Re: [bisected] Clocksource tsc unstable git

[Peter Zijlstra]

On Tue, 2010-11-09 at 14:21 +0100, Markus Trippelsdorf wrote:
> 
> You've missed the fact that Peter already has a patch that fixes the
> problem, but he never bothered to post it in this thread.
> 
Right, forgot about that, sorry.

Still Heiko has a good point, I've queued the below after I realized
that PREEMPT_ACTIVE would still require two tests and the on_rq bit is
actually sufficient.

---
Subject: sched: Fix runnable condition for stoptask
From: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
Date: Tue Nov 09 14:36:52 CET 2010

Heiko reported that the TASK_RUNNING check is not sufficient for
CONFIG_PREEMPT=y since we can get preempted with !TASK_RUNNING.

He suggested adding a ->se.on_rq test to the existing TASK_RUNNING
one, however TASK_RUNNING will always have ->se.on_rq, so we might as
well reduce that to a single test.

Reported-by: Heiko Carstens <heiko.carstens@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
LKML-Reference: <new-submission>
---
 kernel/sched_stoptask.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6/kernel/sched_stoptask.c
===================================================================
--- linux-2.6.orig/kernel/sched_stoptask.c
+++ linux-2.6/kernel/sched_stoptask.c
@@ -26,7 +26,7 @@ static struct task_struct *pick_next_tas
 {
 	struct task_struct *stop = rq->stop;
 
-	if (stop && stop->state == TASK_RUNNING)
+	if (stop && stop->se.on_rq)
 		return stop;
 
 	return NULL;

--
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/