Another synchronization mechanism that is provided in the Linux kernel is the Semaphore. It provides the ability for multiple kernel processes/threads to access a resource simultaneously using a “count” value. If the “count” value is 1, then the semaphore is similar to a mutual exclusion lock (Mutex). However, a Semaphore does not have the same ownership rules that a Mutex supports.
Some of the characteristics of the Semaphore is provided below
| 1 | Semaphore locks can sleep |
| 2 | Semaphore locks have a count value which allows greater than one instance of the lock to be acquired |
| 3 | Semaphores are not usually used in ISRs or time critical flows as the semaphore lock can go to sleep |
| 4 | Semaphore which has a “count” value of 1 is termed a binary Semaphore. |
| 5 | The “count” value during initialization depicts how many instances of the semaphore can be obtained |
In the next couple of articles, we will look at the semaphore structure, different semaphore APIs and also look at a simple usage example of a semaphore.
Pingback: Mutex API List and Sample API Code | Hitch Hiker's Guide to Learning