On signaling a condition variable, what happens to mutex ownership?

Question

1. Thread T1 locks mutex M and waits on condition C (releasing M).

2. Thread T2 locks mutex M and signals condition C.

In literature it's stated that once C is signaled, T1 wakes up and reacquires M.

I was left doubtful for what happens with T2 since it had M acquired before signaling C.

What happens to T2 regarding M which was acquired before signaling?

1. Does T1 gets ready but blocked, executing only after T2 releases M, for it to be reacquired?

2. Does T1 gets ready and executes while T2 gets blocked till M gets released, for it to be reacquired?

3. They just execute concurrently, T1 with M acquired, T2 with M released (or also acquired?)

4. ?

Answer 1

Signalling the condition C wakes up the thread T1, then it has to reacquires the mutex M by locking it. In order to do so, T2 must release M first.

So the answer is, T1 wakes up on signal, waits to acquire M, T2 releases M, T1 acquires M, T1 and T2 execute concurrently.