Skip to content

device/tegu: fingerprint F6c-F10 (works) and GNSS G3's third and fourth tries - #138

Merged
derekross merged 15 commits into
derekross:mainfrom
W4HO-ham:tegu-plans-fp-gnss-20261028
Oct 29, 2026
Merged

derekross merged 15 commits into
derekross:mainfrom
W4HO-ham:tegu-plans-fp-gnss-20261028

Conversation

@W4HO-ham

@W4HO-ham W4HO-ham commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

The two follow-ups from #131. FINGERPRINT-PLAN's status and results now go past F6c to F10. The calibration is accepted, the circle lights, a finger enrolls, it matches while other fingers don't, and the lock screen unlocks by fingerprint, all on the owner's integration kernel (LHBM is derekross/linux-omarchy-mobile#126; the userspace fixes are #135). MODEM-PLAN's GNSS section has G3's third try (the RF rails on) and fourth try (gnss-type = <2>). Neither cures the engine-start fault. The night that followed, G3's fifth try onward, the SPI link, the CP link hang and GNSS parked, is split out into #153 at the maintainer's suggestion. Docs only.

Follow-ups from derekross#131's review: the fingerprint plan's status moves past F6c (calibration accepted, the circle lit, a finger enrolled, matching still failing), and MODEM-PLAN records that neither the RF rails nor the AoC's gnss-type cure Kepler's engine-start fault.
…ts first frames

tg4c's post stage writes the normal LHBM bands after the effective delay, so stock captures under the normal circle; the port kept overdrive group 0 throughout. The palmtop now runs the normal bands, and re-enrolling under them is next.
…s by fingerprint

Under the panel's normal LHBM bands and a fresh enrollment, the enrolled finger matches and others don't, through the HAL and through fprintd, and the lock screen unlocks with it.
@W4HO-ham W4HO-ham changed the title device/tegu: fingerprint F6c-F9 and GNSS G3's third and fourth tries device/tegu: fingerprint F6c-F10 (works) and GNSS G3's third and fourth tries Oct 29, 2026
@derekross

Copy link
Copy Markdown
Owner

Thank you. This is a very readable record. The calibration header finding, the overdrive-only-first-frames reading of tg4c, and then F9 and F10 passing are a big step for the palmtop. Congratulations on the first fingerprint unlock on the 9a. The GNSS write-ups are just as useful: the R3/PC reading of the fault narrows things down a lot, and the next step is observation only and attended, as it should be.

A few things before I merge. They're all small edits to the text:

  1. Say which kernel the "works" depends on. The status line says "works: F0-F10 passed", but the run used 7.3.0-rc6-omamobile-tegu-s2mpg15-m1-g991042e2eff3-dirty. Two kernel pieces it relies on aren't merged anywhere yet. The DSI reads are open as drm: google: gs-dsim-cmd: Support DCS reads, and tg4c checks the panel ID with them linux-omarchy-mobile#125, and tg4c's LHBM with lhbm_od isn't in any PR yet (you said it will follow AGENTS, device/tegu: a fallback to slot A on the 9a costs the Linux root #125). Could the status say "works on the owner's kernel", and name those two pieces with their state? Please write the kernel one as derekross/linux-omarchy-mobile#125, because a bare number in this repository links to our own AGENTS, device/tegu: a fallback to slot A on the 9a costs the Linux root #125. That way nobody reads the plan and expects fingerprint on a kernel built from the published tree.
  2. fprintd "masked in the install image": which image? Nothing in this repository masks fprintd (no hit in rootfs/, tools/ or the packages). If it was your own image or a local step, a few words saying so would help. If it came from an image we build, I'd like to know, because then it affects every phone.
  3. The title and description are now behind the head. The last commit adds F9 and F10 passing, so "fingerprint F6c-F9" could become "F6c-F10", and the description's "matching still fails" is out of date. The title becomes the squash subject, so it's worth fixing.

Not blocking:

I checked the rest: no identifiers (the calibration's URL and id stay private, as the plan says), the factory calibration only went into the chroot's private mfg_data.img copy, and nothing writes mfg_data or persist. The tegu state line (on: 1080x2424@120) and the three fixes match what #135 changes. The earlier CI failure on this PR wasn't yours: Node in the CI image couldn't load libsimdjson.so.34, and other PRs pushed around the same time hit it too. CI on the new head is still running.

