# HDMI acceptance and timing deferral — 2026-10-09

Evidence class: operator acceptance linked to previously committed serial,
sysfs and build identity. No new board read, reboot, camera capture or latency
measurement was performed for this closeout.

The operator observed HDMI → panel → HDMI with the expected input modes:

> came up on hdmi. worked fine, touchpad worked. unpliugged hdmi, panel came up, touch worked

After the faster worker trial:

> that's great it's perfect. replugging hdmi is slower than i'd like (couple of seconds) but unplugging -> display was almost instant

The contextual qualification records the requested cable cycle and real-touch
All Apps, Overview and Terminal → Home checks, then the response:

> it works great land it

Source records: `fast-runtime-operator-report.json`,
`fast-runtime-qualification.json`, `fast-runtime-qualified-state.json` and
`fast-runtime-serial-boot.json` in this directory. They retain trial boot ID
`cff529ad-6aa6-4a31-b9af-337bbc783521`, HDMI portrait transform and virtual
trackpad. This is physical operator acceptance; the named navigation checks
use contextual affirmation, without invented per-gesture traces.

The operator then directed:

> k what's next ignore latency measurement good enough for now what proposals can we close

Precise sampled ≤1s HPD-to-connector and ≤3s HPD-to-enabled targets are therefore
deferred, not performed or passed. `fast-runtime-latency.json` remains an empty
transition capture with `INCOMPLETE_OR_SLOW`; it establishes neither a measured
failure nor success. The approximate visible observations support the original
30-second ceiling for the accepted transitions. Monitor lock time is not
separately measured. Independent performance proposals retain their own gates.

## Installed artifact

Implementation/default promotion landed in
`6b4fe555645ac553914549fcd926d1f0902b65d1`. The normal installed system/profile is
`/nix/store/yl3si5ak6yi709yg1fqsnwgq0zn4xfcs-nixos-system-nixos-26.11.20260919.20b1ddd`.
Matching kernel:
`/nix/store/f3rnipvbc5vqam8kwdmn3wgcl5rbz9yx-linux-riscv64-unknown-linux-gnu-7.3.0-rc5/Image`,
SHA256 `9d94102c65c8e4fd6b8880ef7b19b8bee5fa91c85596384319f0104bd89509cb`.
Normal bundle is `/nix/store/jsqi00cgac7v773xi3r1zc6ppib3sc7p-k230-coherent-shell-boot-files`;
built default image is `/nix/store/gw9p4yhc2yyp026bgmn635m26bhlrzhv-k230-sd-image.img`.

`normal-hotplug-install-serial.json` proves recoverable normal installation and
ordinary autoboot, retaining protected stage 1, DT selectors and rollback roots.
`normal-hotplug-runtime-state.json` identifies installed boot
`4bf73b24-a16a-4786-96fb-f1288244d96f`, enabled HDMI and active shell/input services.
`default-hotplug-image-check.json` proves the built image matches the qualified
bundle. No whole-image flash is claimed. The volatile trial and subsequent
normal autoboot remain distinct evidence classes.

## Integration ownership

Worktree `/home/jadams/tmp/k230-hdmi-close-final`, branch
`closeout/hdmi-accepted-2026-10-09`, base
`ba5217ca72f05fbc49ba818418737e8549ae5685`. Owned paths: the HDMI change/archive,
this evidence, its work-board entry and necessary references, and CLI-generated
`display/hdmi` and `display/touch` specs. No source changes, build slot or board
reservation. The operator subsequently canceled Settings reboot scope and authorized the
landscape successor. See `scope-decision-2026-10-09.md` for the exact decision
and task mapping. No canceled or transferred gate is reported as performed.

Narrow closeout checks: `openspec validate plugging-in-hdmi-moves-the-display --strict`,
then (only after scope preservation) `openspec archive plugging-in-hdmi-moves-the-display --yes --json`,
`openspec validate --all --strict`, `python3 scripts/build_site.py` and
`python3 tools/blob-scan.py --no-vendor`. Record exact successor/task mapping and
sync verification before final integration.