Skip to content

Detect T1 Touch ID and keep T1Bridge's fingerprint driver in setup - #14627

Open
makoni wants to merge 2 commits into
omacom:quattrofrom
makoni:t1bridge-touch-id
Open

makoni wants to merge 2 commits into
omacom:quattrofrom
makoni:t1bridge-touch-id

Conversation

@makoni

@makoni makoni commented Oct 8, 2026 •

Copy link
Copy Markdown

Summary

On 2016-2017 Touch Bar MacBook Pros (MacBookPro13,2/13,3/14,2/14,3), Touch ID works on Linux through T1Bridge, which ships its own matched libfprint-t1bridge/fprintd-t1bridge pair. Omarchy's fingerprint setup doesn't know about it, in two ways:

  • omarchy-hw-fingerprint returns false, so the Fingerprint menu entry and the first-run invitation never show up, and omarchy-setup-security-fingerprint stops at "No fingerprint sensor detected". Touch ID isn't a USB reader of its own: it sits behind the T1's Secure Enclave, inside the "iBridge" device (05ac:8600).
  • If detection did pass, setup would run pacman -S --needed --noconfirm --ask 4 -- libfprint-git fprintd usbutils. libfprint-git and fprintd conflict with T1Bridge's pair, so --ask 4 would replace it and leave the sensor with no driver.

This change:

  • Detects T1 Touch ID from sysfs. An Apple device with an interface named Apple USB SEP Interface counts as a reader. The T1 exposes that interface only in its macOS USB configuration, which T1Bridge selects at boot. In the default configuration it's absent and libfprint can't drive the sensor, so detection stays false there.
  • Keeps T1Bridge's driver. When libfprint-t1bridge and fprintd-t1bridge are installed, setup skips the package step instead of replacing them.
  • Doesn't re-enroll an existing Touch ID print. The T1 rejects a second enrollment of a finger it already holds (enroll-duplicate), and T1Bridge's own setup usually enrolls one, so a Touch ID user who already has prints goes straight to verification. Nothing changes for other readers.

The PAM and lock-screen part (setup_pam_config, omarchy-apply-lock) is unchanged and runs as before, after verification.

The Touch ID support quattro gained since this was opened is for Apple Silicon Macs. The aurora driver is built into libfprint-git only on aarch64 (the x86_64 build is unchanged, per its PKGBUILD) and needs /dev/sep-bio, which a T1 Mac doesn't have. On x86 no platform package names a T1 reader in fingerprint-readers, and the enable-fingerprint hook is registered only for Apple Silicon, so setup still authenticates sudo and resolves that hook first, which does nothing on a T1.

Testing

  • test/shell.d/hw-fingerprint-test.sh: a T1 exposing its SEP interface is detected; the same iBridge in its default configuration is not; the interface name on a non-Apple device is not.
  • test/shell.d/fingerprint-package-test.sh: with T1Bridge's pair installed, setup never calls pacman; with no prints it enrolls; with a print it skips enrollment and verifies before PAM and lock setup; a failed verification leaves PAM untouched.
  • After merging the current quattro (11 November), both new test groups still fail on the quattro scripts and pass with this change. ./test/cli, fingerprint-invitation-test.sh and security-polkit-faillock-test.sh pass. I didn't rerun the full ./test/shell suite.
  • On hardware, 11 November: MacBookPro14,3, Omarchy 4.5.0-3, linux-omarchy 7.2.15-1, T1Bridge 0.1.12 with one finger enrolled. The installed omarchy-hw-fingerprint returns 1 there and this branch's returns 0. I ran this branch's two scripts against the installed Omarchy (PATH pointing at them, OMARCHY_PATH=/usr/share/omarchy). Setup printed "Using T1Bridge's fingerprint driver for Touch ID.", made no pacman call, said "Touch ID already has an enrolled fingerprint." instead of enrolling, verification matched, and it configured the lock screen. pacman -Q showed the same packages before and after: libfprint-t1bridge and fprintd-t1bridge, and no libfprint-git.
  • Earlier, 9 October, on the branch before the merge (Omarchy 4.0.4, kernel 7.2.5-3): the lock screen unlocked with the fingerprint on the first touch after an s2idle cycle, sudo fprintd-enroll -f left-index-finger "$USER" completed, and enrolling an already-enrolled finger returned enroll-duplicate.

@greptile-apps

greptile-apps Bot commented Oct 8, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerView in GreptileConfidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable defects were found.

Summary

Adds T1 Touch ID detection and preserves T1Bridge’s installed fingerprint drivers.

  • Omarchy detects T1 Touch ID when its usable USB interface appears.
  • T1 Touch ID setup keeps its driver and reuses enrolled prints.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A[Detect fingerprint reader] --> B{T1Bridge pair installed?}
  B -->|Yes| C[Keep installed drivers]
  B -->|No| D[Install standard drivers if needed]
  C --> E{User already has a print?}
  D --> F[Enroll a fingerprint]
  E -->|Yes| G[Verify fingerprint]
  E -->|No| F
  F -->|Success| G
  G -->|Success| H[Configure PAM and lock screen]
  G -->|Failure| I[Stop without changing PAM]
Loading

Reviews (2) · Last reviewed commit: "Merge remote-tracking branch 'origin/qua..." · Reviewed by Greptile

@makoni

makoni commented Oct 9, 2026

Copy link
Copy Markdown
Author

A caveat I found after opening this, so nobody is surprised by it.

The last step of setup, omarchy-apply-lock, turns on fingerprint unlock for the lock screen, and on a T1 that currently breaks Touch ID if the machine stays locked for a while. Each lock-screen session makes T1Bridge reload its keybag into the T1's Secure Enclave, and the lock starts a new session every ~31 s while nobody is at the machine. On this MacBookPro14,3 the Secure Enclave started rejecting the keybag after 23 and 34 minutes of lock. From then on every fingerprint request failed until a full power-off. The bug itself is T1Bridge's (standardagents/t1bridge#14, details); the trigger is the lock re-arming forever, which #12461 would bound to six sessions (comment). Tuning pam_fprintd didn't help, since T1Bridge appears to end each match at 30 s on its own.

I left the code as it is. The lock screen is the right place to fix this, and #12461 would bound it for every reader, so special-casing Touch ID in setup would only paper over it. Until #12461 or a T1Bridge fix lands, the workaround on a T1 is to remove /etc/pam.d/omarchy-lock-fingerprint after setup. Detection, the kept T1Bridge packages and the skipped re-enrollment behave as described above, and sudo by fingerprint keeps working here with the lock-screen file removed.

@omarchybot

Copy link
Copy Markdown
Collaborator

Thanks for the thorough write-up and the hardware run on the MacBookPro14,3.

This pull request has not been reviewed further yet, because the two models that classify each item disagreed on what it is. Claude Opus 5.5 read it as an enhancement: on quattro nothing breaks for a T1 owner, the fingerprint menu just doesn't offer a reader that only works through T1Bridge's out-of-tree driver, so this adds support for a new driver stack. Codex Medium read it as a compatibility fix: a working Touch ID sensor on specific hardware that fingerprint setup refuses, and would leave without a driver if detection ever passed. Both readings are reasonable, and which one applies decides how it is handled — a compatibility fix gets verified as a bug, an enhancement waits for the maintainer to say they want it — so nothing was run on a worker, no labels were added, and the diff has not had a line-by-line review.

No other open pull request or issue about T1 Touch ID or fingerprint setup turned up. This is waiting on the maintainer to settle what it is.

# Conflicts:
#	bin/omarchy-setup-security-fingerprint
#	test/shell.d/hw-fingerprint-test.sh
@omarchybot omarchybot added the enhancement New feature or request label Nov 18, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants