Test-and-set Lock uses an atomic value for indicating that some thread has engaged the lock and is executing its critical section (CS).
Each thread that wants to enter CS tries to engage lock, and checks to see if it was the one who managed to engage it (with an atomic operation). This has no effect if it was already engaged. If it managed to engage it, it proceeds to its CS, otherwise it just retries.
Once the thread is done with CS, it simply disengages the lock.
As all thread repeatedly attempt to engage the lock for themselves, it leads a storm on the processor memory bus (since the atomic operation ignores the cache). Since bus traffic is always high, it makes it difficult for the lock holder to disengage his lock (due to traffic). Also this scheme does not provide first-come-first- served fairness. Hence, this type of lock is only suitable for educational purposes.
Course: Concurrent Data Structures, Monsoon 2020
Taught by: Prof. Govindarajulu Regeti
lock():
1. When thread wants to access critical
section, it tries to engage lock, for
itself, with an atomic operation. This
has no effect if it was already engaged.
If it managed to engage it, it proceeds
to its CS.
2. If not, it retries again.unlock():
1. When a thread is done with its critical
section, it simply sets the "locked" state
to false.See TASLock.java for code, Main.java for test, and repl.it for output.