Skip to content

arm64: dts: exynos: google: gs201-panther: add Bluetooth - #196

Merged
derekross merged 1 commit into
derekross:omamobilefrom
nesmeck:omamobile-gs201-panther-bt
Nov 4, 2026
Merged

derekross merged 1 commit into
derekross:omamobilefrom
nesmeck:omamobile-gs201-panther-bt

Conversation

@nesmeck

@nesmeck nesmeck commented Nov 4, 2026

Copy link
Copy Markdown
Contributor

Bluetooth for the Pixel 7. The board file enables the BCM4389's Bluetooth core on the APM USI1 UART, wired as on the Pixel 7 Pro:

  • The pieces:
    • cmu_apm (the read-only subset, for the UART clocks);
    • usi_apm1 and serial_apm1;
    • the controller as brcm,bcm4389-bt, at max-speed 3000000.
  • The pins: BT_REG_ON on gpp16-2 and BT_DEV_WAKE on gpp16-3, both outputs with no pull, and the host wake on gpa6-6, an EINT pulled down, rising edge. They match the Pixel 7's live device tree (goog,nitrous's shutdown, device-wakeup and host-wakeup GPIOs) and the Pixel 7 Pro's.
  • Separate from Wi-Fi: WL_REG_ON is gph2-4, so Bluetooth powers up alone. Wi-Fi comes in angelrenefm's arm64: dts: exynos: google: gs201-panther: add Wi-Fi #195, which will sit on top of this, as agreed on omarchy-mobile#342.

Tested on a Pixel 7, step by step (BT0 to BT3 of device/panther/BLUETOOTH-PLAN.md, which comes with the package side in a later omarchy-mobile PR). First on kernel 7.3.0-rc6-omamobile-push1-gbb6c8a7ff379 (the running kernel plus this commit), then on the kernels since (now 7.3.0-rc6-omamobile-push1-gdb341d1e29e7):

  • The UART and the controller:
    • serial_apm1 registers as ttySAC0 (181b0000), with the serdev device serial0-0;
    • hci_uart_bcm binds and reads "BCM4389C1";
    • the phone's own patch, "BCM4389C1 ES1PX_GG_P10 FW:1c340ab760", downloads in about 6.6 s with btbcm.patchram_nowait=1;
    • btmgmt info shows version 11 (Bluetooth 5.2), Broadcom, BR/EDR and LE.
  • The address: the Pixel 7 Pro's omamobile-bt-addr, run by udev at boot, sets the bootloader's /chosen/config/bt_addr.
  • Devices:
    • a Logitech MX Master 3 over BLE drives the shell, with UHID and HID++ (omarchy-mobile#304);
    • Sony WF-1000XM5 headphones play over A2DP, and their buttons work over AVRCP.
  • At boot: Bluetooth comes up on its own before the shell, and the shell's panel works.
  • Unchanged: the display, touch, the charger, the gauge and the USB rescue.
  • Known lines: six "Failed to read codec capabilities (-22)" after each power-on, as on the Pixel 7 Pro.

Not tested: calls over the headset profile (SCO), suspend with Bluetooth on, and power.

Checks: tools/kernel-pr-check --build OUT --merge origin/omamobile passes everything:

  • a clean merge onto 5df2fd3, which has the 7a's Bluetooth;
  • the sign-off;
  • checkpatch --strict with no warnings;
  • 9 device trees with W=1 and dtbs_check, with no new warnings.

🤖 Generated with Claude Code

Enable the BCM4389's Bluetooth core on the APM USI1 UART, as on the
Pixel 7 Pro: CMU_APM for the UART clocks, USI_APM1 and its UART, and the
controller with BT_REG_ON and BT_DEV_WAKE on gpp16-2 and gpp16-3 and its
host wake on gpa6-6, rising edge. The pins and their pulls are those of
the Pixel 7's live device tree (goog,nitrous: shutdown, device-wakeup and
host-wakeup GPIOs), which match the Pixel 7 Pro's.

Assisted-by: Claude Opus 5.5
Signed-off-by: Alejandro Garcia <skipper.pelygro@gmail.com>
@derekross
derekross merged commit 0a6f97a into derekross:omamobile Nov 4, 2026
1 check passed
@derekross

Copy link
Copy Markdown
Owner

Thank you, @nesmeck. The nodes are the 7 Pro's exactly, and the pins match Google's own Pixel 7 tree: in kernel/devices/google/gs201 (branch android-gs-pantah-6.1-android16), gs201-common-bluetooth.dtsi has BT_REG_ON on gpp16-2, BT_DEV_WAKE on gpp16-3 and the host wake on gpa6-6 with the same pulls, and the Pixel 7's and Pixel 7 Pro's common files both include it (through gs201-common.dtsi). That's a citation anyone can check, if you'd like to add it to the message.

Checks: tools/kernel-pr-check --build --merge passes, CI is green, and dt-validate on my side gives 0 messages for this commit and for this commit together with #195 (a deliberately broken copy was rejected, so the check ran). The two merge cleanly in either order.

On timing: as we agreed on #342, the package side should land with or before the kernel enable. Without omamobile-bt-addr, the controller runs on the chip's placeholder address rather than the phone's own, since the kernel doesn't treat the BCM4389's default as invalid. I've merged this first and accepted the placeholder in between, since it's harmless on one phone; the device-panther update with the address unit (with #423's fixed rule) and BLUETOOTH-PLAN can follow whenever it's ready.

One small thing if you push again: the kernel's coding-assistants.rst gives the trailer as Assisted-by: LLM [TOOLS], so Assisted-by: LLM rather than the model name.

Thanks, this makes the Pixel 7's radios a clean pair of commits.

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