Skip to content

Input: salixp: add gated ALIX support on Bluejay - #229

Merged
derekross merged 3 commits into
derekross:omamobilefrom
latentoperator:omamobile-bluejay-alix
Nov 15, 2026
Merged

derekross merged 3 commits into
derekross:omamobilefrom
latentoperator:omamobile-bluejay-alix

Conversation

@latentoperator

@latentoperator latentoperator commented Nov 9, 2026 •

Copy link
Copy Markdown
Contributor

Add ALIX as a separately matched variant with a bounded AP-bus ownership handshake, followed by the existing identity, CRC, system-info/repair-refusal gates and polling input. No alix_stage parameter or shared address-based USI probe hold is submitted, and USI/SPI/touch device-tree nodes remain disabled.

The header now accurately describes both variants' common probe path: assert reset, enable vdd then avdd, release reset, then check identity. ALIX requests AP bus ownership before that sequence; device-core pinctrl and SPI-controller setup precede peripheral probe.

The fresh normal-path Pixel 6a run and physical two-contact/release check have now passed in the signed integration (runtime section below). The exact tested integration is published below; the review corrections are included in the November 14 revision.

November 14 review revision

The PR now contains three commits: binding, driver and disabled board description. The reset/supply header correction is folded into the ALIX driver commit, whose subject uses Input: salixp:. The header again documents the extra reset GPIO toggle after the system-info read. The revised commit message records the accepted November 12 integration run and its limits.

The human added their own DCO sign-offs to the reviewed revision. Published head: 4092ddf029a73856491e8993bd3e6300994880e4. Signing changed no source bytes. Functional C is identical to the previous submitted head; the source delta is the restored header sentence, with the existing merged memory reservations retained by the rebase. Independent review, the signed-head policy/checkpatch and clean-merge gates, a W=1 driver build, ten Google DT builds with no new warnings, the targeted binding check and silent Bluejay dt-validate all pass. Checkpatch reports only the existing new-file MAINTAINERS reminder. Fresh published-head hosted CI passed on 4092ddf029a73856491e8993bd3e6300994880e4 (sign-offs, checkpatch, W=1 and device-tree/binding gates). No new hardware run is claimed.

When #230 follows, preserve both includes in input-then-Wi-Fi order.

Runtime update — November 12–13, 2026

November 12 fresh normal-path ALIX result:

Tested signed integration: 720a495233734e4328a09c33c095d72a8291e1c4, based on fc7086a68fcaaed8960d09f1bdf6c51c9926c759; uname -r = 7.3.0-rc6-omamobile-bluejay-review-20261111, with its matching Image, stage DTs and modules. This is an integration result containing the submitted changes, rather than a boot of each PR in isolation.

Qualified the revised path in separate t0 USI, t1 SPI-only and t2 full-ALIX boots, without the removed shared USI probe hold or alix_stage option. The final normal probe bound the revised controller/input modules with zero SPI errors and zero SPI timeouts.

The owner's physical contact check passed: initially all contacts released, two simultaneous contacts held for at least one second, then all released. The read-only observer recorded contact counts only, with no coordinates or event stream retained. This was headless (GUI_running=false), so it does not establish compositor delivery or release-to-frame latency.

This supplies the requested run of the cleaned-up path. The reset-order header correction is already in the unchanged submitted head. The exact tested integration is available for review; no Fold/SALIXP runtime or suspend qualification is inferred. These were attended headless stages with storage guards, a PID-1-owned 20-second watchdog, host logging and fresh SSH postchecks. They do not qualify the revised desktop/GPU path, native panel modesetting, suspend, endurance or a release installer. The historical graphical desktop has been restored. No kernel commit or DCO metadata changed; AUTO_WRITE=no and masked automatic boot-success remain in force.

Prior source/build validation

Validation on current upstream base fc7086a: independent source review found no remaining concrete blocker. The common integration passes policy, strict checkpatch without errors, W=1 builds of seven touched C files, ten Google device trees with no new dtbs_check warnings and four binding checks. New-file coverage warnings and one continuation-alignment check are disclosed. A full Image, Bluejay DTB and 1,828 modules also build for 7.3.0-rc6-omamobile-bluejay-review-20261111; matching dependency indexes and module resolution pass. The full build reports two format warnings in unchanged AoC ALSA code.

The human certified the final patches under DCO on November 11. Independent checks verify every author sign-off, all 22 commit objects including integration copies, and identical source trees and exact diffs before/after certification. Signing changed metadata only; the frozen source-build evidence remains applicable. Fresh published-head CI passed on 44e6d4bd3c50237feaad082a7cb4c9014de0baba (sign-offs, checkpatch, W=1 and device-tree/binding gates as applicable).

The original tested source c934e31 and known-good phone setup remain preserved. AUTO_WRITE=no, the automatic boot-success mask and the GS101 slot-marking hold remain in force. No firmware, private captures, credentials, device identifiers or calibration data are submitted. Development/source status is documented in Bluejay UPSTREAM, with historical bring-up isolated in mobile #565.

Published tested integration — November 14, 2026

The exact signed integration cited in the November 12–13 runtime report is now available on the fork's bringup/bluejay-reviewed-tested-20261111 branch, at 720a495233734e4328a09c33c095d72a8291e1c4. The commit, source tree and existing human sign-offs are unchanged; this publishes the source for inspection and adds no new runtime qualification.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris — ALIX on its own compatible keeps the Fold's SALIXP path intact, nicely done. The USI probe hold worries me: it's a knob in a driver every Tensor phone uses, keyed by an address. Could you tell us what goes wrong if USI11 configures at normal probe time, so we can fix that in the device tree instead? I'd also drop the staging parameter once stage 3 is the tested path.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. Dropping the USI probe hold and the staging parameter leaves this cleanly confined to ALIX. Once it has run on your 6a in this form, it's ready from my side. Please check that the updated header comment on reset order still matches what the SALIXP path does.

