.2.98 after #397 and #373 is taking 0.2.99. The DEVICES cheetah fingerprint cell reads "works": #358 made it on by default. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * omamobile-modem 0.1.16: the start script's test stops after the marker decision The case "the boot after a refused one starts the modem again" (and the --retry case) ran lib/omamobile-modem-start past the marker refusals and relied on the machine to stop it: on the laptop and in CI no sleep inhibitor is held, so the script refused there before any driver. On the phone, where tools/phone-install-package builds the package as an unprivileged user, the inhibitor is held, so the test went on into the real bring-up and failed on /sys/module/pcie_aspm/parameters/policy (permission denied) and then on omamobile-vendor's "run as root". The script takes a test-only OMAMOBILE_MODEM_TEST_STOP_AFTER_CHECKS=1 (documented in its header; no unit or install step sets it) and exits 0 right after the marker decision, before the link driver's sysfs read, the inhibitor wait, the first modprobe and the profile. The two cases use it and assert the decision itself: exit 0, the stale marker moved aside with its refused copy kept, no latch written, the earlier reason left for the tile (it is removed only once every refusal has passed, which the stop point comes before). The test also checks that the stop point sits between the marker decision and the sysfs read, the inhibitor and the latch, and that no unit sets the variable. Verified: packages/omamobile-modem/test/all as uid 1000, with and without a systemd-inhibit stand-in that reports the inhibitor held (the phone's case); tools/pr-check origin/main. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * release notes: the carrier APNs come from the SIM, not by hand Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
modem_device_check required ro.hardware=raven on every gs101 phone, so a Pixel 6 (google,gs101-oriole) or 6a (google,gs101-bluejay) with its own props.conf would have been refused. The board now comes from the first google,gs101-<board> compatible, and props.conf must name that board with platform gs101. Other SoCs return at once, as before. Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The crash notice is omamobile-modem 0.1.17, and the sitd exit lands in whichever omamobile-modem-sit release derekross#389 merges as (0.3.28 went to derekross#364). Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
bfec539 to
6540758
Compare
|
Rebased onto today's main after #373, #382 and #398: omamobile-modem is now 0.1.17 and the sitd exit lands as modem-sit 0.3.30. Two things met #382's overlay: |
|
Thanks for the quick turnaround and for carrying the rebases onto #382 and #398. I checked each point against the new head: the board now comes from Before I merge:
Small things, fine in the same push:
On numbers: omamobile-modem 0.1.17 is also used by #403, and modem-sit 0.3.30 by other open PRs too. I'll set whichever lands second at merge, and I'll write the squash message myself, since the first commit still says nothing changes on cheetah. |
…urnal, §11 notes Review on derekross#389: - profile_setup: blkid finding no type no longer stops the start under set -e before the "is ..., not ext4 or f2fs" message; it prints it. - notify() says in the journal whom it notified ("notified N of M session(s)", or why none), so a crash report can confirm it. - MODEM-PLAN §11: a dated line that the sleep inhibitor is in the package since omamobile-modem 0.1.13 after all, and where the three CPIF suspend fixes it describes are pushed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Booting derekross#389's head on raven (2026-11-04, ge1e6adb69bbf): omamobile-modem is READY once CPIF's ports exist and the CP comes ONLINE about 3 s later, so omamobile-sitd, ordered only after the modem unit, started at once and exited ("CP is not ONLINE") at every boot. The wait for ONLINE that derekross#398 added is a device-cheetah drop-in, so raven never had it. device-raven now ships the same two drop-ins: sitd's ExecStartPre waits for ONLINE (up to cbd's own 120 s), and the data fallback is skipped without the modem's port. Cheetah is unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks for checking cheetah's path again. All in 4da03c5 and e97fbe7:
phone-smokephone-smoke: ravenKernel
