Changelog¶
What changed in each release of the driver. Entries under Breaking need source changes in code that used the affected API.
[Unreleased]¶
Changed¶
- The repository moved to kokeb-labs-drivers; URLs follow.
- Hardware page: the pinout, the block diagram and the application circuit are now figures drawn from
docs/hardware.yml, which also generates the pin table (replacing the hand-written one) and the parts table. Every pin, block and part cites the datasheet section and page; values the datasheet does not give are marked as not confirmed.kokeb-docs hardware --kicad-netlistexports the circuit. - Synced with the driver template:
tools/preflight.shruns clang-tidy over every file CI lints, as CI does, and the hardware-lessons review adds the switch-off, channel-in-doubt and fuzz-skip lessons.
[0.1.0] - 2026-11-09¶
Added¶
- Driver: the first version of the Infineon TLE92466ED driver (datasheet Rev. 1.2): start-up with the RESN pulse or t_POR, INIT_DONE, the ICVID identity and the start-up faults reported and cleared; Config and Mission Mode; per-channel mode, current setpoint, integrating current controller, PWM frequency controller, dither, slew rate and diagnosis; parallel pairs; channel enables written from a fresh read of CH_CTRL; feedback under the freeze of section 4.10.5; supply and temperature readings; every diagnosis register, a fault latch and the clears; the OFF-state overcurrent check; the self-test; the SPI watchdog; the external clock; raw register reads.
- Protocol: 32-bit frames with CRC-8 SAE J1850 and a two-frame access with an anchor read that checks every reply (CRC, width, R/W echo, status, stuck and looped-back lines, repeated words), bounded retries of reads, resynchronisation after a failure, reset detection.
- Safe switch-off: a disable writes even when the CH_CTRL read fails,
EnterConfigMode()and a refusedEnterMissionMode()drive EN low whatever happens on the bus, andGetDroppedChannels()reports the channels the chip switched off on its own, from validated reads only. The CH_CTRL copy never reports a channel on that is off, whether a write's reply carries the register before or after the write (section 5.2.1); the host tests run under both readings. A write whose reply failed may have taken effect: every copy the driver decides on without I/O then holds the more dangerous value or is unknown, and a guard on an unknown value returnsWrongStatewith nothing sent (docs/contract.md, "Ambiguous writes"). - Release gate: bench step 1 of
docs/limitations.md(how the chip clears its flags) must pass before a final release; until then the driver is released only as a release candidate. - Port:
SpiInterfaceand theSpiPortconcept with exact-signature hooks; optionalGpioSet(),GpioRead()andLog(); the bus requirements asconstexpr. - Tests and examples: a chip model written from the datasheet, host tests of every operation and failure path, compile-fail tests, documentation snippets, a fuzz target with its corpus, the host example and the ESP32 example.