@latentoperator

Copy link
Copy Markdown
Contributor Author

November 12 fresh normal-path ALIX result:

Tested signed integration: 720a495233734e4328a09c33c095d72a8291e1c4, based on fc7086a68fcaaed8960d09f1bdf6c51c9926c759; uname -r = 7.3.0-rc6-omamobile-bluejay-review-20261111, with its matching Image, stage DTs and modules. This is an integration result containing the submitted changes, rather than a boot of each PR in isolation.

Qualified the revised path in separate t0 USI, t1 SPI-only and t2 full-ALIX boots, without the removed shared USI probe hold or alix_stage option. The final normal probe bound the revised controller/input modules with zero SPI errors and zero SPI timeouts.

The owner's physical contact check passed: initially all contacts released, two simultaneous contacts held for at least one second, then all released. The read-only observer recorded contact counts only, with no coordinates or event stream retained. This was headless (GUI_running=false), so it does not establish compositor delivery or release-to-frame latency.

This supplies the requested run of the cleaned-up path. The reset-order header correction is already in the unchanged submitted head. Leaving draft for your review; no Fold/SALIXP runtime or suspend qualification is inferred. These were attended headless stages with storage guards, a PID-1-owned 20-second watchdog, host logging and fresh SSH postchecks. They do not qualify the revised desktop/GPU path, native panel modesetting, suspend, endurance or a release installer. The historical graphical desktop has been restored. No kernel commit or DCO metadata changed; AUTO_WRITE=no and masked automatic boot-success remain in force.

@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. Without the USI probe hold and the stage parameter, this is confined to ALIX, and your 6a run with two contacts held and released is the check I wanted. Our gate passes on today's omamobile (2db2bc3), dt-validate finds nothing on gs101-bluejay.dtb, and CI is green.

Three small things before it leaves draft:

  1. The last commit, "Input: salixp: clarify shared reset and supply ordering", only rewrites the header the ALIX commit added. Kernel PRs land commit by commit, so could you fold it into "Input: salixp - add the ALIX variant…"? Please also use the file's prefix, Input: salixp:.
  2. The rewrite dropped "It also toggles the reset GPIO once, after the system-info read", and the driver still does that. That header is the list of everything the driver sends to a part that holds its own calibration, so please keep the sentence.
  3. As on the other Bluejay PRs: please push 720a4952… to your fork, so the run can be tied to this head.

When this and #230 have both landed, whichever goes second will conflict on the include line in gs101-bluejay.dts. Keep both includes, gs101-bluejay-input.dtsi and then gs101-bluejay-wifi.dtsi.

Add the distinct ALIX compatible and its active-low AP request and
acknowledgment GPIOs. Keep the SALIXP binding unchanged.

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
Match ALIX chip ID 36:48 separately from SALIXP 36:40 and negotiate
AP bus ownership with a bounded 500 ms acknowledgment before hardware
initialization. The normal path keeps the identity, CRC, system-info
and repair-refusal gates, then starts the qualified polling input path.

Drop the development stage parameter and leave the shared USI driver
unchanged. The address-based hold staged early bring-up; the saved
evidence does not establish a normal-probe hardware defect.

The revised normal path ran on the Pixel 6a in the signed integration
linked below on November 12, 2026: two contacts held and released, with
zero SPI errors or timeouts. The folded header
correction changes documentation only. The board nodes stay disabled;
this does not qualify Fold/SALIXP runtime or suspend.

Link: https://android.googlesource.com/kernel/google-modules/touch/fts_touch/+/2a0089b3c7b9f1f1dbb52a1460a26621e50f61cf/
Link: 720a495
Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
Describe the SPI/USI path, regulators, reset and AP ownership wiring
from the Bluejay and GS101 GPL trees. Keep USI, SPI and the touchscreen
disabled until a separately reviewed overlay is hardware-qualified.
No shared USI probe parameter is required by this source description.

Link: https://android.googlesource.com/kernel/devices/google/bluejay/+/refs/tags/android-16.0.0_r0.9/dts/gs101-bluejay-touch.dtsi

Assisted-by: LLM
Signed-off-by: Chris Gray <chrisgray70@gmail.com>
@latentoperator
latentoperator marked this pull request as ready for review November 14, 2026 16:06
@derekross

Copy link
Copy Markdown
Owner

Thank you, Chris. All three points are done: the header fix is folded into the ALIX commit under Input: salixp:, the sentence about the second reset toggle is back, and 720a4952… is on your fork. I compared the files by blob. The binding and gs101-bluejay-input.dtsi are identical to the tested tree. salixp.c differs only by the restored header sentence, and the board file only by #230's include. So your 6a run counts for this head. Our gate passes on today's omamobile (da171ea), and CI is green with dtbs_check and dt_binding_check.

I'll merge this one. When #230 follows, it will conflict on the include line in gs101-bluejay.dts: keep both includes, gs101-bluejay-input.dtsi and then gs101-bluejay-wifi.dtsi.

@derekross
derekross merged commit fe8689b into derekross:omamobile Nov 15, 2026
1 check passed
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