Physical PostSample comparison: only worker entry observed

The temporary matching PostSample boot reached its fresh 7.3.0-rc5 banner, exact received arguments, /bin/sh entry and primary Bash prompt. One receipt command was sent. The bounded 180.0975-second capture observed only the worker-entry breadcrumb: no receipt, first-post-sleep, numeric samples or new PostSample points arrived. No normal return or parser protocol error was observed. This is an incomplete diagnostic observation, not ordinary mainline boot acceptance or an identified cause.

Root operated the board/UART exclusively from worktree /home/jadams/tmp/k230-mainline-probe-integration, branch integrate/mainline-probe-path, controller revision e7c2b1a9, using the matching full build from frozen 75cc49df. The intervening commits contain host evidence, not source changes. Result records exact artifact identities, UTC times, return 2 and private capture SHA256/size. Raw UART, nonce, boot IDs, protected baseline and transfer URL remain private. No camera or glass proof was obtained during this candidate test.

Exact/full host proof and actual positive controller preparation passed before hardware access. Transfer/import staged eight new paths with return zero; the closure was GC-rooted. Fresh protected preflight checked normal system/profile/kernel/init, eight boot hashes, services and registration absence. The controller rebooted normal Linux into U-Boot, checked load sizes/CRCs and booted the matching Image/initrd/DT with the exact qualified volatile policy. Persistent normal boot selection stayed unchanged.

python3 tools/mainline-drm-initrd-shell-trial.py \
  --mode minimal --same-image-shell-pid1 --uart-progress \
  --uart-progress-breadcrumbs --uart-progress-post-sample \
  --bundle /nix/store/fjmxf6kn1yq0xk9v783amgymybhcrwkb-k230-mainline-drm-trial-boot-files \
  --manifest "$PRIVATE_RUN/candidate-manifest.json" \
  --normal-report "$PRIVATE_RUN/normal-report.json" \
  --log "$PRIVATE_RUN/trial.private.log" \
  --result "$PRIVATE_RUN/result.private.json"

The actual controller’s completed structured result qualifies the prompt and records receipt UNKNOWN; the raw-line monitor is not authoritative. Exactly one stimulus was attempted. No retry, proc command, guard, extra input or candidate reboot followed. Fixed records contains only the observed constant line. Entry arrived before the stimulus; host arrival order is not hardware execution timing.

The selected source places entry immediately before the first msleep(5000). Entry presence proves reaching its SBI output attempt, not return from that call. The absent first-post-sleep therefore leaves entry-call return, sleep/wakeup, later execution and later output unresolved. No numeric counter delta can be computed in this run. The previous Breadcrumbs run reached sample 1 and returned a Bash receipt; this run did not repeat that progress. Neither run establishes a stable stall location or causation by the two new markers. The new markers are only attempted after sample 1’s output and after the third sleep, respectively; absence provides no evidence that those calls were reached.

The source-only boundary review identifies schedule_timeout_uninterruptible, its timeout timer callback and task wakeup as independent milestones worth observing. Any next comparison should record progress in memory independently of the reporter’s SBI writes and account for the remaining shared serial output limit. Another chain of ECALL breadcrumbs alone cannot settle their own return. No IRQ/clock/firmware fault is claimed from missing serial output.

The capture completed and board/UART reservation was released. The subsequently requested NEW operator reset passed fresh guarded normal postflight: distinct boot identity, exact system/profile/kernel/init, eight boot hashes, three active services and registration absence. Home IPC returned zero; a separate reviewed camera image shows Home clock/icons/background, with angle/glare/focus limits. Operator reset receipt. This does not establish automatic return, ordinary mainline /init, usable root or panel/glass acceptance; task5b.5 stays UNVERIFIED. The historical controller status records the capture’s missing return; the separate receipt establishes later manual recovery.

Physical evidence08ea7370 passed CI37175128998 and both published work/physical pages returned HTTP200 with that exact revision; deployment receipt. Independent review matched the fixed facts and capture digest to protected data. Group5h is reconciled; group5i’s separately landed Memory comparison is next.