28. Review

28.1. Common Concurrency and Synchronization System Calls

Function	Reference	Brief Description
pipe()	here	Creates a pipe for bidirectional interprocess communication
mmap()	here	Allocates a page of virtual memory
pthread_mutex_init()	man 3 pthread_mutex_init	Creates a mutex
pthread_mutex_lock()	here	Locks a mutex
pthread_mutex_unlock()	here	Unlocks a mutex
pthread_mutex_destroy()	man 3 pthread_mutex_destroy	Frees resources allocated to a mutex
sched_yield(), sleep()	here	Call scheduler to switch threads
wakeup(), wakeup_all()	here	Moves threads from the wait queue to the ready queue
sem_init()	here	Initialize a semaphore
sem_wait()	here
sem_post()	here

28.2. Review Questions

What is the purpose of synchronization in concurrent programming?

To increase the speed of thread execution.To create independent execution entities within a process.To restrict the possible interleavings of threads and control how they use shared resources.To prevent race conditions.

What are the requirements for a solution to the critical section problem?

Concurrency, Deadlock Avoidance, Resource Allocation, and Scalability.Mutual Exclusion, Starvation Avoidance, Synchronization, and Efficiency.Synchronization, Atomicity, Deadlock Prevention, and Fairness.Mutual Exclusion, Progress, Bounded Waiting, and Performance.

How does Peterson's Algorithm differ from other solutions for implementing critical section entry and exit protocols?

Peterson's Algorithm ensures mutual exclusion and progress for two threads.Peterson's Algorithm is the simplest and most efficient solution for critical section synchronization.Peterson's Algorithm guarantees fairness and bounded waiting for all threads.Peterson's Algorithm is a scalable solution that can handle multiple threads and critical sections.

Why are atomic instructions important to locks?

Atomic instructions improve the performance of locks by reducing contention and allowing for more efficient synchronization.Atomic instructions simplify the implementation of locks by providing a standardized interface for concurrent access.Atomic instructions enable locks to work across different processor architectures and ensure portability.Atomic instructions ensure that locks can be acquired and released in an atomic and indivisible manner.

What are the conditions for a deadlock?

All threads require data input at the same time, which makes all threads blockedA thread that requires more resources will wait and hold onto what it currently has until it is granted its required resourcesA resource that has been allocated to one thread cannot be used by any other threadA thread's granted resources cannot be forcibly taken away for a different threadA thread exits while holding onto resourcesThere is a chain of threads where each thread is waiting for the next to give up its resources