Home /Modules/eprom-corrado/COMPLETENESS.md

COMPLETENESS.md

eprom-corrado/docs/COMPLETENESS.md · 54 line(s) · markdown

eprom-corrado — Driver Completeness

Audit status for the EPROM programmer driver family, using the status concepts and vocabulary from the repository-level drivers/DRIVER.COMPLETENESS.md and TERMINOLOGY.md. This record is deliberately conservative: it does not convert assumptions into actual hardware values to appear complete.

Overall status

Provisional path (software-verified; not yet hardware-complete).

The image/checksum library, the operator logic (backup / copy / delete), the EpromControl contract, and the three connector transports are implemented and tested against a fake backend and reference images. The live-hardware USB path (real TL866 bulk framing to a physical EPROM) is not audited here and is not claimed hardware-complete.

Device-specific completion checklist

Mapped against DRIVER.COMPLETENESS.md §"Device-specific completion":

#ItemStatusEvidence / note
1Exact model & revisionActualTarget: 27C256 (32768×8) EPROM in the Corrado ECU; 27C512 twin-tune variant. From Corrado corrado_eprom.h.
2Firmware/interface versionUnknownTL866 firmware revision varies per unit; not enumerated here.
3TransportActualUSB bulk via libusb-1.0; endpoints EP_OUT 0x01 / EP_IN 0x81.
4USB descriptorsActual (IDs)TL866A VID 0x04D8 PID 0xE11C; TL866II+ VID 0xA466 PID 0x0A53. Full descriptor set not re-enumerated.
5Pin/control mappingsNot applicableProgrammer abstracts chip pins; no direct pin control exposed.
6Register/memory mappingsActualFlat 32768-byte image; trailing 2-byte little-endian checksum word.
7HID report layoutsNot applicableBulk device, not HID.
8Command/response pathsProvisionalTL866 bulk read/write/verify framing is a readable reference in Corrado's drivers; cross-check vs. upstream minipro before real flashing.
9I/O data pathsVerified (software)read→image, image→write→verify exercised over the fake backend and reference images.
10State transitionsVerified (software)open → read/write/erase → blank-check → close modelled and tested.
11Error & recovery pathsVerified (software)Explicit failure reasons (ARG/IO/SIZE/CHECKSUM/NO_DEVICE/USB/VERIFY/UNSUPPORTED/TIMEOUT).
12Hotplug/reconnectUnknownNot implemented in this reference.
13Cache/DMANot applicableNo DMA path in the host-side library.
14Unsupported-feature reportingActualUV/OTP 27C parts report UNSUPPORTED for electrical erase (must use a UV eraser); reusable replacements succeed.
15Platform-specific behaviorVerified (software)Linux/macOS/Windows build paths exist; Linux core + vendored-libusb build verified.
16Test evidencePartialChecksum library and connector transports verified against G60/VR6 *_REFERENCE.bin; Process transport verified driving the real compiled CLI. No live-chip test.

What would raise this to hardware-operational / complete

  1. Verify the TL866 bulk framing against a physical unit of a known firmware revision (closes item 8 and part of 16).
  2. Enumerate the full USB descriptor set and firmware version on real hardware (items 2, 4).
  3. Add hotplug/reconnect handling and record its behaviour (item 12).
  4. Record live read/write/verify evidence on a physical 27C256 / reusable part (item 16).

Until then this family is a provisional path behind an explicit evidence boundary — accurate, reusable, and honest about where real-hardware proof is still owed. Never treat the provisional framing as a verified hardware value.