# USB08_8G command-compatibility Release qualification — 2026-07-30

Status: complete
Result: passed
Type: evidence
Updated: 2026-08-01
Next Action: none; retain as historical reactive-fallback evidence, not the current production lifecycle
Campaign: 003-complete-bounded-hpt-slow-export-firmware-ticket
Parent: work/wp/completed/wp-implement-usb-command-class-compatibility-fallback-end-to-end.md

## Candidate identity

- Debug S-record SHA-256:
  `67E539FE0947BFC8C5B5EC57FE154212E23C1515AE0563FE09D29C03123481BA`
- Release S-record SHA-256:
  `838645AF04F6718231AE892A7F2EB16EC6710A8354DA270A6601CFEAC8E6ABE9`
- Release ELF SHA-256:
  `D8183962FB5DB570FD25B9C473E2D02EA78CB436CF007AC3213533A00FA96DD0`
- A/B internal-flash SHA-256:
  `FFBA2B680B07543416D8A6357FDE17A47F433626519824E6182491976FFD1B03`
- Both full 4 MiB flash captures matched the Release S-record exactly.

## Target and media identity

- Controller A J-Link: `69664082`
- Controller A SD CID: `035344535330384780976F0084018559`
- Controller B J-Link: `69660123`
- Controller B SD CID: `035344535330384780976F0085018507`
- USB label: `USB08_8G`
- USB serial: `20111009140821671DA2`
- FAT UUID: `46A9-9922`

## Debug timing qualification

The drive rejects SCSI SYNCHRONIZE CACHE with completed-command
`ILLEGAL REQUEST`. A 5-second passive drain was insufficient: the renamed
image was not visible at the immediate post-cycle verification, although it
appeared after a later controller reboot. The 30-second drain passed but was
operationally excessive.

The selected 10-second bounded drain passed three consecutive full image-save
durability trials on Controller B and three on Controller A. All six trials:

- wrote and renamed exactly 8,388,864 bytes;
- completed a physical USB VBUS cycle;
- established a fresh media generation;
- remounted with acceptance mask `15`;
- verified `8,388,864/8,388,864` bytes after remount; and
- ended with zero FileX and live system errors.

## Release qualification

Controller B then Controller A each passed:

- Release firmware/hash and SD CID identity gates;
- `USB08_8G` mount with acceptance mask `15`;
- one full image-save VBUS/remount/CRC durability cycle;
- one explicit VBUS-off/remount test with `mounted_after=1`; and
- zero live system errors.

Raw evidence directories:

- `work/evidence/hpt-log-store-circular-mirror/controller-b/usb-command-compatibility-release-20260730/`
- `work/evidence/hpt-log-store-circular-mirror/controller-a/usb-command-compatibility-release-20260730/`
- `work/evidence/hpt-log-store-circular-mirror/controller-b/usb-command-compatibility-10s-20260730/`
- `work/evidence/hpt-log-store-circular-mirror/controller-a/usb-command-compatibility-10s-20260730/`

## Independent Raspberry Pi verification

After final Release testing:

- `/dev/sdc1` identified as `USB08_8G`, serial
  `20111009140821671DA2`, UUID `46A9-9922`;
- `fsck.vfat -vn` reported a clean filesystem with two FATs and no repair;
- `7520553_SMI_00_05_00.imx` size was 8,388,864 bytes;
- image SHA-256 was
  `4E0E1C3DD651B557FD6B71840C60E6D1DA8493E413D39EFE825C7853F8536CD6`;
- `W00.wav` SHA-256 was
  `37A9DC665366F06D8081B23ECF88BDB30CF81546C3A10EDA1BE827DE2D006B92`;
- `W01.wav` SHA-256 was
  `E22FC8FFF2CF68878C56EA87660CA2325E0FBF04CF091AAEB112A7F274B1B2F7`;
- no `.TMP` files were present; and
- the drive was left unmounted on the Pi.

## Scope boundary

This evidence qualifies `USB08_8G`, the selected 10-second unsupported-cache
drain, and the A/B Release deployment. Controller B had no retained HPT
records, so a meaningful HPT export was not performed. The broader workpackage
also retains open fault-injection and compliant-SYNCHRONIZE-CACHE coverage.
