Repository navigation
omamobile-gnss 0.2.7: Location on the Pixel 9a - #114
Conversation
0.2.3 made bionic/prebuilt/omamobile-gnss-client mandatory in install and in the package test, but the bionic prebuilts are built by hand and never committed, so a build machine whose prebuilts predate 0.2.3 can no longer package omamobile-gnss. The client is now optional in both: the package builds without it (and says so), and the gnss-slsi profile's check already names it when it's missing.
The Location switch ran only the Pixel 7 Pro's path: omamobile-gnss-start loads the Broadcom driver and starts the gnss profile, the NMEA bridge reads that profile's FIFO, and the service needs its image. On the 9a, GNSS is the modem's Kepler core behind the gnss-slsi profile, so nothing reached GeoClue. The start script now runs gnss-slsi on the 9a (once gnssd listens) with its NMEA to the FIFO, stop picks the profile before the hang marker is looked at, the bridge reads whichever profile's FIFO exists, the units accept either image, and the CLI reports the 9a's image and leaves the Broadcom orbit-prediction timer off there.
|
Thanks for the early look, @W4HO-ham. I read through the Pixel 7 Pro path: Since this changes the 7 Pro's |
|
With #153 parking GNSS on the 9a until the SPI link question is settled, do you want to keep this as a draft, or close it for now and reopen when G3 resumes? Either is fine with me. If you keep it, the |
|
Keep it as a draft, please; it comes back with G3 and takes whatever omamobile-gnss number is free then. I've marked it so. |
The matrix and the Wanted list against what has merged since #140, with each pull request's state checked on GitHub today: - Pixel 7 (panther): the GPU and haptics nodes (kernel #129, #130) were closed because they landed unchanged through kernel #134, merged with the touchscreen node and s2idle by default; device-panther 0.1.1 (#145) carries the GPU's always-on rule, the feedbackd rule and the sleep.conf drop-in. GPU and Haptics stay "testing" and now cite #134, #145 and the 2026-10-28 notes. Touch stays WIP: kernel #134's body records touches driving Hyprland and the PIN screen, but the driver fixes (kernel #133) and the kernel option with the touch notes (#149) are still in review. Suspend: "no" with the s2idle default named, since the notes say it hasn't run. - Pixel 6 Pro (raven): Haptics "no" -> "works": HAPTICS-TORCH-PLAN H5 passed (2026-10-28/29), cs40l25-vibra and the node merged as kernel #114, feedbackd's rule in device-raven 0.1.27 (#113). Display names its open kernel PR, #116. - Pixel 7a (lynx): Bluetooth "no" -> WIP: the merged WIFI-PLAN (#133) records it scanning through its own UART; kernel #98 and #98 are in review. The Wanted line says so too. - Pixel 7 Pro (cheetah) detail: Boot names the AVB footers a chained vbmeta_b needs (NOTES 2026-10-28 and 10-29, #104); Suspend describes the merged tier 1 auto-sleep (#64, off by default behind the omamobile-sleep toggle, P0-P5 on 2026-10-26) in place of the old "until S8" text. Nothing from the open #171, #172, #174-#176. - Wanted: coral's #52 is merged, so its line no longer says "in review"; flame gets a line for its native install in review (#55, #56, #58, #59, split out of #17); the OnePlus 13 (dodge) is listed with #111 and kernel #115, both open. The Pixel 9a row and the Pixel 9 family line are left to #165. Every status word still parses with the website's parseCell. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…base # Conflicts: # packages/omamobile-gnss/version
|
Out of draft: the 9a's receiver gives a fix now, so this path is tested end to end. Merged main (0.2.5 after your 0.2.4); pr-check and the package test pass. What ran tonight on the palmtop, with the package's logic hand-installed as drop-ins over the 0.2.1 package (the same start, bridge and service decisions this PR makes): Two things this PR doesn't cover, both recorded in the 9a's MODEM-PLAN (#202):
The 7 Pro's path is untouched, as before. |
|
Congratulations on the first fix, and thanks for the full write-up. One thing before it merges: the 9a's path starts the session before the hang marker is written, and the modem plan says a GNSS session can still hang the SoC about 170 s after ONLINE. Since socket activation lets any GeoClue client start it, a hang could repeat at every boot. Could you write the marker before |
…n and clears it only on stop A GNSS session on the 9a can still hang the SoC with the modem up, and socket activation lets any GeoClue client start one, so without a marker a hang could repeat at every boot. The marker now goes down before gnss-slsi starts and is cleared only by --stop; the board is read through OMAMOBILE_DT as the shell's setup does, and the test runs both branches with a fake device tree and a stub chroot.