…was masked

Review: the status says it works on the owner's integration kernel, with LHBM open as derekross/linux-omarchy-mobile#126 and the DSI reads as derekross#125, and the fprintd mask came from the palmtop's private bring-up root. The G3 next step is on the charger.
@W4HO-ham

Copy link
Copy Markdown
Contributor Author

Thanks, all three are in the new head:

  1. The status says "works on the owner's kernel", names the integration release, and lists the kernel pieces with links: the LHBM is drm/panel: google-tg4c: Add local high brightness mode linux-omarchy-mobile#126, which turned out not to need DSI reads because it keeps the panel's own bands. The reads (drm: google: gs-dsim-cmd: Support DCS reads, and tg4c checks the panel ID with them linux-omarchy-mobile#125) were only used for the overdrive experiments.
  2. fprintd: the mask was on the palmtop's own root, a /dev/null link from the same batch as the 9a's other bring-up masks (ModemManager, NetworkManager, omamobile-aoc, omamobile-fp). It came from the private bring-up root the palmtop was installed from, not from anything we build. F10 says so now.
  3. The title and description match the head (F6c-F10).

"On the charger" is added to the G3 next step too.

The engine tracks satellites and gnssd's watchdog does the rest. The link garbles single bytes at the 6.24 MHz the SPI really runs at, and stock's OSCCLK clock, 32-bit words and auto CS break the framing.
Stock's OSCCLK clock breaks framing here. Reading in mode 1 lines every frame up on the as-ported clock, but the first bytes after each chip select still arrive damaged, and Kepler resets can take the SoC down.
Kepler resends frames, so the copies can be compared. In mode 1, long frames lose the same bytes at every 64-byte read boundary.
Holding chip select between Kepler's reads ends long frames in zeros, so Kepler wants it released between reads, as stock does.
With feedback delay 3 Kepler's frames are intact but one bit late. A per-read realignment gives clean short frames and satellites; long frames still lose data at each chip-select release.
At stock's clock Kepler's frames are intact but its bit offset drifts with every transfer, more as the clock slows. The CP link hang came back with the slower watcher, so it remains open.
The doorbell writes into the CP's BAR with no lock held across its link check, and the PCIe power-off clears pcie_powered_on outside the transmit lock, so a doorbell can land while the PHY goes into isolation.
The doorbell path is serialized with the PCIe power-off after all; no unguarded BAR access found by reading.
@derekross

Copy link
Copy Markdown
Owner

Thanks. All three points from last time are in. The status says "works on the owner's kernel" and names derekross/linux-omarchy-mobile#126 and #125 with full links. F10 says the fprintd mask came from your private bring-up root. "On the charger" is in the G3 step. The fingerprint half is ready to merge.

Since then the branch has grown by eight commits: G3's fifth to thirty-fourth tries and the CP link hang. That's a lot of careful work, and finding that the "hardfault" is gnssd's own watchdog over a lossy SPI link changes the whole picture. Thank you. Because the PR is squash-merged, though, those commits now need the same care before merge:

  1. The title is behind again. It says "GNSS G3's third and fourth tries", and the head goes to the thirty-fourth plus the link hang. The description's "next step is an attended strace" is out of date too. My suggestion is to move the eight MODEM-PLAN commits (af5d8a0 to d51c2db) to their own PR. Then this one merges now as fingerprint F6c-F10 plus G3's third and fourth tries, and the GNSS night gets its own review. If you'd rather keep them here, please retitle it, e.g. "device/tegu: fingerprint F6c-F10 (works) and GNSS G3 through the thirty-fourth try".
  2. One paragraph is in twice. "Stock's clock, tried (twenty-fourth try)" appears word for word at MODEM-PLAN lines 726 and 740 of the head. The heading "fifteenth to twenty-third tries" also now holds tries 24-34 and the CP link analysis, so a heading or two would help.
  3. The next step for the link hang. "Next, hands-free but at the price of watchdog resets: a trace at each BAR access in CPIF…" plans to reproduce SoC hangs with the modem up and nobody there. Please make it attended and on the charger, as a gated step with Expect, Stop (for example, two resets or any UFS error-handler line), Rollback and risk, and a cap on the number of resets. The tracing change it needs should be on a kernel branch before it boots.
  4. The test code behind the night's results. Several tries ran code that isn't in any public tree: the gnss_spi.ko builds with fb_delay/realign/rx_cpha/hold/osc/div parameters, the module that registered pdma1 and ran a PL330 memcpy, the module that set USI3's mux and divider behind the clock framework, the per-transfer spi->mode switch, and the DTBs with the pdma1 node and with auto CS. The plan points at ~/tegu/gnssif-realign and logs/…/g7-strace.sh, which nobody else can read. Please push them to a bring-up branch on your fork of linux-omarchy-mobile, as you did with docs/review: 34 lessons and 3 corrections from the twenty-fifth and twenty-sixth rounds #127, and link it by commit, before the next GNSS boot. Please also say how each test DTB went onto slot B (bin/omamobile push, I assume).
  5. The partition check for the night. MODEM-PLAN's own Safety section says to check efs, efs_backup, modem_userdata and persist against the backup before every modem session. The night had many modem sessions and several resets, most with the modem up (02:43 at 348 s, 03:15, 05:02, and the 69 s boot with the pdma1 node). Please record that the check passed, either before the sessions or once now, the way M0.1 records it. Linux never writes the NV, so I expect it to match, but the plan should show it.

