Consider the Test-and-Set spinlock on a system with several processors. Is it possible for the spinlock to spin on a cached memory location? Does decreasing the amount of time it takes, on average, for a processor to complete a critical section usually increase or decrease lock contention? Explain why.

Let's, say you have a linked list with a Test-And-Set (TAS) spinlock serializing access to it. Can you use this construction to implement a first-come- first-served (FCFS) lock for n threads? Explain how to do so, or why it is impossible.

Answer rating: 100% (QA)

a locklock m while testandsetm unlocklock m m false b TestAndSet type of instruction are necessary for mutual exclusiononly one process in critical se... View the full answer