Input: mshw0485: add opt-in pen userspace interfaces

[ooaklee]

Summary

• add the small HID-core BUS_SPI definition needed by an MSHW0485 HID child
• split the raw heat-frame misc ABI and the iptsd HIDRAW bridge behind independent options that default to n in Kconfig
• select both options in the Ubuntu qcom-x1e arm64 policy; runtime registration still requires the matching Denali device-tree opt-in
• require a matching device-tree opt-in before either optional endpoint can register
• opt only the shared Surface Pro 11 Denali description into those interfaces

Dependency and status

Review this PR after #82 and #83. Its head is stacked on x/sp11-qspi-touch-core, so the GitHub diff is cumulative until #83 is merged into jg/ubuntu-qcom-x1e-7.2.y. The pen-specific portion is the five commits after 676d8e800267.

This branch does not merge the broad -sp11 staging branch or replay the correction merged through #81. It deliberately builds on the OF-only direct-touch core isolated in #83.

This PR is a draft. A custom build with both options disabled proves that it does not expose the experimental interfaces unintentionally. The Ubuntu qcom-x1e flavour deliberately enables both options, but their matching Denali device-tree properties remain mandatory and the enabled configurations still require the hardware checks below before the topic is ready.

Containment

• CONFIG_TOUCHSCREEN_MSHW0485_HEAT_FRAME and CONFIG_TOUCHSCREEN_MSHW0485_IPTSD are independent and default to n
• the Ubuntu qcom-x1e arm64 annotations explicitly select both options as y; this downstream flavour policy does not change their Kconfig defaults
• the heat-frame misc device additionally requires microsoft,enable-heat-frame-abi
• the HIDRAW iptsd bridge additionally requires microsoft,enable-iptsd-bridge
• only the common Surface Pro 11 Denali DTSI carries those properties
• with both options disabled, the compiled driver retains SP11: add protocol-9 QSPI and OF-only MSHW0485 touch #83's touch-only initialization, falling-edge IRQ handling, and input path, with no misc-device or HID child references
• the HID child connects with HID_CONNECT_HIDRAW only; it does not create a second touchscreen input device
• only the core DFT reports are relayed to iptsd; diagnostic report 0x07 and privacy-sensitive report 0x6e remain off the production HIDRAW path
• the existing MSHW0485 driver remains the sole transport, panel-mode, recovery, and touch owner; the generic SPI-HID transport is not carried
• no non-Denali board opts into either interface

Provenance

The HID-core commit preserves Jarrett Schultz as the author of the current upstream v4 patch and retains its exact published chain:

• Signed-off-by: Dmitry Antipov <dmanti@microsoft.com>
• Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
• Signed-off-by: Jingyuan Liang <jingyliang@chromium.org>

The one-line HID-BPF enum mirror is explicitly marked as Leon's current-tree adaptation. Signed-off-by: Leon Silcott <leon@boasi.io> is appended to that submission; no Jarrett sign-off is invented because the authoritative v4 patch does not contain one.

The remaining four focused commits are authored by Leon Silcott, carry the authorized Boasi sign-off, and retain immutable links to the corresponding integration commits for the heat-frame and iptsd work.

The final Ubuntu policy commit links the original target-integration commit and preserves Leon Silcott as author with the authorized Boasi sign-off.

The implementation commit also adds a focused MICROSOFT SURFACE G6 TOUCHSCREEN DRIVER maintainer entry covering the driver, binding, ABI documentation, and UAPI header.

Validation

• git diff --check 676d8e800267..HEAD for the incremental pen topic
• strict checkpatch is clean for code and style; the imported BUS_SPI commit reports only MISSING_SIGN_OFF for nominal author Jarrett because the exact published trailer chain intentionally contains no Jarrett sign-off
• ARM64 ubuntu_x1e_defconfig W=1 builds in all four configurations: both options disabled, heat only, iptsd only, and both enabled
• disabled-build symbol audit confirms there are no misc-device, heat-ABI, HID child, or HIDRAW references in the MSHW0485 object
• heat-only symbol audit confirms the misc ABI is present without HID-child/HIDRAW references
• DT binding/schema validation for the two opt-in properties
• scripts/get_maintainer.pl resolves the new driver, binding, ABI documentation, and UAPI to Leon Silcott and the Linux input list
• successful builds of both Surface Pro 11 Denali DTBs and the non-SP11 Dell XPS 13 X1E DTB
• Ubuntu qcom-x1e annotation export selects both options as y; olddefconfig preserves them and the annotations checker reports no MSHW0485 policy drift
• source audit confirms that both Kconfig options retain default n; compiled-DT audit confirms that only Denali carries the matching runtime opt-ins

Hardware checks still required

• first confirm the exact SP11: add protocol-9 QSPI and OF-only MSHW0485 touch #83 QSPI/direct-touch build boots and retains stable multitouch and suspend/resume behavior
• boot a custom build with both optional kernel settings disabled and repeat those direct-touch checks
• enable only the heat-frame option and verify /dev/g6ts-heat, record framing, polling, generation changes, controller recovery, suspend/resume, and removal while a reader is open
• enable only the iptsd option and verify the HIDRAW sibling, stylus input through iptsd, feature handling, bounded rebind after recovery, and absence of a duplicate touchscreen device
• confirm diagnostic report 0x07 and privacy-sensitive report 0x6e are not emitted on the HIDRAW path
• boot the standard Ubuntu qcom-x1e configuration, with both options selected by annotations, and repeat pen, touch, reset/recovery, suspend/resume, and removal stress tests
• run the corresponding direct-touch smoke test on both supported SP11 variants; non-SP11 systems remain compile-time guarded because their device trees do not opt in