Re: semaphore and spin locks

[Thomas Hofer <th@monochrom.at>]

atul wrote (Dienstag, 10. Oktober 2000 03:25):

> Afa i know semaphores provide a very good mechanism for
> synchronization so why spin locks are needed n how they differ from
> semaphores..??

I'm just a kernelnewbie, so correct me if I'm wrong:

A spin lock is a primitive mechanism to prevent two or more processes 
accessing a certain data structure at the same time. The lock is an 
integer that's initially set to 0. A process which wants to enter the 
critical area has to set the lock to 1 using an atomic "if 0 then set 
1"-operation. If the lock is in use, the process has to busy-wait, 
spinning in a tight loop. ==> CPU-time-consuming if you have to wait 
long, but very efficient if you don't have to wait.

A semaphore on the other hand is a more complicated mechanism consiting 
of a count, a wait-queue and a lock. Processes which wait for a 
semaphore don't do busy waiting, but have to do a context switch, which 
takes some time. So a simple spin lock can be more efficient if you 
expect that there is a good chance that no process has to wait long for 
the lock. But you have to use it with care.

Thomas.
-
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
IRC Channel:   irc.openprojects.net / #kernelnewbies
Web Page:      http://www.kernelnewbies.org/