Packages: hyprland 0.57.0-3.1, libcamera 0.7.2-4.22, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamessages 0.2.1-1, omamobile-apps 0.2.2-1, omamobile-audio 0.5.15-1, omamobile-device-raven 0.1.32-1, omamobile-firstboot 0.1.3-1, omamobile-gnss 0.2.7-1, omamobile-keyring 20261026-1, omamobile-mirror 0.1.1-1, omamobile-mmd 0.1.2-1, omamobile-modem 0.1.17-1, omamobile-modem-sit 0.3.30-1, omamobile-omarchy 0.1.31-1, omamobile-owl-widgets 0.1.2-1, omamobile-sensors 0.2.1-1, omamobile-shell 0.2.94-1, omamobile-voice 0.1.0-1, omaphone 0.2.2-1, omarchy 4.0.6-1, omatel-agent 0.2.4-2, quickshell 0.3.2.r7.g2b4501a-1 Small things:
device-raven 0.1.31 is taken by #400 too; please renumber at merge as you said. |
The runtime only knew cheetah's S5300. The Pixel 6 Pro (raven, gs101) runs the same stock rfsd and cbd in the chroot, with two differences: - its NV partitions are ext4, not f2fs, so each copy is mounted as what it is (blkid), and f2fs is only loaded when a copy needs it; - props.conf names cheetah, so on a gs101 board the profile checks that the per-phone props.conf says raven and gs101 before any daemon runs, or cbd and rfsd would take the other board's paths. omamobile-device-raven already provides omamobile-modem-policy, so the package installs on raven as it is; nothing here changes what cheetah runs (its copies are f2fs and its board check returns at once). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On raven the modem's PCIe link died 93 minutes into a boot. Nothing said so: sitd exited, mobile data had no bearer, and the next crash record's kernel lines were only this tool's own umts_boot0 opens, since the dead link's one repeated message had pushed everything else out. - A critical notification in every session when the CP leaves ONLINE: no calls, texts or data until a restart, and where the record is. - The record's kernel lines take the PCIe link driver's lines too, drop CPIF's open/release lines for this tool's ioctls, and fold a repeated message into one line and a count, so the start of a flood survives. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
When the CP crashed under it, sitd died with a traceback through the SIT read (EOFError: SIT channel closed). If the CP is no longer ONLINE after an EOFError, OSError or RuntimeError, it now exits with "the modem went away (...): calls, SMS and data are off until the phone restarts"; omamobile-modem.service has the crash record and the notification. With the CP still ONLINE the error is raised as before. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
What the modem watcher and sitd changes in this series answer: two suspends with the CP ONLINE, two crashes, and the kernel fixes on test since. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
modem_device_check required ro.hardware=raven on every gs101 phone, so a Pixel 6 (google,gs101-oriole) or 6a (google,gs101-bluejay) with its own props.conf would have been refused. The board now comes from the first google,gs101-<board> compatible, and props.conf must name that board with platform gs101. Other SoCs return at once, as before. Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The crash notice is omamobile-modem 0.1.17, and the sitd exit lands in whichever omamobile-modem-sit release derekross#389 merges as (0.3.28 went to derekross#364). Review of derekross#389. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ekross#373, derekross#398 and derekross#416 main took modem-sit 0.3.29 (derekross#373) and 0.3.30 (derekross#416), and modem 0.1.15 and 0.1.16 (derekross#382, derekross#398), while this waited: the sitd exit lands as 0.3.31 and the modem changes as 0.1.17. The NV overlay's note now says the copies are f2fs on gs201 and ext4 on gs101; it works on the mounted copies, so nothing else in it changes for raven. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
test/nv-overlay checks that the overlay is applied after the NV copies are mounted and before the checksum check, by finding the mount line in profile_setup. That line is now `mount -t "$t"` (each copy mounted as what it is, ext4 on gs101), so the search came back empty and the test ended non-zero after every check had passed. CI caught it; my local runs only read the ok lines. