> "the lock duration is strictly bounded; it is determined by the application or the thread-scheduler" Practically speaking - wouldn't the exact duration be he

> "the lock duration is strictly bounded; it is determined by the application or the thread-scheduler" Practically speaking - wouldn't the exact duration be heavily influenced by the countless layers of abstraction beneath the application itself (kernel, scheduler, hardware, speculative execution, etc)? If so, can we ever truly make the claim that the "duration is strictly bounded"?