Skip to content

arm64: dts: apple: t6050-k116s: Add the Touch ID sensor - #2204

Open
axboe wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
axboe:k116s-touchid
Open

axboe wants to merge 1 commit into
aurora-silicon:integration/aurorafrom
axboe:k116s-touchid

Conversation

@axboe

@axboe axboe commented Nov 18, 2026

Copy link
Copy Markdown
Contributor

Thomas, this is the node from your K116s notes, unchanged. The K116s has the J714s's Mesa sensor: spi2, chip select 0, 8 MHz, sensor-id 0x3356, power on NUB GPIO 21, data ready on AP GPIO 138 (checked against this machine's ADT). Pin 138 leaves the reserved ranges.

Tested on a K116s (firmware 27.2), on integration/aurora 0abc2f3 plus my K116s series, with your libfprint 1.94.100+aurora13 and fprintd 1.94.5+aurora4 packages:

  • The sensor binds and capture is interrupt-driven.
  • aurora-touchid-setup builds both files from the factory records.
  • Enroll and verify work; the enrolled finger is restored after a reboot.
  • A verify matches after a deep sleep and wake, without soc: apple: sep: calibrate the Touch ID sensor again after a system sleep #2202.
  • 3 of 3 deep sleeps resume with the node enabled.
  • xART guard refusals stay at 0.

Not tested: wrong-finger rejection and lockout.

The K116s has the J714s's Mesa sensor: spi2, chip select 0, 8 MHz,
sensor-id 0x3356, powered by NUB GPIO 21, data ready on AP GPIO 138
(ADT /arm-io/spi2/mesa). Describe it as on the J714s and take pin 138
out of the reserved ranges.

The node is Thomas Glanzmann's, from his K116s notes.

Suggested-by: Thomas Glanzmann <thomas@glanzmann.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
@RyanTheTide

Copy link
Copy Markdown
Contributor

Thanks, this is a clean change. The node matches the J714s block, the reserved-range edit opens only pin 138, and the K116s is already in the T6050 SEP profile. So the PR is device tree only. We are also moving the SEP profile choice to the SoC compatible, after which no T6050 machine needs a board admit.

Your test list already covers most of the hardware row. To close it, could you add these, at this PR's head:

  1. After each of the three deep sleeps, a fprintd-verify result on the first try. Your notes show one match after one wake, so we would like 3/3, plus resume dmesg free of SEP, SPI and DART errors.
  2. A fprintd-verify with a finger that is not enrolled, giving verify-no-match. Lockout can wait.
  3. The cold-boot SEP lines (dmesg | grep -i sep: ready time, endpoint count, sensor registered with interrupt-driven capture).

Pasting the command output is enough. After that we will run no-regression rows on our A18 Pro (Neo) and M4 Pro machines. The change is board-local, so we expect no change there.

@axboe

axboe commented Nov 19, 2026

Copy link
Copy Markdown
Contributor Author

Ryan, thanks. Here is the rest of the row. All of it is from a K116s (firmware 27.2) running this commit on integration/aurora 0abc2f3 plus my K116s series; I have not run the PR's head on its own.

1. Verify after a deep sleep. From the journal of 2026-11-18: six deep-sleep wakes, and the first verify after each one matched.

Wake First verify after it
09:59:18 match at 10:02:03
10:16:46 match at 10:16:59
10:47:53 match at 10:57:41
11:17:47 match at 11:19:03
11:25:27 match at 11:25:41
11:42:02 match at 11:45:31

Over every boot since the node went in, journalctl -p warning has two kernel lines that mention SEP, SPI, the sensor or a DART, both from the first boot, before a finger was enrolled ("at least one catacomb/lockout file is absent", "matching unavailable this boot"). Nothing at a resume.

2. A finger that is not enrolled.

$ fprintd-verify
Using device /net/reactivated/Fprint/Device/0
Listing enrolled fingers:
 - #0: right-index-finger
Verify started!
Verifying: right-index-finger
Verify result: verify-no-match (done)

and the driver at the same time:

apple_sep 502600000.sep: verify: no enrolled finger matched the capture (enclave answered user -1, Linux enrolments use user 1000)
apple_sep 502600000.sep: verify: no match; the match operation stays open for the next touch of this authentication

3. Cold boot.

apple_sep 502600000.sep: SEP platform profile: T6050/J714s, K116s, J716c
apple_sep: xART guard level 1: 10 session records exist, 0 created by Linux; the other 10 are protected
apple_sep 502600000.sep: attach: 8 endpoints advertised in 20 messages
apple_sep 502600000.sep: ready in 1443 ms: 13 endpoints (cntl sbio scrd sse hdcp xars sks xarm hibe pnon skdl stac srep), key store open
apple_sep 502600000.sep: sbio: registering sensor at bring-up stage 0
apple-mesa spi0.0: sep sensor: data-ready IRQ 152 configured for interrupt capture
apple_sep 502600000.sep: sensor: interrupt-driven capture

The enrolled finger is restored at every boot ("Touch ID: restored 1 enrolled identity/identities; enrolment survived reboot").

@RyanTheTide

Copy link
Copy Markdown
Contributor

Thanks, that completes the row: first-try matches after six deep-sleep wakes, verify-no-match with an unenrolled finger, and a clean cold boot. Since the PR is device tree only, results from the same commit on top of your K116s series are fine. We are putting it up for review.

For reference, the SoC-based SEP profile choice we mentioned is now #2311. Once it lands, the K116s entry in the profile list is no longer needed, and nothing changes for this PR.

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