Re: RFC: futex_wait() can DoS the tick

[Thomas Gleixner <tglx@linutronix.de>]

On Thu, 11 Jun 2015, Mike Galbraith wrote:
> On Wed, 2015-06-10 at 20:59 +0200, Thomas Gleixner wrote:
> Well, trying to do that like so...
> 
> trace-cmd record -M 8 -p function -e irq:* -e irq_vectors:local* -e timer:hrtimer*  -- sleep 1

> ..capture takes much more than a second...
> 
> LOC:     248161     226536      42091      38892   Local timer interrupts
> LOC:     248381     226783      42092      38901   Local timer interrupts
> LOC:     248668     227038      42092      38903   Local timer interrupts
> LOC:     248963     227277      42092      38904   Local timer interrupts
> LOC:     249214     227515      42092      38905   Local timer interrupts
> LOC:     249486     227732      42092      38905   Local timer interrupts
> LOC:     249720     227961      42092      38905   Local timer interrupts
> 
> ...but more importantly, when it gets cranked up..
> 
> homer:~/tmp # trace-cmd report > report
> homer:~/tmp # grep local_timer_entry report|wc -l
> 252
> 
> ...it scares the problem away.

Can you try the following, please?

Enable function tracer and hrtimer events manually. Then watch the irq
count on cpu3. If it stalls or becomes slow, then stop the trace with

      echo 0 >/sys/kernel/debug/tracing/tracing_on

If the overhead of the function tracer hides the problem, then try just
with hrtimer, sched_switch and irq events.

Thanks,

	tglx