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…urnal, §11 notes Review on derekross#389: - profile_setup: blkid finding no type no longer stops the start under set -e before the "is ..., not ext4 or f2fs" message; it prints it. - notify() says in the journal whom it notified ("notified N of M session(s)", or why none), so a crash report can confirm it. - MODEM-PLAN §11: a dated line that the sleep inhibitor is in the package since omamobile-modem 0.1.13 after all, and where the three CPIF suspend fixes it describes are pushed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Booting derekross#389's head on raven (2026-11-04, ge1e6adb69bbf): omamobile-modem is READY once CPIF's ports exist and the CP comes ONLINE about 3 s later, so omamobile-sitd, ordered only after the modem unit, started at once and exited ("CP is not ONLINE") at every boot. The wait for ONLINE that now ships the same two drop-ins: sitd's ExecStartPre waits for ONLINE (up to cbd's own 120 s), and the data fallback is skipped without the modem's port. Cheetah is unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…s after today's main main has moved past this branch's numbers (omamobile-modem 0.1.20, modem-sit 0.3.34 and device-raven 0.1.34 have shipped), so the three packages take the next ones. Raven's MODEM-PLAN section 11 names the modem's crash notice by the release this lands in, as it already does for modem-sit's, rather than by a number that moves. device-raven's test/all keeps both the speaker-protection block from derekross#401 and this branch's sitd/sit-data drop-in block. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The watcher runs in the profile's private mount namespace, which is made at boot before the user's /run/user/<uid> tmpfs is mounted; inside it that directory stays empty, so after a crash notify() found no session bus and said nothing (review on derekross#389: unit start 14:42:49, /run/user/1000 mounted 14:43:11, the bus socket missing from inside the namespace). notify() now looks for sessions through /proc/1/root/run/user, at each call, and runs the send under nsenter --target 1 --mount when its mount namespace differs from PID 1's. The profile's own isolation is left as it is. On raven on 2026-11-11 the one notice that arrived came from a modem started by hand after login, as this predicts. The test runs notify() with the bus absent, then present: nothing is sent the first time, one send through nsenter the second. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Entering PID 1's mount namespace wasn't enough. On raven, with the modem
started at boot before login, the watcher found the session through
/proc/1/root but the send failed: omarchy-notification-send's
busctl --user, run from inside the profile's PID namespace and cgroup,
was refused ("Failed to connect to user scope bus via local transport:
Operation not permitted").
The watcher now asks PID 1 for the send: systemd-run --uid=<user> with
the session's bus address, waiting for it. PID 1 starts it in the host's
namespaces and its own cgroup, and its output reaches the journal. A
failed send logs its exit code and last line instead of nothing, and
nsenter and runuser are gone.
Tested on raven, 2026-11-11, 7.3.0-rc6-omamobile-push1-ge1e6adb69bbf:
modem started 12:51:42 with no session, the bus absent inside the
watcher's namespace, the user logged in, then the modem stopped:
"notified 1 of 1 session(s): Cellular stopped". The first version of
this fix, on the same steps, logged "notified 0 of 1" with the busctl
error above.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
89a3e3c to
4dbfd18
Compare
|
Rebased onto today's main to clear the conflict (only |
|
Thank you, this is the fix I hoped for, and you found the second half yourself: the bus lookup through PID 1's view wasn't enough, and sending through a transient unit keeps the profile's isolation exactly as it was. The test covers the case I asked for (no bus at start, a bus later), CI is green on this head, and thanks for rebasing and renumbering again. One thing before I merge, and your own test shows it. You checked the notice by stopping the modem, and it arrived: "notified 1 of 1 session(s): Cellular stopped". The watcher treats every move away from ONLINE as a crash, and a stop powers the modem off while the watcher is still polling, so a deliberate The stop already leaves a mark before it powers the modem off: the On numbers: omamobile-modem 0.1.21 and device-raven 0.1.35 are also claimed by another open pull request, so if that one lands first I'll move yours up by one at merge. I'll write the squash message with the numbers the files carry; no need to edit the title. |
|
A note on numbers, nothing for you to change in the code: #403 merged today and took When you next merge main into this branch, please take |
derekross#403 merged and shipped omamobile-modem 0.1.21 and device-raven 0.1.35, the numbers this branch also carried. The lines were equal, so the merge of main raised no conflict, and the Pixel 6 Pro's package would have kept its old build. Both take the next numbers. modem-sit stays at 0.3.35; other open pull requests claim it too, and it is set at merge. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks for catching that. Merged main and took |
|
Thanks for taking the numbers: the branch now carries One thing is still open, and I think my note about numbers the next day buried it, sorry: the stop notice from my comment of 11-11. On this head the watcher still calls The stop already leaves its mark before the power-off: When that's in I'll run it on my Pixel 7 Pro: a stop before a push (no notification, the "stopped" line in the journal) and a reboot. The send itself has run on your raven and not on a 7 Pro; once a stop no longer triggers it I have no harmless way to trigger it there, so I'll note it as not exercised on the 7 Pro rather than ask you for more. Then I'll merge, and write the squash message myself. |
The watcher took every move away from ONLINE for a crash, and a stop powers the CP off while it is still polling, so a deliberate `systemctl stop omamobile-modem` (run before every push on the 7 Pro) put "Cellular stopped: The modem crashed" on the screen once the notice could reach the session. profile_stop writes /run/omamobile-modem/intent-$INVOCATION_ID, holding the ID, before the power-off, and the watcher runs under the same invocation. It now makes the same test rfsd's job makes (a 32-hex ID, a regular file that isn't a symlink, its content the ID) and, when it holds, logs that the modem was stopped and sends nothing. The crash record on a stop stays as it is (derekross#532). test/watch-stop.py drives watch() through ONLINE then OFFLINE: with this invocation's intent nothing is sent; without one, with another ID's, or with a symlinked one, one notice. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Thanks, and no need to apologise, it was easy to miss under the numbers. Done in a0a0bf3: the watcher now makes the same test as rfsd's wrapper (a 32-hex
|
|
Thank you, this is exactly the test I had in mind, and you added the symlink case and the missing-ID case on top. I read Nothing more is needed from you. What's left is mine:
One small thing, for a follow-up if you like, not for this PR: the last case in I'll write the squash message myself. |
…0: the Pixel 7 Pro's modem stack on by default (derekross#398) * omamobile-modem 0.1.16, device-cheetah 0.1.43, omamobile-shell 0.2.98: the Pixel 7 Pro's modem stack on by default The maintainer's decision of 2026-11-03: the next release ships the modem working out of the box (calls, texts and data on the tested network, M4.6 through three reboots). The switch of M4.6, omamobile-modem-boot.target, is now on by default on cheetah, and a boot that can't or shouldn't start the modem costs as little as possible: - omamobile-device-cheetah 0.1.43 enables the target with a vendor wants link under multi-user.target.wants, as it does for the AoC and Wi-Fi units; the four units the target wants get no link of their own, and omamobile-modem itself still enables nothing (its test holds it to that; tegu and raven are untouched). Off: systemctl mask the target. The stack's packages (omamobile-modem-sit, omamobile-mmd) join cheetah's ROOTFS_PACKAGES so the release image carries them. - A phone that isn't provisioned costs nothing: the modem and inhibit units carry ConditionPathExists= for the chroot, the NV checksums and the board id (what tools/phone-provision puts there), so both are skipped, not failed, and the inhibitor isn't held. The inhibitor needs its own copy because the modem unit's Requires= pulls it in even when that unit's conditions skip it. sitd skips itself without /dev/umts_ipc0; sit-data now does too, through a device drop-in: without it a condition-skipped sitd counted as started for sit-data's Requires=, and the fallback would have failed after 60 s and restarted every 10 s all boot. The start script checks the same three paths first, for a start by hand. - The modem-up guard clears itself after one refused boot: a stale marker still refuses that boot and logs where to look (journalctl -b -1), then moves to modem-up.refused-<time>-<pid> (the last three kept) so the next boot starts the modem again. A phone that hangs with the modem up every time loses the modem on every other boot at most, never for good. /etc/omamobile/modem-hold restores the old hard stop; --retry still clears both markers; the bring-up latch keeps its hard stop. - Every refusal leaves one short line in /run/omamobile/modem-refused, and the shell's Mobile data tile shows it under its name instead of "No modem" (omamobile-shell 0.2.98, tile-state.sh): no new notification path, the journal has the rest. Versions: derekross#372 and derekross#380 claim device-cheetah 0.1.42, derekross#389 omamobile-modem 0.1.14 and derekross#382 0.1.15, main has omamobile-shell 0.2.97 (derekross#394); hence 0.1.43, 0.1.16 and 0.2.98. Docs: RELEASES (ROOTFS_PACKAGES, decision 12 amended, checklist step 6), the release notes draft (the modem known issue becomes what works and what still needs a hand), INSTALL, DEVICES, and MODEM-PLAN M4.6b with the runbook: install, two reboots, a refused-boot drill by planting a stale marker, the hold flag. