Unsafe mutex implementation? No real multithreading?

[SerraAvenger]

while (#PARAM(lock)) {
    Wait(1.0, c_timeGame);
}
#PARAM(lock) = true;
#SUBFUNCS(actions)
#PARAM(lock) = false;

wth?

 

[Lyerae]

What is this?

 

[SerraAvenger]

What is this?

The "Critical section" from GUI

 

[Ancanus]

Seems to do what it's supposed to do.
Prevents threads from running a critical section more than one at a time.

Were you expecting a queue behaviour for the next thread to go in? It shouldn't be hard to implement as a separate function.

 

[SerraAvenger]

Seems to do what it's supposed to do.
Prevents threads from running a critical section more than one at a time.

If there's real multithreading, it doesn't.

// A
while (#PARAM(lock)) {
    Wait(1.0, c_timeGame);
}

// B
#PARAM(lock) = true;
#SUBFUNCS(actions)
#PARAM(lock) = false;

Thread 1, 2

1A
2A
=> 1B, 2B parallel => FAIL.

Were you expecting a queue behaviour for the next thread to go in? It shouldn't be hard to implement as a separate function.

No, I was expecting a thread-safe semaphore.

 

[Ancanus]

You mean that... #PARAM(lock) is originally false, thread 1 clears the while condition, a context switch occurs before thread 1 can set #PARAM(lock) to true, thread 2 clears the while condition and now both are inside the critical section?

Yeah, that would suck. However, we are assuming that context switches may occur at any time. Perhaps Blizzard implemented it in a way that context switches only occur on waits? But it wouldn't be real multithreading then...

 

[SerraAvenger]

Yeah, that would suck. However, we are assuming that context switches may occur at any time. Perhaps Blizzard implemented it in a way that context switches only occur on waits? But it wouldn't be real multithreading then...

That's exactly what I was saying =)