Timeline for Is a mutex lock always implemented as spin waiting?
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 9, 2020 at 18:44 | comment | added | Tim | I meant softwareengineering.stackexchange.com/questions/418800/… | |
| Nov 9, 2020 at 18:42 | comment | added | JacquesB | @Tim: This is answered in the second quote in the question. | |
| Nov 9, 2020 at 18:26 | comment | added | Tim | I meant: what differences are between a binary semaphore and a mutex lock, to a user? | |
| Nov 9, 2020 at 18:06 | comment | added | Useless | Depends on contention patterns and whether the user would like their cores and electricity to be doing something more productive than waiting for each other. It's a question in its own right, but try some research some before posting it. | |
| Nov 9, 2020 at 18:06 | comment | added | JacquesB | @Tim: A spin lock takes up processor resources when spinning, a block does not since the thread is suspended by the OS. But blocking requires a system call, so it has some overhead. This is why an implementation may combine them. If a lock is only locked for at short while, it is more efficient to spin wait. But if it is locked for longer than the system call overhed, it is better to block. | |
| Nov 9, 2020 at 17:56 | comment | added | Tim | @JacquesB Both a binary semaphore and a mutex lock can be implemented as spin waiting, blocking or both. What differences are between them, to a user? | |
| Nov 9, 2020 at 17:42 | comment | added | JacquesB | @Tim: A semaphore can also be spin waiting or blocking or both, just like a mutex. | |
| Nov 9, 2020 at 17:22 | comment | added | Robert Harvey | @Tim: I suggest you study some of these mechanisms; there is source code available on the Internet for various mutex implementations. As it is, we're mostly talking about word definitions here, which won't get you very far. Once you understand the actual mechanisms, the meaning of the word definitions should become readily apparent. | |
| Nov 9, 2020 at 17:20 | comment | added | Tim | Thanks. If a mutex lock can be implemented as block waiting, is a mutex lock implemented as such the same as a binary semaphore? (Stalling's OS book says a mutex lock and a binary semaphore differ in whether the process that locks the mutex (sets the value to zero) must be the one to unlock it. It doesn't mention whether they differ in spinning waiting only. See the quote.) | |
| Nov 9, 2020 at 17:12 | history | answered | JacquesB | CC BY-SA 4.0 |