// the find
eclipse-threadx/threadx
Eclipse ThreadX is an advanced real-time operating system (RTOS) designed specifically for deeply embedded applications.
Eclipse ThreadX is a preemptive RTOS for microcontrollers, originally Azure RTOS from Microsoft, now donated to the Eclipse Foundation. It targets deeply embedded systems where you're counting bytes of RAM and nanoseconds of interrupt latency — not a general-purpose embedded OS. Port coverage is genuinely broad: 30+ architecture variants across Cortex-M, Cortex-A, RISC-V, ARC, Xtensa, and even win32/linux for host-side testing.
Preemption-threshold scheduling is a real differentiator — it lets you bound priority inversion without full mutex overhead, which most RTOSes don't offer. The error-checking layer (txe_* files) is a separate compilation unit you can strip from release builds, so you get debug-time parameter validation without paying for it in production. The module system (common_modules) gives you MPU-enforced isolation for dynamically loaded code, which is rare at this price point. FreeRTOS and POSIX compatibility shims in the utility layer mean migrating existing code doesn't require a full rewrite.
The codebase still says 'Microsoft Corporation' in author headers throughout — the Eclipse transition is governance-level, not a real community fork, and contributor diversity is thin. The Modules feature (dynamic loading with MPU isolation) only covers a handful of Cortex-M/A targets; if you're on RISC-V or ARC you don't get it. The byte pool allocator is a first-fit with coalescing — it will fragment under long-running workloads and there's no compaction; block pools are the right answer but the API makes this easy to ignore until it bites you in production. Documentation lives across three separate repos (rtos-docs, getting-started, samples) which makes it genuinely hard to find what you need.