Skip to content

device/panther: a bring-up overlay for GNSS - #861

Open
angelrenefm wants to merge 2 commits into
derekross:mainfrom
angelrenefm:panther-gnss-overlay
Open

angelrenefm wants to merge 2 commits into
derekross:mainfrom
angelrenefm:panther-gnss-overlay

Conversation

@angelrenefm

Copy link
Copy Markdown
Contributor

The overlay that turns on the Pixel 7's GNSS receiver, for linux-omarchy-mobile#268 (the node, its pins and rails, all disabled, described as the Pixel 7 Pro's).

  • device/panther/overlays/gnss.dtso sets status = "okay" on the sub-PMIC, LDO9S-LDO11S, USI5 and its SPI, by path. That is cheetah's gnss-pmic-test.dtso and gnss-test.dtso in one. It isn't in DTB_OVERLAYS, so no Pixel 7 gets it by default.
  • device/panther/NOTES.md has a dated entry with where the description comes from and what was tested.

Tested:

Not tested: a boot with it through omamobile-boot-update (it needs #268 merged and published first).

tools/pr-check passes.

I used AI help (Claude) to work on this.

The Pixel 7 describes its BCM4775x receiver as the Pixel 7 Pro does
(one gs201-common-gps.dtsi in Google's source), so the kernel gets
cheetah's node for panther, disabled, and this overlay turns on what it
needs: the sub-PMIC, LDO9S-LDO11S, USI5 and its SPI. Not in
DTB_OVERLAYS. The dated NOTES entry has what was tested on a Pixel 7.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@derekross

Copy link
Copy Markdown
Owner

Thank you, Ángel. I built panther's device tree from linux-omarchy-mobile#268 and applied your overlay to it: it changes exactly the six status lines you list and nothing else, and it applies together with fingerprint-sensor.dtso. Using paths rather than labels and leaving it out of DTB_OVERLAYS are both right. Google's gs201-common-gps.dtsi is where you say it is.

Two additions to the NOTES entry before I merge it:

  1. The kernel release and the package version for the test. The entry says "a local kernel whose node is the same as the PR's", on 2026-11-04. Could you add that boot's uname -r and the omamobile-gnss version, and the indoor run you describe on omamobile-shell 0.2.68: a wake by the Pixel 9a's power key lights the lock screen #268 (1089 sentences, no fix) with its date and release?
  2. The blacklist. The overlay's comment says bcm4775x-gnss is blacklisted and omamobile-gnss-start loads it. That blacklist ships in omamobile-gnss, which the Pixel 7's image and device package don't install, so on a phone without it the driver would load by itself at the first boot with the overlay. Please add a few lines to the entry: install omamobile-gnss first; what the first boot with the overlay should show (the driver not loaded, the three vdd_gnss_* rails registered and off, USB and the fingerprint sensor as before); and the way back (the DTB without the overlay).

Two small things in the same push: name the kernel pull request by number (linux-omarchy-mobile#268) and say the overlay needs a kernel with it. On today's published kernel fdtoverlay refuses it, because the three rails aren't there yet. That's fine since nothing applies it by default, but it means I'll merge this together with #268, not before.

When you do boot a kernel with #268 and this overlay, a dated result with the tools/phone-smoke report would be a good follow-up. I don't need it for this merge.

…to try it

Review on derekross#861 and linux-omarchy-mobile#268: mcu_resp and nstandby are on
gph2, whose pin controller is disabled in gs201.dtsi, so the overlay turns
it on. The NOTES entry names the kernel PR, says omamobile-gnss goes in
first (its blacklist), what the first boot should show and the way back,
and gives the kernel release and omamobile-gnss versions of the
2026-11-04 run and of the 2026-11-17 indoor one.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@angelrenefm

Copy link
Copy Markdown
Contributor Author

Thank you, Derek. Both additions are in the NOTES entry. Both runs were on 7.3.0-rc6-omamobile-push1-g44048a9aa795: the 2026-11-04 outdoor one with omamobile-gnss 0.2.7 (built on the phone), and the 2026-11-17 indoor one (1089 sentences, no fix) with 0.2.9. The blacklist steps and the way back are in. The kernel PR is named by number, and the entry says the overlay needs a kernel with it (on one without, fdtoverlay refuses it). The overlay now also turns on gph2's pin controller, as you suggested on linux-omarchy-mobile#268.

I used AI help (Claude) to work on this.

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