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * omamobile-modem 0.1.16, device-cheetah 0.1.43, omamobile-shell 0.2.100: review round 2, the modem-up guard bounded The review of derekross#398 asked for three changes and a few corrections. The self-clearing modem-up guard was unbounded: a phone that hangs because the modem is up went up, hang, refused, up, hang, refused, for ever, with only a tile line on the boots in between. Now the refused copy the script already keeps is the counter: one automatic retry, then a hold. refuse_unclean keeps the marker and refuses "(held)" when a modem-up.refused-* copy exists (no clean stop since the last refusal); --stopped on this boot's marker removes the copies, so a clean stop resets the count; --retry clears both markers and the copies. The test's "four unclean boots in a row" block becomes "the second refusal holds, a clean stop resets", and M4.6b gets drill B9b for it. systemctl is-enabled reads "disabled" for a unit started by a vendor wants link alone (it reports /etc and /run links only), so B6's pass and stop lines expected the wrong answer and would have stopped a correct install; B6 now expects "enabled" before and "disabled" after the disable, and proves the link with ls -l of the /usr/lib link and systemctl list-dependencies --plain multi-user.target. The README, the target's header and the install message say the same, so nobody reads "disabled" and skips the mask. The read-only NV checksum check (M3.5) opens B6 as well as closing the runbook (AGENTS.md section 1). B8's is-active expectation is "failed inactive inactive active" (mmd runs). omamobile-shell takes 0.2.100: main is 0.2.98 after derekross#397 and derekross#373 is taking 0.2.99. The DEVICES cheetah fingerprint cell reads "works": derekross#358 made it on by default. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * omamobile-modem 0.1.16: the start script's test stops after the marker decision The case "the boot after a refused one starts the modem again" (and the --retry case) ran lib/omamobile-modem-start past the marker refusals and relied on the machine to stop it: on the laptop and in CI no sleep inhibitor is held, so the script refused there before any driver. On the phone, where tools/phone-install-package builds the package as an unprivileged user, the inhibitor is held, so the test went on into the real bring-up and failed on /sys/module/pcie_aspm/parameters/policy (permission denied) and then on omamobile-vendor's "run as root". The script takes a test-only OMAMOBILE_MODEM_TEST_STOP_AFTER_CHECKS=1 (documented in its header; no unit or install step sets it) and exits 0 right after the marker decision, before the link driver's sysfs read, the inhibitor wait, the first modprobe and the profile. The two cases use it and assert the decision itself: exit 0, the stale marker moved aside with its refused copy kept, no latch written, the earlier reason left for the tile (it is removed only once every refusal has passed, which the stop point comes before). The test also checks that the stop point sits between the marker decision and the sysfs read, the inhibitor and the latch, and that no unit sets the variable. Verified: packages/omamobile-modem/test/all as uid 1000, with and without a systemd-inhibit stand-in that reports the inhibitor held (the phone's case); tools/pr-check origin/main. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d * release notes: the carrier APNs come from the SIM, not by hand Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…rst (derekross#410) CI's test job took 934 s on derekross#389, three suites most of it in a row (omamobile-audio 316 s, omamobile-shell 263 s, omamobile-modem-sit 138 s). The suites now run up to OMAMOBILE_TEST_JOBS at once (default: the CPUs, at most 8; 1 runs them one at a time), the slow ones first, each suite's output printed together when it ends. A result is written aside and moved into place, so a report never reads it half written; a suite that ends without a result fails as "not ok - NAME (no result)"; and each suite stops after OMAMOBILE_TEST_TIMEOUT seconds (default 1800) and fails with its log printed, since output is held until a suite ends.