|
Done and pushed, package tests pass:
Still owed: one run from this package build on the palmtop rather than the hand-installed drop-ins. The palmtop is on another test right now; I'll paste that run here when it's done. |
The 9a's GNSS cell goes from "no" to WIP: a position fix on 2026-10-30 and Omarchy's Location switch serving it on 2026-10-31, started by hand (46 of 52 GGA sentences fixed, 9-13 satellites, first fix about 80 s after a client connects). It needs kernel #150 (merged) and #114, and a session lasts as long as the modem's PCIe link, one to five minutes, which is the open item. The record lands with #202.
0.2.5 moved ExecStop to omamobile-gnss-start --stop, which picks the profile, but the older assertion still demanded the 7 Pro's chroot stop line, so check() failed on the palmtop. It now checks that ExecStop runs the start script and that --stop ends in the profile's own stop.
|
The run you asked for, from the package build and not the hand drop-ins: omamobile-gnss 0.2.5 built and installed on the palmtop with
Two things the run found, both fixed on the branch: my hand-copied One caveat that isn't this package's: about a minute after the session ended the palmtop reset itself with nothing in pstore, the modem-link hang that #202 describes as the open item. The marker was already cleared by then, so the next boot's start wasn't skipped, which is the right outcome for a reset after a clean stop. |
|
Thanks for the run from the package build. 145 of 153 fixed in a moving car, through the real units, is exactly the evidence I asked for, and it's good the run found the conflicting hand copies and the stale Reading it again, I found one problem in how the marker and the new
So the marker holds GNSS off for one activation, not "until One way to fix it is to clear the marker in Two smaller things:
With the stop path fixed, I'll switch Location on and off on my 7 Pro, as promised, and merge. |
…on of this boot systemd runs ExecStop after a start that skipped on a leftover marker too, so --stop deleted the marker and the next GeoClue client started a full session again. The 9a's start now writes /run/omamobile-gnss/session with the marker, and --stop clears both only when that file exists.
|
Thanks for catching that. You're right that |
|
Thank you, that's the fix I had in mind, and the test covers exactly the case that worried me: a leftover marker with an empty As promised, I'll switch Location on and off on my 7 Pro next and merge if that's clean. |
…y default, fingerprint opt-in, GNSS to testing (#248) With omamobile-device-tegu 0.2.3 the charge limit, the AoC and NFC start at boot from the package, so the Pixel 9a's Charging, AoC, Speakers/mics, Sensors and NFC cells move to "works" (on by default), and Fingerprint to "works (opt-in)", as on the 7 Pro. Each was verified on a boot of its kernel pull request's head built from the shared config fragment; the record is #244's entry, which gives each boot's uname -r. The Boot cell records the 10-31 unmark and re-mark, and GNSS moves to testing with the modem started by hand (#114, #202, kernel #172). The kernel parts are in review, as each cell says. Closes #248 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
One thing from trying to build this on my 7 Pro: the new 7 Pro case in -e "s|/sys/firmware/devicetree/base/soc@0/usi@109400c0/spi@10940000/status|'"$tmp"'/no-spi5|" \\With that the suite passes on the phone. I can add it at merge if you'd rather not push again. My Location on/off run waits for a laptop build of the package (its bionic prebuilts aren't built on the phone), so it comes a little later. |
…the phone's On a Pixel 7 Pro with the GNSS overlay, the test read the real SPI5 status, found it okay and went on to load the real modules. The SPI5 path now points into the test's temporary directory like the others.
