Mutex vs Semaphore
Mutex:
Is a key to a toilet. One person can have the key – occupy the toilet – at the time. When finished, the person gives (frees) the key to the next person in the queue.
Officially: “Mutexes are typically used to serialise access to a section of re-entrant code that cannot be executed concurrently by more than one thread. A mutex object only allows one thread into a controlled section, forcing other threads which attempt to gain access to that section to wait until the first thread has exited from that section.”
Ref: Symbian Developer Library
Semaphore:
Is the number of free identical toilet keys. Example, say we have four toilets with identical locks and keys. The semaphore count – the count of keys – is set to 4 at beginning (all four toilets are free), then the count value is decremented as people are coming in. If all toilets are full, ie. there are no free keys left, the semaphore count is 0. Now, when eq. one person leaves the toilet, semaphore is increased to 1 (one free key), and given to the next person in the queue.
Officially: “A semaphore restricts the number of simultaneous users of a shared resource up to a maximum number. Threads can request access to the resource (decrementing the semaphore), and can signal that they have finished using the resource (incrementing the semaphore).”
Difference Between Semaphores and Mutex
After reading though the material above, some pretty clear distinctions should have emerged. However, I’d like to reiterate those differences again here, along with some other noticeable differences between semaphore and Mutex.
- A semaphore can be a Mutex but a Mutex can never be semaphore. This simply means that a binary semaphore can be used as Mutex, but a Mutex can never exhibit the functionality of semaphore.
- Both semaphores and Mutex (at least the on latest kernel) are non-recursive in nature.
- No one owns semaphores, whereas Mutex are owned and the owner is held responsible for them. This is an important distinction from a debugging perspective.
- In case the of Mutex, the thread that owns the Mutex is responsible for freeing it. However, in the case of semaphores, this condition is not required. Any other thread can signal to free the semaphore by using the
sem_post()function. - A Mutex, by definition, is used to serialize access to a section of re-entrant code that cannot be executed concurrently by more than one thread. A semaphore, by definition, restricts the number of simultaneous users of a shared resource up to a maximum number
- Another difference that would matter to developers is that semaphores are system-wide and remain in the form of files on the filesystem, unless otherwise cleaned up. Mutex are process-wide and get cleaned up automatically when a process exits.
- The nature of semaphores makes it possible to use them in synchronizing related and unrelated process, as well as between threads. Mutex can be used only in synchronizing between threads and at most between related processes (the pthread implementation of the latest kernel comes with a feature that allows Mutex to be used between related process).
- According to the kernel documentation, Mutex are lighter when compared to semaphores. What this means is that a program with semaphore usage has a higher memory footprint when compared to a program having Mutex.
- From a usage perspective, Mutex has simpler semantics when compared to semaphores.
- A Mutex allows serial access to a resource, whereas semaphores, in addition to allowing serial access, could also be used to access resources in parallel. For example, consider resource R being accessed by n number of users. When using a Mutex, we would need a Mutex “m” to lock and unlock the resource, thus allowing only one user at a time to use the resource R. In contrast, semaphores can allow n number of users to synchronously access the resource R.
- The advantage of semaphores over other synchronization mechanisms is that they can be used to synchronize two related or unrelated processes trying to access the same resource.
As per operating system terminology, the mutex and semaphore are kernel resources that provide synchronization services (also called as synchronization primitives). Why do we need such synchronization primitives? Won’t be only one sufficient? To answer these questions, we need to understand few keywords. Please read the posts on atomicity and critical section. We will illustrate with examples to understand these concepts well, rather than following usual OS textual description.
The producer-consumer problem:
Note that the content is generalized explanation. Practical details will vary from implementation.
Consider the standard producer-consumer problem. Assume, we have a buffer of 4096 byte length. A producer thread will collect the data and writes it to the buffer. A consumer thread will process the collected data from the buffer. Objective is, both the threads should not run at the same time.
Using Mutex:
A mutex provides mutual exclusion, either producer or consumer can have the key (mutex) and proceed with their work. As long as the buffer is filled by producer, the consumer needs to wait, and vice versa.
At any point of time, only one thread can work with the entire buffer. The concept can be generalized using semaphore.
Using Semaphore:
A semaphore is a generalized mutex. In lieu of single buffer, we can split the 4 KB buffer into four 1 KB buffers (identical resources). A semaphore can be associated with these four buffers. The consumer and producer can work on different buffers at the same time.
Misconception:
There is an ambiguity between binary semaphore and mutex. We might have come across that a mutex is binary semaphore. But they are not! The purpose of mutex and semaphore are different. May be, due to similarity in their implementation a mutex would be referred as binary semaphore.
Strictly speaking, a mutex is locking mechanism used to synchronize access to a resource. Only one task (can be a thread or process based on OS abstraction) can acquire the mutex. It means there will be ownership associated with mutex, and only the owner can release the lock (mutex).
Semaphore is signaling mechanism (“I am done, you can carry on” kind of signal). For example, if you are listening songs (assume it as one task) on your mobile and at the same time your friend called you, an interrupt will be triggered upon which an interrupt service routine (ISR) will signal the call processing task to wakeup.
General Questions:
1. Can a thread acquire more than one lock (Mutex)?
Yes, it is possible that a thread will be in need of more than one resource, hence the locks. If any lock is not available the thread will wait (block) on the lock.
2. Can a mutex be locked more than once?
A mutex is a lock. Only one state (locked/unlocked) is associated with it. However, a recursive mutex can be locked more than once (POSIX complaint systems), in which a count is associated with it, yet retains only one state (locked/unlocked). The programmer must unlock the mutex as many number times as it was locked.
3. What will happen if a non-recursive mutex is locked more than once.
Deadlock. If a thread which had already locked a mutex, tries to lock the mutex again, it will enter into the waiting list of that mutex, which results in deadlock. It is because no other thread can unlock the mutex. An operating system implementer can exercise care in identifying the owner of mutex and return if it is already locked by same thread to prevent deadlocks.
4. Are binary semaphore and mutex same?
No. We will suggest to treat them separately, as it was explained signalling vs locking mechanisms. But a binary semaphore may experience the same critical issues (e.g. priority inversion) associated with mutex. We will cover these later article.
A programmer can prefer mutex rather than creating a semaphore with count 1.
5. What is a mutex and critical section?
Some operating systems use the same word critical section in the API. Usually a mutex is costly operation due to protection protocols associated with it. At last, the objective of mutex is atomic access. There are other ways to achieve atomic access like disabling interrupts which can be much faster but ruins responsiveness. The alternate API makes use of disabling interrupts.
6. What are events?
The semantics of mutex, semaphore, event, critical section, etc… are same. All are synchronization primitives. Based on their cost in using them they are different. We should consult the OS documentation for exact details.
7. Can we acquire mutex/semaphore in an Interrupt Service Routine?
An ISR will run asynchronously in the context of current running thread. It is not recommended to query (blocking call) the availability of synchronization primitives in an ISR. The ISR are meant be short, the call to mutex/semaphore may block the current running thread. However, an ISR can signal a semaphore or unlock a mutex.
8. What we mean by “thread blocking on mutex/semaphore” when they are not available?
Every synchronization primitive will have waiting list associated with it. When the resource is not available, the requesting thread will be moved from the running list of processor to the waiting list of the synchronization primitive. When the resource is available, the higher priority thread on the waiting list will get resource (more precisely, it depends on the scheduling policies).
9. Is it necessary that a thread must block always when resource is not available?
Not necessarily. If the design is sure ‘what has to be done when resource is not available‘, the thread can take up that work (a different code branch). To support application requirements the OS provides non-blocking API.
For example POSIX pthread_mutex_trylock() API. When the mutex is not available the function will return immediately where as the API pthread_mutex_lock() will block the thread till resource is available.
Nice articles on the topic:
- MUTEX VS. SEMAPHORES – PART 1: SEMAPHORES
- MUTEX VS. SEMAPHORES – PART 2: THE MUTEX
- MUTEX VS. SEMAPHORES – PART 3 (FINAL PART): MUTUAL EXCLUSION PROBLEMS
From part 2:
The mutex is similar to the principles of the binary semaphore with one significant difference: the principle of ownership. Ownership is the simple concept that when a task locks (acquires) a mutex only it can unlock (release) it. If a task tries to unlock a mutex it hasn’t locked (thus doesn’t own) then an error condition is encountered and, most importantly, the mutex is not unlocked. If the mutual exclusion object doesn't have ownership then, irrelevant of what it is called, it is not a mutex.
From: