Re: + kthread-add-a-missing-memory-barrier-to-kthread_stop.patch added to -mm tree

["Dmitry Adamushko" <dmitry.adamushko@gmail.com>]

On 23/02/2008, Linus Torvalds <torvalds@linux-foundation.org> wrote:
>
>
>  On Sat, 23 Feb 2008, Oleg Nesterov wrote:
>  >
>
> > Yes, but still I suspect wmb() is not enough. Note that try_to_wake_up()
>  > first checks (reads) the task->state,
>  >
>  >       if (!(old_state & state))
>  >               goto out;
>  >
>  > without the full mb() it is (in theory) possible that try_to_wake_up()
>  > first reads TASK_RUNNING and only then sets CONDITION. IOW, STORE and
>  > LOAD could be re-ordered.
>
>
> No. The spinlock can have preceding stores (and loads, for that matter)
>  percolate *into* the locked region, but a spinlock can *not* have loads
>  (and stores) escape *out* of the region withou being totally broken.

it's not a LOAD that escapes *out* of the region. It's a MODIFY that gets *in*:

(1)

MODIFY(a);

LOCK

LOAD(b);

UNLOCK


can become:

(2)

LOCK

MOFIDY(a)
LOAD(b);

UNLOCK

and (reordered)

(3)

LOCK

LOAD(a)
MODIFY(b)

UNLOCK

and this last one is a problem. No?


>
>                 Linus
>

-- 
Best regards,
Dmitry Adamushko