On Tue, 2007-01-02 at 12:25 -0500, Robert P. J. Day wrote: > On Tue, 2 Jan 2007, Ming Zhang wrote: > > > Hi All > > > > In this document coming with kernel source, it has > > > > " int atomic_inc_and_test(atomic_t *v); > > int atomic_dec_and_test(atomic_t *v); > > > > These two routines increment and decrement by 1, respectively, the > > given atomic counter. They return a boolean indicating whether the > > resulting counter value was zero or not. > > > > It requires explicit memory barrier semantics around the operation > > as above." > > > > Do this mean the code who call such functions should take care about > > the barrier? or such functions take care for us already? > > as i read it, it means that the memory barriers must be supplied by > *someone* -- in this case, they should be part of the functions so you > don't have to supply them yourself. > > an easy way to see this is to search the kernel source tree and see > examples of that routine, to see how often memory barriers are added: > > $ grep -r -A5 -B5 atomic_inc_and_test * | less > > as you can see, there don't appear to be explicit memory barriers, so > you shouldn't have to add them. make sense. also i read the doc again, should be in that way. thanks! > > rday -- http://blackmagic02881.wordpress.com/ -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
