Operating System - Galvin - Chapter 3

Question

Galvin, 9th edition on page 146 states that

“Shared memory can be faster than message passing, since message-passing systems are typically implemented using system calls and thus require the more time-consuming task of kernel intervention. In shared-memory systems, system calls are required only to establish shared memory regions. Once shared memory is established, all accesses are treated as routine memory accesses, and no assistance from the kernel is required.”

 

I agree that shared memory is bound to be faster, but aren’t the mutex and semaphore calls require Kernel intervention?

Comments

Do we always need semaphores to access shared memory?

And how are semaphores implemented?

---

wait(semaphore *S) { 
    S->value--;
    if (S->value < 0) {
        add this process to S->list;
        block();
    }
}

signal(semaphore *S) {
    S->value++;
    if (S->value <= 0) {
        remove a process P from S->list;
    wakeup(P);
    }
}

Page 216 says that block() and wakeup are basic system calls provided by OS. And system calls require kernel intervention.

---

That's a strong semaphore, you have described that requires a queue. If it is between two processes under your control, then you can create a custom mutex by using the test-set operation on a shared space. That way you can avoid calls to the kernel.