Skip to content

omamobile-device-husky 0.1.13: the Bluetooth controller gets the phone's address - #752

Merged
derekross merged 1 commit into
derekross:mainfrom
gitColinX:husky-bt-addr
Nov 15, 2026
Merged

derekross merged 1 commit into
derekross:mainfrom
gitColinX:husky-bt-addr

Conversation

@gitColinX

@gitColinX gitColinX commented Nov 14, 2026 •

Copy link
Copy Markdown
Contributor

The 8 Pro's BCM4398 comes up on its placeholder address (43:98:D0:..), the same on every unit, as #199 and #722 noted. Meanwhile the bootloader writes the phone's own address into /chosen/config/bt_addr at every boot, as it does on the 7 Pro and the 9a. This adds the 9a's address step to the 8 Pro's device package.

What changes

  • bin/omamobile-bt-addr, vendor/systemd/system/omamobile-bt-addr@.service and vendor/udev/rules.d/80-omamobile-bt-addr.rules: the 9a's three files, with BCM4398 in place of BCM4383 and the 8 Pro's placeholder in the comment. The rule starts the unit when an hci_uart_bcm controller appears. The script powers the controller on once, sets the public address with btmgmt, waits for the index to come back, and checks that the address took. The address is read at run time; nothing about it is in the package.
  • install stages them. test/all checks they're staged and that the script reads the bootloader's DT. Run from the repository, it also checks that the three files match the 9a's apart from the chip name; it skips that check when the package is built alone, as the cheetah comparisons do.
  • README row, version 0.1.13.

Tested

  • test/all passes in the repository and as a lone copy of the package (as tools/phone-install-package builds it).
  • tools/pr-check origin/main: all ok.
  • On my unit (G1MNW), published omamobile at 82f31d3a7, the package built and installed on the phone with tools/phone-install-package:
    • /chosen/config/bt_addr is present, colon-separated hex, unicast and globally administered;
    • the controller's driver chain includes hci_uart_bcm, which the rule matches;
    • before: the controller was on the 43:98:D0 placeholder; systemctl start omamobile-bt-addr@hci0 finished cleanly; after: the controller's address matches the bootloader's, and it is powered.
    • As expected, the pairing made on the placeholder address is gone (BlueZ keeps pairings per adapter address), so devices paired before this have to be paired once more.
    • Then one restart: udev started omamobile-bt-addr@hci0 at 4.06 s and it finished at 7.37 s; the controller came up powered on the bootloader's address; running, no failed units.
    • On the new address, the AirPods Pro paired again from the Settings page, connected on a2dp-sink, and a quiet tone played through them was heard.

Not tested

  • Calls over Bluetooth (the Handsfree profile).

Related

  • extract-vendor-firmware: husky follow-ups from #722 #746 adds a README line warning that pairings made on the placeholder won't survive the address step. If this merges first, that line can go; I'll update whichever lands second.
  • The Bluetooth row in ROADMAP and DEVICES is Lef's, so I haven't touched it here.

…e's address

The BCM4398 comes up on its placeholder address (43:98:D0:..), the same
on every 8 Pro, while the bootloader writes the phone's own address into
/chosen/config/bt_addr at every boot, as on the 7 Pro and the 9a. This
takes the 9a's omamobile-bt-addr, its udev rule (hci_uart_bcm) and its
unit, with the chip named: the script powers the controller on once,
sets the public address with btmgmt, waits for the index to come back
and checks the address. The test compares the three files with the 9a's
when run from the repository.

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

Copy link
Copy Markdown
Owner

Thank you, Colin. This is the right way to bring the address step over: the 9a's script, unit and rule with only the chip's name changed, a test that keeps them that way from the repository and still lets the package build alone, and a phone run that covers the boot path, the address check and a pairing on the new address. I checked the three files against the 9a's (only the names in comments differ), ran the package's test on main with this merged, and found no address in the diff or the description beyond the shared placeholder. The version is free and CI is green, so I'll merge it.

One thing for later, which isn't yours to fix here: the rule's KERNEL=="hci[0-9]*" also matches the per-connection devices (hci0:1 and so on), so each Bluetooth connection starts the unit once more and leaves a failed omamobile-bt-addr@hci0:N behind. The 7 Pro, 6 Pro and 9a have the same rule, and #423 fixes all three with ENV{DEVTYPE}=="host"; I'll make sure husky's copy goes into that one as well, so the four stay identical.

On #746: once this is in, the README paragraph there about pairing again "until this package has its own Bluetooth address step" can go, as you offered.

@derekross
derekross merged commit 89983f4 into derekross:main Nov 15, 2026
2 checks passed
gitColinX added a commit to gitColinX/omarchy-mobile that referenced this pull request Nov 15, 2026
derekross#752 gave the package its Bluetooth address step, so the paragraph
that said a pairing wouldn't survive "until this package has its own
address step" now says that pairings made before 0.1.13 have to be
made again once. The Wi-Fi line named kernel derekross#156 as a draft; it
merged through kernel derekross#187 on 2026-11-09.

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