Re: vmstat: On demand vmstat workers V8

[Christoph Lameter]

On Tue, 5 Aug 2014, Sasha Levin wrote:

> I can easily trigger it by cranking up the cpu hotplug code. Just try to
> frequently offline and online cpus, it should reproduce quickly.

Thats what I thought.

The test was done with this fix applied right?


Subject: vmstat ondemand: Fix online/offline races

Do not allow onlining/offlining while the shepherd task is checking
for vmstat threads.

On offlining a processor do the right thing cancelling the vmstat
worker thread if it exista and also exclude it from the shepherd
process checks.

Signed-off-by: Christoph Lameter <cl@xxxxxxxxx>

Index: linux/mm/vmstat.c
===================================================================
--- linux.orig/mm/vmstat.c	2014-07-30 09:35:54.602662306 -0500
+++ linux/mm/vmstat.c	2014-07-30 09:43:07.109037043 -0500
@@ -1317,6 +1317,7 @@ static void vmstat_shepherd(struct work_
 {
 	int cpu;

+	get_online_cpus();
 	/* Check processors whose vmstat worker threads have been disabled */
 	for_each_cpu(cpu, cpu_stat_off)
 		if (need_update(cpu) &&
@@ -1325,6 +1326,7 @@ static void vmstat_shepherd(struct work_
 			schedule_delayed_work_on(cpu, &per_cpu(vmstat_work, cpu),
 				__round_jiffies_relative(sysctl_stat_interval, cpu));

+	put_online_cpus();

 	schedule_delayed_work(&shepherd,
 		round_jiffies_relative(sysctl_stat_interval));
@@ -1380,8 +1382,8 @@ static int vmstat_cpuup_callback(struct
 		break;
 	case CPU_DOWN_PREPARE:
 	case CPU_DOWN_PREPARE_FROZEN:
-		if (!cpumask_test_and_set_cpu(cpu, cpu_stat_off))
-			cancel_delayed_work_sync(&per_cpu(vmstat_work, cpu));
+		cancel_delayed_work_sync(&per_cpu(vmstat_work, cpu));
+		cpumask_clear_cpu(cpu, cpu_stat_off);
 		break;
 	case CPU_DOWN_FAILED:
 	case CPU_DOWN_FAILED_FROZEN:
--
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/