|
Thanks for running it there. Pushed as your one-liner (efe7fd5): |
|
Thanks for pushing it so quickly. efe7fd5 is exactly the line I suggested: the 7 Pro case now reads its SPI5 status from the test's temporary directory, so it takes the "nothing to do" branch on any machine and never gets near the real The only thing left is my own Location on/off run on the 7 Pro. It waits for a laptop build of the package because of the bionic prebuilts, and I'll merge once it's clean. |
# Conflicts: # packages/omamobile-gnss/version
|
#359 took 0.2.6, so I merged main in and moved this to 0.2.7 (the version file was the only conflict); the title follows. |
|
Thanks, @W4HO-ham. I compared the new head with the one I last looked at, merged onto today's main: the package is identical apart from the version, now 0.2.7 after #359, and the title follows. CI and pr-check are green. Merging; I'll do my own Location on/off run on the 7 Pro with the next build and report here if anything's off. Since the branch has merge commits, I'm writing the squash message myself. |
The matrix re-read against what merged and ran on hardware since 2026-11-02: - Pixel 7: Bluetooth and the torch move to testing (kernel #196, #197, #426); fingerprint moves to WIP (kernel #207; the package side, #458, is in review); AVB footers and reboot=warm noted. - Pixel 8a: a new row, from first light and UFS behind the guard (kernel #192, #206; #414). - Pixel 6 Pro: GPU scaling and cooling (kernel #189) and UFS suspend (kernel #188) merged. - Pixel 9a: GPU levels merged through kernel #96; GNSS's moving-car session (#114) and the link driver (kernel #185). - Pixel 7a: Bluetooth's device tree merged (kernel #98). The Pixel 8 Pro's row is left to #472. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The matrix and the Wanted list against what has merged since derekross#140, with each pull request's state checked on GitHub today: - Pixel 7 (panther): the GPU and haptics nodes (kernel derekross#129, derekross#130) were closed because they landed unchanged through kernel derekross#134, merged with the touchscreen node and s2idle by default; device-panther 0.1.1 (derekross#145) carries the GPU's always-on rule, the feedbackd rule and the sleep.conf drop-in. GPU and Haptics stay "testing" and now cite derekross#134, derekross#145 and the 2026-10-28 notes. Touch stays WIP: kernel derekross#134's body records touches driving Hyprland and the PIN screen, but the driver fixes (kernel derekross#133) and the kernel option with the touch notes (derekross#149) are still in review. Suspend: "no" with the s2idle default named, since the notes say it hasn't run. - Pixel 6 Pro (raven): Haptics "no" -> "works": HAPTICS-TORCH-PLAN H5 passed (2026-10-28/29), cs40l25-vibra and the node merged as kernel derekross#114, feedbackd's rule in device-raven 0.1.27 (derekross#113). Display names its open kernel PR, derekross#116. - Pixel 7a (lynx): Bluetooth "no" -> WIP: the merged WIFI-PLAN (derekross#133) records it scanning through its own UART; kernel derekross#98 and derekross#98 are in review. The Wanted line says so too. - Pixel 7 Pro (cheetah) detail: Boot names the AVB footers a chained vbmeta_b needs (NOTES 2026-10-28 and 10-29, derekross#104); Suspend describes the merged tier 1 auto-sleep (derekross#64, off by default behind the omamobile-sleep toggle, P0-P5 on 2026-10-26) in place of the old "until S8" text. Nothing from the open derekross#171, derekross#172, derekross#174-derekross#176. - Wanted: coral's derekross#52 is merged, so its line no longer says "in review"; flame gets a line for its native install in review (derekross#55, derekross#56, derekross#58, derekross#59, split out of derekross#17); the OnePlus 13 (dodge) is listed with derekross#111 and kernel derekross#115, both open. The Pixel 9a row and the Pixel 9 family line are left to derekross#165. Every status word still parses with the website's parseCell. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
) The 9a's GNSS cell goes from "no" to WIP: a position fix on 2026-10-30 and Omarchy's Location switch serving it on 2026-10-31, started by hand (46 of 52 GGA sentences fixed, 9-13 satellites, first fix about 80 s after a client connects). It needs kernel derekross#150 (merged) and derekross#114, and a session lasts as long as the modem's PCIe link, one to five minutes, which is the open item. The record lands with derekross#202.
