Re: [PATCH v7 1/4] spinlock: A new lockref structure for locklessupdate of refcount

[Linus Torvalds]

On Thu, Aug 29, 2013 at 7:30 PM, Benjamin Herrenschmidt
<benh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> Or we can keep both completely separate like Linus does on x86.

I did it that way mainly to minimize the patch.

I agree with you that it probably makes sense to layer them the other
way around from what Michael's patch did, iow implement
arch_spin_is_locked() in terms of arch_spin_value_unlocked().

That said, on power, you have that "ACCESS_ONCE()" implicit in the
*type*, not in the code, so an "arch_spinlock_t" is fundamentally
volatile in itself. It's one of the reasons I despise "volatile":
things like volatility are _not_ attributes of a variable or a type,
but of the code in question. Something can be volatile in one context,
but not in another (one context might be locked, for example).

              Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/