[ooaklee]

Summary

Restore automatic module alias resolution for the MSHW0485 SPI touchscreen on
the exact released 7.2.5 v8 carrier, and allocate a distinct v9 qualification
ABI.

The released v8 driver retained its OF and HID match tables but omitted the
established SPI device-ID table. Device Tree coldplug exposes the controller
as spi:mshw0485, so modprobe -R spi:mshw0485 could not resolve the module
and a fresh boot left spi0.0 unbound. Explicitly loading
mshw0485_touch bound the controller and restored physically confirmed touch.

Change

• add the established mshw0485 SPI device ID;
• export it with MODULE_DEVICE_TABLE(spi, ...);
• attach it to the existing SPI driver's id_table;
• preserve the existing OF and HID match paths;
• allocate 7.2.5-jg-0sp11v9 for isolated build and hardware qualification.

This branch is based directly on released v8 commit
8ba2b75713feb5c72bbc11dd8cce4048e23cdeea. It deliberately does not pull the
broader 7.2.0 integration line or add a forced-module-load policy.

Validation completed

• both formatted commits pass strict kernel checkpatch.pl with 0 errors,
  0 warnings, and 0 checks;
• the SP11 integration validator passes;
• the v8-to-v9 delta contains only the driver and changelog;
• public-content and commit-message hygiene scans pass;
• current integration and known-good v23 confirm the same SPI ID pattern;
• OF matching remains first and the existing Denali/compatible probe guards
  remain unchanged;
• the legacy SPI-HID aliases do not claim spi:mshw0485.

Required before release

• build full packages and verify modinfo, modules.alias, staged depmod,
  vermagic, and signature policy;
• preserve the known-good fallback and validate normal/recovery boot entries;
• perform a fresh v9 boot without any global force-load rule;
• verify automatic binding, touch/pen input nodes, HIDRAW/IPTSD startup, and
  absence of the prior SPI-ID warning;
• physically confirm touch;
• retire the temporary v8-only workaround after first acceptance and record a
  second clean boot.

Parent parity tracker: ooaklee/linux-surface-pro-11-oe#33

Dedicated regression tracker: ooaklee/linux-surface-pro-11-oe#69

[ooaklee]

Build qualification update — 2026-10-11

The first full-package build of the v9 branch completed successfully.

• source revision: d4504d974f6ae826797c8db51ffb0a8d46909c92
• source tree: 0a4b9dcfea4bde6c5a0c2d215f0677cee4c933f0
• ABI: 7.2.5-jg-0sp11v9-qcom-x1e
• build target: binary-indep binary-qcom-x1e
• local patches: none
• exported set: image, modules, flavour headers, common headers, build
  manifest, package manifest, and closed SHA256SUMS
• all six exported checksum entries verify

The staged package/module checks confirm:

• mshw0485_touch.ko.zst is present at the exact v9 ABI path;
• module name and vermagic match the v9 ABI;
• modinfo contains spi:mshw0485, both existing OF aliases, and both
  existing SP11 HID aliases;
• generated modules.alias contains
  alias spi:mshw0485 mshw0485_touch;
• modprobe -R spi:mshw0485 resolves uniquely to mshw0485_touch;
• the touchscreen module is PKCS#7/SHA-512 signed;
• the OLED Denali DTB is present and structurally valid;
• no packaged forced-load workaround is present.

An additional all-module hardening audit found the same policy result in v8
and v9: each package contains 7,816 modules, all 7,731 non-staging modules are
signed, and the same 85 modules beneath drivers/staging/ are deliberately
unsigned by the Ubuntu packaging hook. The normalized full module-path set
and unsigned-path set are identical across the two packages. Both touchscreen
modules are signed, while v9 alone gains spi:mshw0485. This is therefore an
inherited package policy rather than a regression from this change; it is not
being represented as an all-modules-signed result.

A clean Lexr build of the supported five-package set, including exact-ABI
external-DTB boot support, is now in progress. Fresh-boot hardware
qualification remains pending, so this PR stays draft.

[ooaklee]

Local v9 qualification update (hardware validation pending)

The five-package Lexr bundle now builds cleanly from the exact PR head:

• source revision: d4504d974f6ae826797c8db51ffb0a8d46909c92
• source tree: 0a4b9dcfea4bde6c5a0c2d215f0677cee4c933f0
• ABI: 7.2.5-jg-0sp11v9-qcom-x1e
• build source: HTTPS Git branch codex/fresh-boot-kernel-7-2-5-issue
• local patches: none
• requested boot image mode: source
• effective DTB delivery: external-required
• packages: image, modules, boot support, ABI headers, and common headers

Validation passed:

• Lexr v0.5.0-rc.5 kernel inspect --package-set all
• all five package checksums
• closed eight-file bundle inventory and provenance/schema checks
• exact target module path, name, vermagic, PKCS#7/SHA-512 signature metadata, and unique spi:mshw0485 -> mshw0485_touch resolution
• both declared external DTBs and exact-ABI boot-support trigger/hooks
• no forced-load workaround in reviewed package policy paths
• 7,731 non-staging modules carry uniform signature metadata; the exact 85 unsigned staging exceptions match the reviewed pre-change policy

An independent GLM-5.2 review approved the exact two-commit delta with no required changes. The requested DeepSeek review could not run because that provider currently has no active credentials, so this is not presented as a two-model approval.

Next gate is a one-shot hardware boot with the known-good kernel kept as the persistent fallback. The first v9 boot must verify automatic binding before any manual modprobe: loaded module, SPI modalias resolution, driver symlink, input node, absence of the old missing-spi_device_id warning, physical touch/pen, and suspend/resume. This PR remains draft until that evidence is captured.

[ooaklee]

Closed as an inactive/consolidated review vehicle during the 20 October kernel queue cleanup—not merged, newly hardware-qualified or accepted for upstream by this closure.

Previously recorded v8 acceptance evidence stands unchanged; this closure only retires the review vehicle.

The original source head and contributor history remain retained. Where a closed topic branch is removed, its verified archive/pr-69-20261020 tag preserves the exact head for recovery. Released/recovery/dependency branches remain untouched.

The separate tested-source staging route and all outstanding upstream reconstruction, provenance, fork-CI and deviation/promotion gates are tracked in OE #33. Experimental reverse-engineering PRs remain separate and unmerged. Beta/default branches and installed kernels are unchanged.

Retained exact source: d4504d974f6ae826797c8db51ffb0a8d46909c92.
Dated consolidation record: ooaklee/linux-surface-pro-11-oe#33 (comment)

This duplicate review vehicle contains the exact bcf63b5/d4504d9 commits already retained by #75 and original v26. No separate v9 fresh-boot checklist is claimed complete, and OE issue #69 remains open.