…y default, fingerprint opt-in, GNSS to testing (derekross#248) With omamobile-device-tegu 0.2.3 the charge limit, the AoC and NFC start at boot from the package, so the Pixel 9a's Charging, AoC, Speakers/mics, Sensors and NFC cells move to "works" (on by default), and Fingerprint to "works (opt-in)", as on the 7 Pro. Each was verified on a boot of its kernel pull request's head built from the shared config fragment; the record is derekross#244's entry, which gives each boot's uname -r. The Boot cell records the 10-31 unmark and re-mark, and GNSS moves to testing with the modem started by hand (derekross#114, derekross#202, kernel derekross#172). The kernel parts are in review, as each cell says. Closes derekross#248 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* omamobile-gnss 0.2.4: build without the Pixel 9a's HAL client 0.2.3 made bionic/prebuilt/omamobile-gnss-client mandatory in install and in the package test, but the bionic prebuilts are built by hand and never committed, so a build machine whose prebuilts predate 0.2.3 can no longer package omamobile-gnss. The client is now optional in both: the package builds without it (and says so), and the gnss-slsi profile's check already names it when it's missing. * omamobile-gnss 0.2.5: Location on the Pixel 9a The Location switch ran only the Pixel 7 Pro's path: omamobile-gnss-start loads the Broadcom driver and starts the gnss profile, the NMEA bridge reads that profile's FIFO, and the service needs its image. On the 9a, GNSS is the modem's Kepler core behind the gnss-slsi profile, so nothing reached GeoClue. The start script now runs gnss-slsi on the 9a (once gnssd listens) with its NMEA to the FIFO, stop picks the profile before the hang marker is looked at, the bridge reads whichever profile's FIFO exists, the units accept either image, and the CLI reports the 9a's image and leaves the Broadcom orbit-prediction timer off there. * omamobile-gnss 0.2.5: the 9a writes the hang marker before its session and clears it only on stop A GNSS session on the 9a can still hang the SoC with the modem up, and socket activation lets any GeoClue client start one, so without a marker a hang could repeat at every boot. The marker now goes down before gnss-slsi starts and is cleared only by --stop; the board is read through OMAMOBILE_DT as the shell's setup does, and the test runs both branches with a fake device tree and a stub chroot. * omamobile-gnss: the stop assertion follows ExecStop to the start script 0.2.5 moved ExecStop to omamobile-gnss-start --stop, which picks the profile, but the older assertion still demanded the 7 Pro's chroot stop line, so check() failed on the palmtop. It now checks that ExecStop runs the start script and that --stop ends in the profile's own stop. * omamobile-gnss: a stop clears the 9a's hang marker only after a session of this boot systemd runs ExecStop after a start that skipped on a leftover marker too, so --stop deleted the marker and the next GeoClue client started a full session again. The 9a's start now writes /run/omamobile-gnss/session with the marker, and --stop clears both only when that file exists. * omamobile-gnss test: the 7 Pro case reads a scratch SPI5 status, not the phone's On a Pixel 7 Pro with the GNSS overlay, the test read the real SPI5 status, found it okay and went on to load the real modules. The SPI5 path now points into the test's temporary directory like the others.
The matrix re-read against what merged and ran on hardware since 2026-11-02: - Pixel 7: Bluetooth and the torch move to testing (kernel derekross#196, derekross#197, derekross#426); fingerprint moves to WIP (kernel derekross#207; the package side, derekross#458, is in review); AVB footers and reboot=warm noted. - Pixel 8a: a new row, from first light and UFS behind the guard (kernel derekross#192, derekross#206; derekross#414). - Pixel 6 Pro: GPU scaling and cooling (kernel derekross#189) and UFS suspend (kernel derekross#188) merged. - Pixel 9a: GPU levels merged through kernel derekross#96; GNSS's moving-car session (derekross#114) and the link driver (kernel derekross#185). - Pixel 7a: Bluetooth's device tree merged (kernel derekross#98). The Pixel 8 Pro's row is left to derekross#472. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Location on the Pixel 9a through Omarchy's switch. #110 is in main now, so the diff against main is this PR's own change. On the 9a the Kepler core gives fixes with the modem up (MODEM-PLAN G3), and a session from the package build fixed 145 of 153 times in a moving car (results in the thread).
Today the Location switch (
omamobile-gnss on) only works on the Pixel 7 Pro.omamobile-gnss-startloads the Broadcom driver and starts thegnssprofile, and the NMEA bridge reads that profile's FIFO. The service also needs that profile's image. On the 9a, GNSS is the modem's Kepler core behind thegnss-slsiprofile (#70), so nothing would reach GeoClue even with a fix.The changes:
google,zumapro-tegu),omamobile-gnss-startrunsgnss-slsiwith its NMEA to the FIFO, once gnssd listens on@GNSSND; otherwise it says why and exits 0.--stoppicks the profile and is handled before the hang marker, so a leftover marker never skips a stop.omamobile-gnss-nmeareads whichever profile's FIFO exists, and its unit waits for either and may open both.omamobile-gnss.serviceruns with either image (two|conditions), and stops through the start script./run/omamobile-gnss/sessionbeside it), so a leftover marker from a reset survives the skipped start and theExecStopsystemd runs after it.omamobile-gnssreports the 9a's image and doesn't turn on the Broadcom orbit-prediction timer there.The Pixel 7 Pro's path is unchanged: no tegu compatible, so the same driver load and
gnssprofile, with the first FIFO. Tests: new checks intest/allfor each piece, and the bridge's FIFO choice checked against a temporary tree with only the 9a's FIFO.tools/pr-checkis clean.