Re: High context switch rate, ksoftirqd's chewing cpu

[Tejun Heo <tj@kernel.org>]

Hello,

On Sat, Dec 01, 2012 at 12:06:41PM +0100, Zlatko Calusic wrote:
> I have good news. The patch fixes the regression!
> 
> To doublecheck and provide you additional data, I updated to the latest Linus
> kernel (commit 7c17e48), recompiled (WITHOUT the patch), rebooted and this is
> what vmstat 1 looks like:

Awesome, can you please test the following patch too?  Thanks!

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 042d221..94964d1 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -1364,6 +1364,11 @@ static void __queue_delayed_work(int cpu, struct workqueue_struct *wq,
 	BUG_ON(timer_pending(timer));
 	BUG_ON(!list_empty(&work->entry));
 
+	if (!delay) {
+		__queue_work(cpu, wq, &dwork->work);
+		return;
+	}
+
 	timer_stats_timer_set_start_info(&dwork->timer);
 
 	/*
@@ -1417,9 +1422,6 @@ bool queue_delayed_work_on(int cpu, struct workqueue_struct *wq,
 	bool ret = false;
 	unsigned long flags;
 
-	if (!delay)
-		return queue_work_on(cpu, wq, &dwork->work);
-
 	/* read the comment in __queue_work() */
 	local_irq_save(flags);