Re: clock drift in set_task_cpu()

From: Philby John
Date: Mon Aug 09 2010 - 10:56:52 EST


On Mon, 2010-08-09 at 18:47 +0530, Jack Daniel wrote:
> On Thu, Aug 5, 2010 at 3:28 PM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> > On Wed, 2010-07-21 at 17:10 +0530, Jack Daniel wrote:
> >> On a Xeon 55xx with 8 CPU's, I found out the new_rq->clock value is
> >> sometimes larger than old_rq->clock and so clock_offset tends to warp
> >> around leading to incorrect values.
> >
> > What values get incorrect, do you observe vruntime funnies or only the
> > schedstat values?
>
> Just the schedstat values, did not observe anything wrong with vruntime.
>
> >
> >> You have very correctly noted in
> >> the commit header that all functions that access set_task_cpu() must
> >> do so after a call to sched_clock_remote(), in this case the function
> >> is sched_fork(). I validated by adding update_rq_clock(old_rq); into
> >> set_task_cpu() and that seems to fix the issue.
> >
> > Ah, so the problem is that task_fork_fair() does the task placement
> > without updated rq clocks? In which case I think we should at least do
> > an update_rq_clock(rq) in there (see the below patch).
>
> Yes, this is indeed the problem and your patch seems to fix the issue.
>
> >
> >> But I noticed that
> >> since CONFIG_HAVE_UNSTABLE_SCHED_CLOCK is already set, if
> >> (sched_clock_stable) in sched_clock_cpu() will yield to true and the
> >> flow never gets to sched_clock_remote() or sched_clock_local().
> >
> > sched_clock_stable being true implies the clock is stable across cores
> > and thus it shouldn't matter. Or are you saying you're seeing it being
> > set and still have issues?
> >
>
> Please ignore these comments, initial debugging set me on the wrong
> foot, to suggest that TSC is unstable.
>
> > diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
> > index 9910e1b..f816e74 100644
> > --- a/kernel/sched_fair.c
> > +++ b/kernel/sched_fair.c
> > @@ -3751,6 +3751,8 @@ static void task_fork_fair(struct task_struct *p)
> >
> > raw_spin_lock_irqsave(&rq->lock, flags);
> >
> > + update_rq_clock(rq);
>
> As you rightly pointed out above, updating the clocks in
> task_fork_fair() will rightly fix the issue. Can get rid of rest of
> the update_rq_clock() functions as they (like you said), are expensive
> and I tested commenting them out.

>From 1bc695bc2ac6c941724953b29f6c18196a474b8f Mon Sep 17 00:00:00 2001
From: Philby John <pjohn@xxxxxxxxxx>
Date: Mon, 9 Aug 2010 18:19:08 +0530
Subject: [PATCH] sched: ensure rq->clock get sync'ed when migrating tasks

In sched_fork() when we do task placement in ->task_fork_fair()
ensure we update_rq_clock() so we work with current time. This has
been noted and verified on an Intel Greencity (Xeon 55xx)

Signed-off-by: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
Signed-off-by: Philby John <pjohn@xxxxxxxxxx>
---
kernel/sched_fair.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 806d1b2..48bc31c 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -3751,7 +3751,7 @@ static void task_fork_fair(struct task_struct *p)
unsigned long flags;

raw_spin_lock_irqsave(&rq->lock, flags);
-
+ update_rq_clock(rq);
if (unlikely(task_cpu(p) != this_cpu))
__set_task_cpu(p, this_cpu);

--
1.6.3.3.333.g4d53f




--
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/