# USB 32 KiB matrix Round 1 Pi precheck: USB01_2G and USB05_4G

Status: complete
Type: evidence
Updated: 2026-08-01
Next Action: none; retain as historical precheck evidence for the completed matrix

Project Map: work/maps/map-smi-firmware-stabilization-and-qualification.md
Workpackage: work/wp/completed/wp-restore-broad-usb-compatibility-baseline.md
Checklist: work/checklists/checklist-stage-1b-four-drive-usb-compatibility-baseline.md

## Scope

This is the repeat read-only media gate immediately before Round 1 of the
isolated 32 KiB four-drive matrix. The original baseline already includes a
complete direct read of each drive. This gate reconfirms identity, capacity,
FAT consistency, retained test content, current-insertion kernel state, and
final unmounted state without repeating that non-destructive but
time-consuming whole-device read.

## Identity and capacity

| Canonical label | Pi device | Model / identity | Capacity | Partition / UUID |
| --- | --- | --- | ---: | --- |
| `USB05_4G` | `/dev/sda` | General UDisk, revision `5.00`, VID:PID `abcd:1234`, descriptor serial unavailable | 4,026,531,840 bytes | MBR `0xb820a071`; FAT32 LBA partition `2,048..7,864,319`; `362E-8170` |
| `USB01_2G` | `/dev/sdb` | USB DISK 2.0, revision `PMAP`, VID:PID `13fe:4300`, descriptor serial unavailable | 2,013,265,920 bytes | MBR `0x43837690`; FAT32 LBA partition `2,048..3,932,159`; `D45F-70AA` |

`sg_readcap` independently reported 512-byte logical blocks and the same final
LBAs and capacities. Linux device names are session-local; canonical labels,
capacity, model, VID:PID, filesystem UUID, and partition identity govern the
move.

## Read-only filesystem gate

- `fsck.fat -n -v` reached the final filesystem sector on both drives, found
  two FATs, found no structural corruption, and left each filesystem
  unchanged. Exit status `1` is solely the preserved Windows convention that
  stores the canonical label in the root directory while leaving the
  boot-sector label blank.
- Both FAT copies are byte-identical on each drive:
  - `USB05_4G`: each 3,923,456-byte FAT hashes to
    `3BC8B4639D45551C54A5F1C1E7E3893F00542E5D7837DC182BE3E368DD45873D`.
  - `USB01_2G`: each 1,961,472-byte FAT hashes to
    `37EB3CA8B6C1E43669867FEF692CC53DE2DD3852CD38F310AB0F1D8D94F05545`.
- Read-only mounts confirmed the expected common files on both drives:
  - `7520553_SMI_00_05_00.imx`, 8,388,864 bytes, SHA-256
    `908966E3460C1594D21DE0B379BE2B06F1D51057A7EC5D1C25D0540F2381D7D9`;
  - `W00.wav`, 198,566 bytes, SHA-256
    `37A9DC665366F06D8081B23ECF88BDB30CF81546C3A10EDA1BE827DE2D006B92`;
  - `W01.wav`, 176,516 bytes, SHA-256
    `E22FC8FFF2CF68878C56EA87660CA2325E0FBF04CF091AAEB112A7F274B1B2F7`;
  - two 225-byte PID captures per drive, all SHA-256
    `6BCB89F2EFACC503B950EF00151A51AF8353C40B53781110C14DDC441737FC3A`.
- The prior CSVs remain complete and have the expected sizes and record
  counts: `R1A_USB01.csv` 3,990 bytes / 65 records,
  `R2B_USB01.csv` 3,739 bytes / 61 records,
  `R1B_USB05.csv` 3,739 bytes / 61 records, and
  `R2A_USB05.csv` 3,990 bytes / 65 records.
- No unexpected `.tmp`, `.part`, or `.bak` publication artifact exists.
- The kernel log from the current insertion contains no reset, timeout,
  medium, device-offline, buffer-I/O, block-I/O, sense-key, descriptor-read,
  or rejecting-I/O error.
- Both partitions were unmounted before the checks and are unmounted after
  them.

## Disposition

`USB01_2G` and `USB05_4G` pass the Round 1 entry gate unchanged. They are ready
for the powered-controller production workflow on the exact deployed isolated
32 KiB candidate. No repair, format, VBUS action, controller power action, or
specimen-specific mitigation was used.
