Re: [Xen-devel] BUG: NOW() seems to (sometimes) go backwards!

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 08.06.16 at 15:43, <dario.faggioli@xxxxxxxxxx> wrote:
> On Wed, 2016-06-08 at 04:42 -0600, Jan Beulich wrote:
>> How big of system are we talking about? I'm asking to assess the
>> overhead of adding some cross-CPU checks to get_s_time_fixed()
>> (in a debugging patch), logging relevant values if non-monotonic
>> output gets detected. (On too big a system, the overhead here
>> might end up masking the problem.)
>> 
> Yeah, I sort of tried doing something like that already, but was
> logging the wrong thing (I was not yet suspecting a problem with
> scaling).

Yeah, I did also realize that with you observing the problem even
within a core, the cross-CPU checking could be limited to just the
other SMT sibling, which should bring the overhead down quite a
bit. Otoh, with it - as you say - often just being a couple of nano-
seconds, even that may then already be too much.

Speaking of overhead: The tracing itself that you use does not
influence the picture in any way, does it?

Considering how easily you say you see the problem, it ought to
be more widespread. I'm considering to try to do some such
instrumentation to see whether I can see the issue anywhere.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel