i'd like to clarify the different varieties of semaphores and mutexes and if someone can fill in the blanks, i'd appreciate it. as i read it, first there were general semaphores, defined in the include/asm-*/semaphore.h files. these are general counting semaphores, and can be used in places only where contending tasks could sleep, so they're not suitable for, say, interrupt handlers. (they can't be used in interrupt context.) in those same semaphore.h files, there are also special cases to declare binary semaphores using DECLARE_MUTEX and init_MUTEX. however, while these advertise the word "mutex", they are simply a general semaphore with possible values of 0 and 1 (a binary semaphore), so they have all the same power and restrictions of a normal semaphore, there's just a convenient syntax to declare the restricted binary case. and, finally, there is the newer actual "mutex", described in Documentation/mutex-design.txt, which has simpler semantics and some restrictions that semaphores don't. is this about right so far? so, in terms of cleaning up mutexes in the source tree, it's possible to replace a "binary" semaphore with an actual mutex, as long as that binary semaphore is being used in such a way that all of the restrictions of mutexes are satisfied (and, certainly, not all of them are). also, it's unfortunate that the semaphore.h files declare binary semaphores using macros that contain the word "mutex" since that just *begs* to be confused with actual mutexes. rday -- ======================================================================== Robert P. J. Day Linux Consulting, Training and Annoying Kernel Pedantry Waterloo, Ontario, CANADA http://fsdev.net/wiki/index.php?title=Main_Page ======================================================================== -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
