>> The case that Manfred brought up was perfectly valid - we could do
>> __raw_atomic_foo() functions to do it if there is a need. We could
>> also just decide that atomic_foo() does not guarantee ordering.
Bret> Since you already have separate functions for memory barriers, I
Bret> would think not guarenteeing ordering would be the sensible
Bret> thing to do. Keep it simple.
Thats how I thought writel() was supposed to work. However making them
guarantee ordering means that new users do now get `cheated' by
mistake and the people who looked into it can use optimized versions.
Keeping it simple could be interpreted both ways 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/