What is true about the functions wait$()$ and post$()$ for condition semaphores?

By calling wait $()$ we decrease the semaphore by one, and post$()$ increases the semaphore by $1$

The thread calling wait$()$ goes to the ready state, while the thread calling signal$()$ sends a signal to weak up anyone sleeping in the queue of the condition variable

The thread calling wait$()$ needs to held the lock, while the thread calling signal automatically acquires the lock

When a thread wake up $($from a wait$(])$ and start running automatically retake the lock, and a thread before calling signal $()$ needs to already have the lock