ISO 14229 UDS diagnostics
on Zephyr RTOS.

// why EDS for Zephyr

UDS on Zephyr — done right.

On Zephyr the diagnostics problem is mostly a binding problem. The RTOS already gives you a CAN driver model, a network stack and a settings subsystem — so the work is wiring generated UDS handlers to a devicetree device correctly, not writing drivers. EDS generates those handlers from a 30-line YAML description and hands them a const struct device *:

ecu:
  name: BCM
  addressing:
    can_id_request:  0x7DF
    can_id_response: 0x7E8

sessions:
  - default
  - extended
  - programming

services:
  - id: 0x22   # ReadDataByIdentifier
  - id: 0x2E   # WriteDataByIdentifier
  - id: 0x27   # SecurityAccess
  - id: 0x31   # RoutineControl
  - id: 0x34   # RequestDownload
  - id: 0x36   # TransferData
  - id: 0x37   # RequestTransferExit
  - id: 0x19   # ReadDTCInformation
  - id: 0x14   # ClearDiagnosticInformation

dids:
  - id: 0xF190
    name: VIN
    length: 17
    read_sessions: [default, extended]
  - id: 0xF187
    name: PartNumber
    length: 10
    read_sessions: [default, extended]
    write_sessions: [extended]
    write_security: [1]

python3 tools/codegen.py \
  --config diagnostics_config.yaml \
  --out build/ \
  --safety-wrappers \
  --asil-level B \
  --test-gen

build/
  uds_server.c/.h      # UDS dispatcher + Zephyr thread
  uds_dids.c/.h        # DID read/write handlers
  uds_dtcs.c/.h        # DTC storage (NVM-backed)
  uds_sessions.c/.h    # session state machine
  uds_security.c/.h    # AES-128-CMAC SecurityAccess
  uds_safety.c/.h      # ASIL-B 5-step validation
  test_uds.py          # pytest suite (CI-ready)

# west.yml — add EDS as a Zephyr module
manifest:
  projects:
    - name: eds
      url: https://github.com/Xaloqi/EDS
      revision: v1.18.1
      path: modules/eds

# prj.conf
CONFIG_EDS=y
CONFIG_EDS_ISOTP=y
CONFIG_EDS_DOIP=y

// service coverage

All 19 UDS services, on either RTOS.

Service coverage is a property of the generator, not of Zephyr — the same 19 services are produced for a FreeRTOS build from the same YAML. The full list, the generated file layout and the codegen flags live on the EDS product page.

What is Zephyr-specific is everything below: how those handlers reach a CAN controller, how DTCs survive a reboot, and how you run the whole thing with no hardware at all.

// safety

ASIL-B artifacts, and what Zephyr contributes.

Every generated handler runs a 5-step validation chain before its body — DID existence, session permission, security level, access permission, data length — produced by codegen rather than written by hand. The chain, the deviation log and the traceability matrix are covered on the Safety & security page.

The Zephyr-specific part is the evidence you can produce cheaply: native_sim runs the full generated pytest suite on the host in CI, so the ASIL-B behaviour is exercised on every commit rather than only on a bench. The Professional tier adds the safety manual, HARA and RTM built against that same configuration.

// pricing

Free runtime. Two paid tiers.

The GPL v2 runtime is always free on GitHub. Developer adds the code generator, test generation, ARXML import and the AI CLI; Professional adds the ASIL-B safety artifacts and the harness. Annual licence, offline after activation.

// further reading