Re: [PATCH] sched: don't clear PF_THREAD_BOUND in select_fallback_rq

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

* Qiang Huang | 2013-04-25 17:01:18 [+0800]:

So I looked again at this.

>This is revert of "sched-clear-pf-thread-bound-on-fallback-rq.patch"
>(commit 0d939066acdcb in v3.4-rt),.
>
>Select_fallback_rq() can be easilly called during system boot, because
>select_task_rq_fair() just return task_cpu(p) for bounded kernel threads,
>which is 0 during system boot and not in tsk_cpus_allowed, so
>select_fallback_rq() is called and PF_THREAD_BOUND is cleared. In my
>box, 1/3 bounded kernel threads will clear that flag after boot.

Please tell me _which_ threads lose this flag. I don't see this…

>And it will cause problems, for example:
># for pid in `ps -e -o pid`; do taskset -p -c 0-15 $pid; done
>this command will cause system hung.

I tracked this down and I grabbed a brown paper bag…
So to trigger this, it seems enough to 
| taskset -p -c 5 2
| echo 0 > /sys/devices/system/cpu/cpu5/online

and the fix is:

Subject: [PATCH] kernel/cpu: fix cpu down problems

During forward porting I mixed up those two calls. The CPU up/down
worked usually. However a migrate_disable() in one of the notifier could
freeze the system.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 kernel/cpu.c |   13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index d44dea3..0784023 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -557,11 +557,6 @@ static int __ref _cpu_down(unsigned int cpu, int tasks_frozen)
 	}
 
 	cpu_hotplug_begin();
-	err = cpu_unplug_begin(cpu);
-	if (err) {
-		printk("cpu_unplug_begin(%d) failed\n", cpu);
-		goto out_cancel;
-	}
 
 	err = __cpu_notify(CPU_DOWN_PREPARE | mod, hcpu, -1, &nr_calls);
 	if (err) {
@@ -569,8 +564,16 @@ static int __ref _cpu_down(unsigned int cpu, int tasks_frozen)
 		__cpu_notify(CPU_DOWN_FAILED | mod, hcpu, nr_calls, NULL);
 		printk("%s: attempt to take down CPU %u failed\n",
 				__func__, cpu);
+		goto out_cancel;
+	}
+
+	err = cpu_unplug_begin(cpu);
+	if (err) {
+		printk("cpu_unplug_begin(%d) failed\n", cpu);
+		__cpu_notify(CPU_DOWN_FAILED | mod, hcpu, nr_calls, NULL);
 		goto out_release;
 	}
+
 	smpboot_park_threads(cpu);
 
 	/* Notifiers are done. Don't let any more tasks pin this CPU. */
-- 
1.7.10.4

Does this fix your problems.

Sebastian
--
To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html