Not blocking:

  • The two boots that reset at about 34 s, right after Bluetooth started, before any GNSS step, would be easier to find in NOTES than inside the GNSS section.
  • The rfsd restart hazard you found matches AGENTS' "never restart omamobile-modem" gotcha (from raven). If you agree, a 9a line there would be a good small follow-up.

The policy check and CI pass on the new head. I found no identifiers and no vendor files, and nothing in the new text writes persist, devinfo, the RPMB or a partition outside slot B. The RAM dump and kepler.bin stay private, as the text says.

Two kernel changes give the GNSS SPI stock's 3.072 MHz and one uncut 64-byte read per block. Mode 0 reads stay marginal, mode 1 reads lose one byte per back-to-back block; a logic analyzer is the next tool.
W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Oct 29, 2026
…erekross#127 line

The 10-28 evening entry still says the sleep hook has to power the NFC adapter off, and that fingerprint matching doesn't work. Both are out of date: the ST54L's NFC mode (linux-omarchy-mobile#131, derekross#137) makes the power-off hook unnecessary, and one try of such a hook failed to wake, and matching works since FINGERPRINT-PLAN F9/F10 (derekross#138). Point forward from both bullets so nobody builds the hook from the old note, and wrap the sentence about derekross#127 like the rest of the entry.
The ten commits with G3's fifth to forty-fourth tries, the CP link hang and GNSS parked are reviewed on derekross#153 now, so this pull request goes back to fingerprint F6c-F10 and G3's third and fourth tries: MODEM-PLAN is the file as of 6fe8313 again.
@W4HO-ham

Copy link
Copy Markdown
Contributor Author

Split as you suggested. The night's ten commits are in #153, with the twenty-fourth try in once, headings for tries 24-34 and the CP link hang, the BAR trace rewritten as an attended, on-the-charger step with a stop rule and a cap on resets, the partition check recorded (all six match the 10-28 backup), and the bring-up branch for the test code on its way. The two 34 s resets after Bluetooth started are a dated entry in NOTES there.

This one is back to fingerprint F6c-F10 plus G3's third and fourth tries. I couldn't move the branch back without a force-push from here, so the head d3a1dc0 puts MODEM-PLAN back to 6fe8313's content in one plain commit; the squash gives the same diff as the fingerprint commits alone. The description is trimmed. I'll add a 9a line to the rfsd restart gotcha in AGENTS as a separate small PR.

W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Oct 29, 2026
…ile#138 now

Kernel derekross#135 went up without a sign-off and is replaced by derekross#138 with the same two commits; the parked paragraph links the live one.
@derekross

Copy link
Copy Markdown
Owner

Thanks for splitting this so quickly. The net diff is now exactly what the title says: fingerprint F6c-F10 and G3's third and fourth tries. MODEM-PLAN is byte for byte the file from round 29. The policy check and CI pass on the new head, and the fingerprint half was already ready, so this one is good to merge.

One thing about how it lands. The branch still has the ten GNSS commits plus the commit that takes them out again, and this repository's squash merge puts every commit message into the squash body. As it stands, main would get ten GNSS commit messages under a fingerprint title. Either is fine with me:

  • drop the eleven commits yourself (reset the branch to 6fe8313 and force-push), or
  • leave it, and I'll write the squash body by hand when I merge.

