2007-02-18 22:20:17

by Oleg Nesterov

Subject: [PATCH] fix refrigerator() vs thaw_process() race

refrigerator() can miss a wakeup, "wait event" loop needs a proper memory
ordering.

Signed-off-by: Oleg Nesterov <[email protected]>

--- WQ/kernel/power/process.c~WAKE 2007-02-18 22:56:49.000000000 +0300
+++ WQ/kernel/power/process.c 2007-02-19 01:04:26.000000000 +0300
@@ -46,8 +46,10 @@ void refrigerator(void)
recalc_sigpending(); /* We sent fake signal, clean it up */
spin_unlock_irq(&current->sighand->siglock);

- while (frozen(current)) {
- current->state = TASK_UNINTERRUPTIBLE;
+ for (;;) {
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ if (!frozen(current))
+ break;
schedule();
}
pr_debug("%s left refrigerator\n", current->comm);

2007-02-18 23:33:16

by Rafael J. Wysocki

Subject: Re: [PATCH] fix refrigerator() vs thaw_process() race

On Sunday, 18 February 2007 23:17, Oleg Nesterov wrote:
> refrigerator() can miss a wakeup, "wait event" loop needs a proper memory
> ordering.
>
> Signed-off-by: Oleg Nesterov <[email protected]>

ACK

> --- WQ/kernel/power/process.c~WAKE 2007-02-18 22:56:49.000000000 +0300
> +++ WQ/kernel/power/process.c 2007-02-19 01:04:26.000000000 +0300
> @@ -46,8 +46,10 @@ void refrigerator(void)
> recalc_sigpending(); /* We sent fake signal, clean it up */
> spin_unlock_irq(&current->sighand->siglock);
>
> - while (frozen(current)) {
> - current->state = TASK_UNINTERRUPTIBLE;
> + for (;;) {
> + set_current_state(TASK_UNINTERRUPTIBLE);
> + if (!frozen(current))
> + break;
> schedule();
> }
> pr_debug("%s left refrigerator\n", current->comm);
>
>
>

--
If you don't have the time to read,
you don't have the time or the tools to write.
- Stephen King

2007-02-19 10:50:32

by Pavel Machek

Subject: Re: [PATCH] fix refrigerator() vs thaw_process() race

Hi!

> refrigerator() can miss a wakeup, "wait event" loop needs a proper memory
> ordering.
>
> Signed-off-by: Oleg Nesterov <[email protected]>
>
> --- WQ/kernel/power/process.c~WAKE 2007-02-18 22:56:49.000000000 +0300
> +++ WQ/kernel/power/process.c 2007-02-19 01:04:26.000000000 +0300
> @@ -46,8 +46,10 @@ void refrigerator(void)
> recalc_sigpending(); /* We sent fake signal, clean it up */
> spin_unlock_irq(&current->sighand->siglock);
>
> - current->state = TASK_UNINTERRUPTIBLE;
> + set_current_state(TASK_UNINTERRUPTIBLE);


Looks okay to me... but this one liner would be exactly as effective,
right?

Pavel

> schedule();
> }
> pr_debug("%s left refrigerator\n", current->comm);

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2007-02-19 12:13:32

by Oleg Nesterov

Subject: Re: [PATCH] fix refrigerator() vs thaw_process() race

On 02/19, Pavel Machek wrote:
>
> > refrigerator() can miss a wakeup, "wait event" loop needs a proper memory
> > ordering.
> >
> > Signed-off-by: Oleg Nesterov <[email protected]>
> >
> > --- WQ/kernel/power/process.c~WAKE 2007-02-18 22:56:49.000000000 +0300
> > +++ WQ/kernel/power/process.c 2007-02-19 01:04:26.000000000 +0300
> > @@ -46,8 +46,10 @@ void refrigerator(void)
> > recalc_sigpending(); /* We sent fake signal, clean it up */
> > spin_unlock_irq(&current->sighand->siglock);
> >
> > - current->state = TASK_UNINTERRUPTIBLE;
> > + set_current_state(TASK_UNINTERRUPTIBLE);
>
>
> Looks okay to me... but this one liner would be exactly as effective,
> right?

I think no, with this one liner we have

while (frozen(current)) {
// ------ WINDOW ------------
set_current_state(TASK_UNINTERRUPTIBLE);
schedule();
}

What if thaw_process() happens in the window above?

We need the barrier exactly because LOAD (check condition) should not
come before STORE (set task->state).

Oleg.

2007-02-20 23:23:25

by Pavel Machek

Subject: Re: [PATCH] fix refrigerator() vs thaw_process() race

Hi!

> > > refrigerator() can miss a wakeup, "wait event" loop needs a proper memory
> > > ordering.
> > >
> > > Signed-off-by: Oleg Nesterov <[email protected]>
> > >
> > > --- WQ/kernel/power/process.c~WAKE 2007-02-18 22:56:49.000000000 +0300
> > > +++ WQ/kernel/power/process.c 2007-02-19 01:04:26.000000000 +0300
> > > @@ -46,8 +46,10 @@ void refrigerator(void)
> > > recalc_sigpending(); /* We sent fake signal, clean it up */
> > > spin_unlock_irq(&current->sighand->siglock);
> > >
> > > - current->state = TASK_UNINTERRUPTIBLE;
> > > + set_current_state(TASK_UNINTERRUPTIBLE);
> >
> >
> > Looks okay to me... but this one liner would be exactly as effective,
> > right?
>
> I think no, with this one liner we have
>
> while (frozen(current)) {
> // ------ WINDOW ------------
> set_current_state(TASK_UNINTERRUPTIBLE);
> schedule();
> }
>
> What if thaw_process() happens in the window above?
>
> We need the barrier exactly because LOAD (check condition) should not
> come before STORE (set task->state).

Aha, ok :-). You already have my ACK ;-).
Pavel

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html