…ses the sleep block (derekross#424) A modem start that is refused before any link driver loads (a stale marker, the bring-up latch, a phone that isn't provisioned) kept the "don't sleep with the modem up" block for the whole boot, although no modem was up. It now releases the block, so the phone can sleep that boot. A link module that fails to load is a refusal with a reason for the Mobile data tile, where the script used to leave through set -e with modprobe's status. A successful start is unchanged. Tested by the contributor on a Pixel 6 Pro. Tested on a Pixel 7 Pro on 2026-11-10 (7.3.0-rc6-omamobile-push1-gc7fb37460f84, this change with derekross#389): with a bring-up marker placed by hand the next boot refused, logged "no modem driver loaded: sleep is no longer blocked this boot", held no inhibitor and loaded no link module; removing the marker and starting the unit brought the CP ONLINE with the block back. A normal boot before it: ONLINE, a call and a text each way.
What and why
The modem runtime only knew the Pixel 7 Pro. On the Pixel 6 Pro (raven) its NV partitions are ext4, not f2fs, so on main it stops at the efs mount before any daemon runs. omamobile-modem now mounts each NV copy as what it is, and on a gs101 board checks that the per-phone props.conf says raven before cbd and rfsd start. Two smaller fixes from running it daily: the user is told when the modem crashes, and sitd ends in one line instead of a traceback when the CP goes away. raven's MODEM-PLAN gains section 11 (the link at system suspend), which these follow.
Device and build
uname -r): 7.3.0-rc6-omamobile-push1-ge1e6adb69bbf (raven's integration kernel: CPIF for gs101 isn't published yet)Checks (no phone needed)
tools/pr-check origin/mainpassestests/allpassestests/packagespasses (or the packages I touched: omamobile-modem, omamobile-modem-sit, on every commit)versionbumped, and the new version is in the commit subjectCo-Authored-By:orAssisted-by:trailer, never aSigned-off-by:On the phone
profile_checkintoprofile_setup(after the NV tmpfs and the modem image are mounted; it won't fire, sinceCONFIG_F2FS_FS=mis in the shared fragment), and the crash record now also keeps the Wi-Fi host'spcie:lines. Two things are new on the 7 Pro too: a modem crash raises a "Cellular stopped" notification, and sitd exits with one line when the CP goes away. The smoke report's CPIF lines are the unpublished kernel's: a channel nobody opened while sitd is off, and an optional GPIO raven's DT doesn't name.userdata, weakens the storage guard, or reads registers from userspace (AGENTS.md §1)tools/phone-smokereportphone-smoke: raven
Kernel
7.3.0-rc6-omamobile-push1-ge1e6adb69bbf, tools at0d328a0, 2026-11-03T11:17Z.Packages: hyprland 0.57.0-3.1, libcamera 0.7.2-4.22, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamessages 0.2.1-1, omamobile-apps 0.2.2-1, omamobile-device-raven 0.1.30-1, omamobile-firstboot 0.1.3-1, omamobile-gnss 0.2.7-1, omamobile-keyring 20261026-1, omamobile-mirror 0.1.1-1, omamobile-mmd 0.1.1-1, omamobile-modem 0.1.17-1, omamobile-modem-sit (number set at merge)-1, omamobile-omarchy 0.1.31-1, omamobile-owl-widgets 0.1.2-1, omamobile-sensors 0.2.1-1, omamobile-shell 0.2.94-1, omamobile-voice 0.1.0-1, omaphone 0.2.2-1, omarchy 4.0.6-1, omatel-agent 0.2.4-2, quickshell 0.3.2.r7.g2b4501a-1
🤖 Generated with Claude Code