I'll merge this one before #153. #153 then needs a small rebase, which I explain over there.

@W4HO-ham

Copy link
Copy Markdown
Contributor Author

Please write the squash body by hand, as you offered. I can't force-push from here today (the branch reset is the one git write my setup refuses), so the eleven commits stay on the branch; the tree is the fingerprint one, as you found.

W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Oct 29, 2026
The clock and the cut are in linux-omarchy-mobile#138, in review, not yet in the kernel; the status line and the morning's section say so. The morning's tries ran images with the night's release string, since the build fixes LOCALVERSION, so the plan names the Image files and their sha256 prefixes instead. The partition check at 10:20 came after the morning's tries, so it covers them. The private paths that were left (the realign tree on maclab, the g7 script, the drafted clock change, the offline tools on the laptop) now point at the public commits of linux-omarchy-mobile#137 and derekross#138. The tool that wrote slot B is named once, plainly: our bin/omamobile push, run by a local wrapper after bin/omamobile images.
W4HO-ham added a commit to W4HO-ham/omarchy-mobile that referenced this pull request Oct 29, 2026
…oot, S1's command and column

S2 says it turns LDO15S off on purpose for one boot against FINGERPRINT-PLAN's rule (that plan gets the matching note once derekross#138 is on main), and step 3 keeps omamobile-fp, the trusted app and storageproxyd stopped before the push and until the rollback, since F6c still writes secure objects. S1 names its command, power-ab with derekross#157's systemd suspend mode (rtcwake -m no plus systemctl suspend, so the NFC hook runs), reads charge_now, column 7, because the 9a's gauge has no charge_counter, and waits for derekross#157. Also: derekross#125's rules are being settled; the four rails read disabled with or without the unused-rail cleanup; stock's source grepped for the four rail names (DT sources and the register header only, no module asks by name); a Polling read before each run; nothing starts the modem at boot, so nothing to mask.
@derekross
derekross merged commit 4d52467 into derekross:main Oct 29, 2026
2 checks passed
derekross pushed a commit that referenced this pull request Oct 29, 2026
* device/tegu: SLEEP-PLAN, an audit of the s2idle drain

The 9a sleeps and wakes, but its drain is high. This audits why from stock's device tree and a snapshot of the palmtop: no cluster or SICD idle states, the GNSS and fingerprint rails always on in the bring-up DTB, and MIF, with a measurement runbook.

* device/tegu SLEEP-PLAN: the firmware's PSCI has no OS-initiated mode

So cluster states on zumapro would be platform-coordinated, with a composite StateID; and the core state's rejection rate, read from cpuidle.

* device/tegu SLEEP-PLAN: start from what's measured

The cable-out s2idle figure (about 140 mA), MIF in sleep (about 57 mA saved), and the cluster power-down tries of 10-28 (cluster 0 hangs; clusters 1-2 save nothing measurable) were already measured; the candidates are re-ranked on them.

* device/tegu SLEEP-PLAN: S2's image, its expectation and rollback

* device/tegu SLEEP-PLAN: the review's preconditions, S2's before and after, the code by commit

The runbook stops if slot B isn't marked successful and doesn't mark it (a devinfo write on the 9a, #125). S1 needs the NFC stop-poll hook active or polling stopped by hand, with the kernel's #131 state and the rule that stays, a Rollback line and pstore after a cycle that doesn't come back. S2 snapshots regulator_summary before and diffs it after, stops if more than the four rails changed, records stock's consumers of them (none in its DT: no phandle on any of the four, the GNSS three on the PMIC's channel mux, LDO15S a spare), keeps the modem off in that boot and stays away from a G3 session. The push wording says what bin/omamobile push writes and that only vendor_kernel_boot differs. The module, the bring-up DT and the S2 change are linked by commit on the fork's tegu-sleep-bringup branch, and the CPU_INFORM hints of 10-28 are credited to the bring-up module that wrote them over the PMU SMC. The measurements link #100's NOTES entry.

