Messages in this thread | | | Subject | Re: ordered memory access | From | Jes Sorensen <> | Date | 30 Sep 1999 16:06:43 +0200 |
| |
>>>>> "Andrea" == Andrea Arcangeli <andrea@suse.de> writes:
Andrea> On 30 Sep 1999, Jes Sorensen wrote: >> atomic_foo() should be ordered, at least that the idea behind it as
Andrea> It isn't. Also the atomic_inc/dec and set_bit clear_bit Andrea> aren't.
Andrea> Only test_and_*_bit and atomic_dec_and_test enforce ordering Andrea> between previous and following code.
Aren't according to what architecture? Some architectures can only implement these by using a spin lock, thus they will certainly ensure ordering. My point here is, should we specify these functions to guarantee ordering, yes or no?
Intuitively I think it makes sense to do so, and I doubt it will be a performance hit at all, but I am just guessing here.
Jes
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |