Are Semaphores the only way to achieve proper mutual exclusion while designing an OS?

[i_hate_toms]

The first three algorithms (turn-taking,P1using/P2using) don't guarantee system stability. So, are Semaphores (binary or otherwise) the only way to implement mutual exclusion for concurrent execution of a family of processes while designing a multiprogramming operating system kernel?

 

[NyxCharon]

Is this homework? Sure sounds like it.

 

[i_hate_toms]

Is this homework? Sure sounds like it.

Thnx for your reply.. Nah not homework, got this university test on "OS Designing concepts" next Monday (31st)

Why'd they schedule a test on the new years eve? Beats me.

A link or some suggestion about inter-process synchronization techniques, and race-condition avoidance strategies, would be highly appreciated.

{i have this book called "Operating Systems -- Concepts And Design", by Milan Milenkovic, 2nd edition, McGraw Hill, if someone here has a copy, it'd be great if u could point out which section to look at. I mean, the whole book is not in the syllabus for this test, and mutex is discussed all thru the book, can't really figure out where to look}