Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 09 Feb 2022 01:17:05 +0000
From:      bugzilla-noreply@freebsd.org
To:        acpi@FreeBSD.org
Subject:   [Bug 173541] High (0.60+, 1.00) idle load averages
Message-ID:  <bug-173541-16045-c7Jw2oW0pL@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-173541-16045@https.bugs.freebsd.org/bugzilla/>
References:  <bug-173541-16045@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D173541

--- Comment #27 from Alexander Motin <mav@FreeBSD.org> ---
@Kubilay Kocak, I don't think this should be related to timecounters.  It m=
ust
be a coincidence.  As I have written 7 years ago in comment 9, I believe it=
 is
a result of some synchronization between load average callouts and some oth=
er
events.  Since you've mentioned HPET, I'd think towards not HPET as
timecounter, but as HPET as eventtimer.  Depending on hardware and number of
CPUs, kernel may have to use single HPET interrupt to deliver timer interru=
pts
to multiple CPUs.  As result, it is very simple to have two or more CPU to =
wake
up from a long sleep same time (one where HPET interrupt is routed to and
another where timer event is actually scheduled), that may cause load avera=
ge
callout to sample simultaneous threads.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-173541-16045-c7Jw2oW0pL>