Implement mutex

Implementing a hybrid mutex in C++

This medium coding problem tests whether you can build a lightweight synchronization primitive that balances busy-waiting and scheduler yielding. Quant firms and low-latency shops use hybrid mutexes to reduce context-switch overhead on contended locks while avoiding CPU waste on long waits.

The core challenge is combining atomic operations with a spin-then-yield strategy: attempt to acquire the lock with a short spin loop (checking the atomic state repeatedly), then deschedule the thread if acquisition fails. You must achieve this within a strict memory footprint, use appropriate memory-ordering semantics (avoiding sequential consistency where weaker guarantees suffice), and prefer early-exit patterns that avoid expensive read-modify-write operations on the common fast path.

• Atomic compare-and-swap and memory ordering (acquire, release, relaxed)
• Spin loops vs. syscall-based waiting (e.g., futex, condition variables)
• Lock-free fast paths and contention backoff
• Correctness under concurrent access and edge cases (non-recursive, single-threaded safety)