Conditional variables (Mutex)

Conditional variables (Mutex)

A mutex may need special operations applied upon some condition being met. (For example processing a list when it is full.) To implement this we use data structure called conditional variable that holds at least:

• A reference to the mutex it is related to, and
• A list of waiting threads.

Then an OS will offer 3 calls on the API:

• wait(mutex, condition): mutex is released and reacquired when the condition is met (i.e. they are notified about the condition).
• signal(condition): notify a single thread waiting on the condition it has been met.
• broadcast(condition): notify all threads waiting on the condition. (Though as they are all waiting on the mutex they only activate one at a time.)