What runs here
On this node class, the operating-system layer is often just a small amount of scheduling code. After startup initialises the hardware, a main loop repeatedly checks inputs, advances state machines, handles communication, and updates outputs. Interrupts handle events that cannot wait for the next loop iteration.
A tiny cooperative scheduler may call functions when their periods expire or events arrive. Each function finishes and returns before another scheduled function runs; interrupts can still pre-empt it. This usually needs neither a separate stack for every function nor a general-purpose task kernel.
| Component | Role |
|---|---|
| Main loop or cooperative dispatcher | Chooses which piece of work runs next |
| Timer and time base | Tracks periodic work, deadlines, and timeouts |
| Interrupt handlers | Capture or service urgent peripheral events |
| Event flags or bounded queues | Pass work from interrupts to the main loop |
| Execution supervision | Checks progress and coordinates watchdog servicing |
These are implementation choices, not mandatory standard modules. A fixed-function ASIC may implement equivalent sequencing in hardware, with no programmable operating system at all.
Why it looks like this
A lamp, flap actuator, or simple sensor has a narrow workload. Limited RAM, flash, peripheral resources, and cost targets favour a small execution model. Hardware timers, capture units, or PWM peripherals can provide precise behaviour without scheduling every edge in software.
Bare-metal software still has timing requirements. Its responsiveness depends on the longest uninterrupted activity, interrupt interference, and how often the loop revisits each function. A short, bounded loop can be straightforward to analyse; a blocking diagnostic operation can invalidate that analysis.
A fixed firmware image with infrequent functional changes can make static scheduling practical. Update frequency is nevertheless a product decision: adding diagnostics or reprogramming may introduce substantial work even when the actuator function stays simple.
The grid labels this node QM, meaning quality management rather than an assigned ASIL development obligation for this cell. That assumption must follow the vehicle’s safety allocation. Neither a small MCU nor the absence of an OS establishes safety suitability; revisit the architecture if safety requirements reach this node. Hazard analysis and the functional safety concept belong to the concept phase of ISO 26262-3:2018.
How it connects
Above this layer, application state machines and platform services receive periodic calls, events, or completion notifications. Explicit interfaces should define command freshness, execution deadlines, and behaviour after communication loss.
Below it, peripheral drivers expose timers, ADC readings, PWM outputs, and communication events. Interrupt handlers should capture essential information and defer longer processing. Shared data needs deliberate protection: even a single-core main loop can race with an interrupt.
On LIN, the commander, traditionally called the master, schedules frame headers. The responder, traditionally called the slave, receives or publishes the response associated with a header. The bus schedule therefore creates local servicing constraints; it does not schedule the node’s entire application. This division is described in the LIN Specification Package 2.2A.
Sleep and wake-up connect execution to the transceiver and power hardware. Define which event wakes the MCU, which clocks and peripherals recover, and when outputs become valid. A 10BASE-T1S variant needs its own communication and timing assessment; Ethernet does not automatically require a full OS.
Compared with other node classes
- HPC application processor: Linux, QNX, or Android supports a much broader application environment, often within virtual machines, with richer memory and process management.
- HPC safety island: AUTOSAR OS or a safety RTOS supports explicitly scheduled safety workloads and the required assurance activities.
- Zone controller: Multicore AUTOSAR OS coordinates more concurrent work, shared resources, and cross-core interactions.
- MCU-based ECU: AUTOSAR OS or another RTOS provides task scheduling and synchronisation as workload complexity grows. AUTOSAR OS defines tasks, interrupts, alarms, and other configured objects in its Operating System specification.
Best practices
- Bound each execution step; express long operations as state machines.
- Budget loop time, interrupt latency, and communication servicing together under worst-case load.
- Keep interrupt handlers short and define ownership of shared buffers.
- Service the watchdog only after checking meaningful execution progress.
- Test startup, brownout recovery, sleep, wake-up, and missing commands on target hardware.
- Reassess the scheduling model whenever new features consume timing or memory margin.
Common pitfalls
- Treating a scheduled LIN bus as proof that local software always meets deadlines.
- Using blocking delays or peripheral polling that stalls unrelated work.
- Assuming
volatilemakes shared accesses atomic or prevents races. - Assuming QM permits undefined failure behaviour or removes the need for robust engineering.