* device/tegu SLEEP-PLAN: the fingerprint stack stays stopped in S2's boot, S1's command and column

S2 says it turns LDO15S off on purpose for one boot against FINGERPRINT-PLAN's rule (that plan gets the matching note once #138 is on main), and step 3 keeps omamobile-fp, the trusted app and storageproxyd stopped before the push and until the rollback, since F6c still writes secure objects. S1 names its command, power-ab with #157's systemd suspend mode (rtcwake -m no plus systemctl suspend, so the NFC hook runs), reads charge_now, column 7, because the 9a's gauge has no charge_counter, and waits for #157. Also: #125's rules are being settled; the four rails read disabled with or without the unused-rail cleanup; stock's source grepped for the four rail names (DT sources and the register header only, no module asks by name); a Polling read before each run; nothing starts the modem at boot, so nothing to mask.
derekross pushed a commit to W4HO-ham/omarchy-mobile that referenced this pull request Oct 29, 2026
device/tegu/MODEM-PLAN.md conflicted on the status line: this branch's
file already carries everything derekross#138 added. It is resolved to this
branch's file, plus the "gnssd's custom GNSS library (2026-10-28,
omamobile-modem 0.1.8)" bullet that derekross#141 added on main, which this branch
never had and a plain checkout of its file would have dropped.
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…th tries (derekross#138)

FINGERPRINT-PLAN F6c-F10: on the Pixel 9a the under-display sensor
captures GOOD frames once touches reach the HAL in panel pixels and the
circle is lit before the capture; the enrolled finger matches and other
fingers are refused. MODEM-PLAN G3: the third and fourth tries at the
GNSS SPI link. The later GNSS tries moved to derekross#153.

Co-authored-by: W4HO-ham <87742609+W4HO-ham@users.noreply.github.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
* device/tegu: SLEEP-PLAN, an audit of the s2idle drain

The 9a sleeps and wakes, but its drain is high. This audits why from stock's device tree and a snapshot of the palmtop: no cluster or SICD idle states, the GNSS and fingerprint rails always on in the bring-up DTB, and MIF, with a measurement runbook.

* device/tegu SLEEP-PLAN: the firmware's PSCI has no OS-initiated mode

So cluster states on zumapro would be platform-coordinated, with a composite StateID; and the core state's rejection rate, read from cpuidle.

* device/tegu SLEEP-PLAN: start from what's measured

The cable-out s2idle figure (about 140 mA), MIF in sleep (about 57 mA saved), and the cluster power-down tries of 10-28 (cluster 0 hangs; clusters 1-2 save nothing measurable) were already measured; the candidates are re-ranked on them.

* device/tegu SLEEP-PLAN: S2's image, its expectation and rollback

* device/tegu SLEEP-PLAN: the review's preconditions, S2's before and after, the code by commit

The runbook stops if slot B isn't marked successful and doesn't mark it (a devinfo write on the 9a, derekross#125). S1 needs the NFC stop-poll hook active or polling stopped by hand, with the kernel's derekross#131 state and the rule that stays, a Rollback line and pstore after a cycle that doesn't come back. S2 snapshots regulator_summary before and diffs it after, stops if more than the four rails changed, records stock's consumers of them (none in its DT: no phandle on any of the four, the GNSS three on the PMIC's channel mux, LDO15S a spare), keeps the modem off in that boot and stays away from a G3 session. The push wording says what bin/omamobile push writes and that only vendor_kernel_boot differs. The module, the bring-up DT and the S2 change are linked by commit on the fork's tegu-sleep-bringup branch, and the CPU_INFORM hints of 10-28 are credited to the bring-up module that wrote them over the PMU SMC. The measurements link derekross#100's NOTES entry.

* device/tegu SLEEP-PLAN: the fingerprint stack stays stopped in S2's boot, S1's command and column

S2 says it turns LDO15S off on purpose for one boot against FINGERPRINT-PLAN's rule (that plan gets the matching note once derekross#138 is on main), and step 3 keeps omamobile-fp, the trusted app and storageproxyd stopped before the push and until the rollback, since F6c still writes secure objects. S1 names its command, power-ab with derekross#157's systemd suspend mode (rtcwake -m no plus systemctl suspend, so the NFC hook runs), reads charge_now, column 7, because the 9a's gauge has no charge_counter, and waits for derekross#157. Also: derekross#125's rules are being settled; the four rails read disabled with or without the unused-rail cleanup; stock's source grepped for the four rail names (DT sources and the register header only, no module asks by name); a Polling read before each run; nothing starts the modem at boot, so nothing to mask.
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