In concurrent programming, what best defines a lock?

• A. A mechanism to protect data from being viewed or changed while it is updated.
• B. A scheduling policy for threads.
• C. A memory allocation unit for objects.
• D. An error-handling construct.

Answer: (A)

A lock is a synchronization primitive that ensures exclusive access to a shared resource so only one thread can enter a critical section at a time. This prevents data races and keeps updates atomic from the perspective of other threads. The usual pattern is to acquire the lock before using the shared data and release it when done; if another thread attempts to acquire it while it’s held, it waits or spins until it becomes available. This mechanism directly protects data during updates, which is the core purpose of a lock. It isn’t about scheduling decisions (that would be a scheduling policy), nor about memory allocation, nor about error handling.