Re: SMP suspend broken due to "swsusp: Change code ordering in disk.c" et al.

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Saturday, 24 February 2007 09:57, Rafael J. Wysocki wrote:
> On Saturday, 24 February 2007 01:31, Johannes Berg wrote:
> > Hi,
> > 
> > > Could you please try the appended patch?  It's somewhat hackish, but may work.
> > 
> > Not before Monday or Tuesday unfortunately, I'm away from the machine.
> 
> Okay, no big deal.  I think I'll be able to reproduce the problem here. ;-)
> 
> > Maybe I can find someone else willing to test who has the same box.
> > 
> > > The idea is to do nothing on CPU unplug if the CPU's worker thread is frozen
> > > and later check in the thread itself if it has been replaced by another one.
> > 
> > Yeah that should work. That patch won't apply to my tree though unless I
> > actually go to -mm2 as well :)
> 
> Ah, sorry.  I'll prepare a version against 2.6.21-rc1.  Apart from this, I
> think there's a better solution, but I have to verify that.

No, this one is better IMO.

Appended is the version against 2.6.21-rc1.  It seems to work for me, but I'd
like someone else to confirm it.

Greetings,
Rafael

 kernel/workqueue.c |   18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

Index: linux-2.6.21-rc1/kernel/workqueue.c
===================================================================
--- linux-2.6.21-rc1.orig/kernel/workqueue.c
+++ linux-2.6.21-rc1/kernel/workqueue.c
@@ -376,8 +376,19 @@ static int worker_thread(void *__cwq)
 
 	set_current_state(TASK_INTERRUPTIBLE);
 	while (!kthread_should_stop()) {
-		if (cwq->freezeable)
-			try_to_freeze();
+		if (try_to_freeze()) {
+			/* We've just left the refrigerator.  If our CPU is
+			 * a nonboot one, we might have been replaced.
+			 * The lock is taken to prevent the race with
+			 * cleanup_workqueue_thread() from happening
+			 */
+			spin_lock_irq(&cwq->lock);
+			if (cwq->thread != current) {
+				spin_unlock_irq(&cwq->lock);
+				break;
+			}
+			spin_unlock_irq(&cwq->lock);
+		}
 
 		add_wait_queue(&cwq->more_work, &wait);
 		if (list_empty(&cwq->worklist))
@@ -539,6 +550,9 @@ static void cleanup_workqueue_thread(str
 	cwq = per_cpu_ptr(wq->cpu_wq, cpu);
 	spin_lock_irqsave(&cwq->lock, flags);
 	p = cwq->thread;
+	if (p && frozen(p))
+		p = NULL;
+
 	cwq->thread = NULL;
 	spin_unlock_irqrestore(&cwq->lock, flags);
 	if (p)