From: "H. Peter Anvin" <hpa@zytor.com>
To: Zachary Amsden <zach@vmware.com>
Cc: Anthony Liguori <anthony@codemonkey.ws>,
Jeremy Fitzhardinge <jeremy@goop.org>,
Alok Kataria <akataria@vmware.com>,
"avi@redhat.com" <avi@redhat.com>,
Rusty Russell <rusty@rustcorp.com.au>,
Gerd Hoffmann <kraxel@redhat.com>, Ingo Molnar <mingo@elte.hu>,
the arch/x86 maintainers <x86@kernel.org>,
LKML <linux-kernel@vger.kernel.org>,
"Nakajima, Jun" <jun.nakajima@intel.com>,
Daniel Hecht <dhecht@vmware.com>,
"virtualization@lists.linux-foundation.org"
<virtualization@lists.linux-foundation.org>,
"kvm@vger.kernel.org" <kvm@vger.kernel.org>
Subject: Re: [RFC] CPUID usage for interaction between Hypervisors and Linux.
Date: Wed, 01 Oct 2008 18:21:56 -0700 [thread overview]
Message-ID: <48E42234.9000808@zytor.com> (raw)
In-Reply-To: <1222909900.7330.106.camel@bodhitayantram.eng.vmware.com>
Zachary Amsden wrote:
>
> I'm not suggesting using the nominal value. I'm suggesting the
> measurement be done in the one and only place where there is perfect
> control of the system, the processor boot-strapping in the BIOS.
>
> Only the platform designers themselves know the speed of the oscillator
> which is modulating the clock and so only they should be calibrating the
> speed of the TSC.
>
No. *Noone*, including the manufacturers, know the speed of the
oscillator which is modulating the clock. What you have to do is
average over a timespan which is long enough that the SSM averages out
(a relatively small fraction of a second.)
As for trusting the BIOS on this, that's a total joke. Firmware vendors
can't get the most basic details right.
> If this modulation really does alter the frequency by +/- 2% (seems high
> to me, but hey, I don't design motherboards), using an LFO, then
> basically all the calibration done in Linux is broken and has been for
> some time. You can't calibrate only once, or risk being off by 2%, you
> can't calibrate repeatedly and take the fastest estimate, or you are off
> by 2%, and you can't calibrate repeatedly and take the average without
> risking SMI noise affecting the lowest clock speed measurement,
> contributing unknown error.
You have to calibrate over a sample interval long enough that the SSM
averages out.
> Hmm. Re-reading your e-mail, I see you are saying the nominal frequency
> may be off by 2% (and I easily believe that), not necessarily that the
> frequency modulation may be 2% (which I still think is high). Does
> anyone know what the actual bounds on spread spectrum modulation are or
> how fast the clock is modulated?
No, I'm saying the frequency modulation may be up to 2%. Typically it
is something like [-2%,+0%].
-hpa
next prev parent reply other threads:[~2008-10-02 1:25 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-10-01 17:14 [RFC] CPUID usage for interaction between Hypervisors and Linux Alok Kataria
2008-10-01 17:21 ` H. Peter Anvin
2008-10-01 17:33 ` Alok Kataria
2008-10-01 17:45 ` H. Peter Anvin
2008-10-01 18:06 ` Jeremy Fitzhardinge
2008-10-01 21:05 ` Alok Kataria
2008-10-01 22:46 ` H. Peter Anvin
2008-10-02 1:11 ` Nakajima, Jun
2008-10-02 1:24 ` H. Peter Anvin
2008-10-03 22:33 ` Nakajima, Jun
2008-10-03 23:30 ` H. Peter Anvin
2008-10-04 0:27 ` Nakajima, Jun
2008-10-04 0:35 ` H. Peter Anvin
2008-10-07 22:30 ` Nakajima, Jun
2008-10-07 22:37 ` H. Peter Anvin
2008-10-07 23:45 ` Jeremy Fitzhardinge
2008-10-08 1:09 ` H. Peter Anvin
2008-10-07 23:41 ` Jeremy Fitzhardinge
2008-10-07 23:45 ` H. Peter Anvin
2008-10-08 0:40 ` Jeremy Fitzhardinge
2008-10-04 8:53 ` Avi Kivity
2008-10-01 17:47 ` H. Peter Anvin
2008-10-01 18:04 ` Jeremy Fitzhardinge
2008-10-01 18:07 ` H. Peter Anvin
2008-10-01 18:12 ` Jeremy Fitzhardinge
2008-10-01 18:16 ` H. Peter Anvin
2008-10-01 18:36 ` Jeremy Fitzhardinge
2008-10-01 18:43 ` H. Peter Anvin
2008-10-01 19:56 ` Jeremy Fitzhardinge
2008-10-01 20:38 ` Chris Wright
2008-10-01 22:38 ` H. Peter Anvin
2008-10-01 21:01 ` Alok Kataria
2008-10-01 21:08 ` Anthony Liguori
2008-10-01 21:15 ` Chris Wright
2008-10-01 21:31 ` Anthony Liguori
2008-10-01 21:23 ` Alok Kataria
2008-10-01 21:29 ` Anthony Liguori
2008-10-01 21:17 ` Jeremy Fitzhardinge
2008-10-01 21:34 ` Anthony Liguori
2008-10-01 21:43 ` Chris Wright
2008-10-02 11:29 ` Avi Kivity
2008-10-01 23:47 ` Zachary Amsden
2008-10-02 0:39 ` H. Peter Anvin
2008-10-02 0:57 ` H. Peter Anvin
2008-10-02 1:11 ` Zachary Amsden
2008-10-02 1:21 ` H. Peter Anvin [this message]
2008-10-02 0:41 ` Anthony Liguori
[not found] ` <48E3BBC1.2050607__35819.6151479662$1222884502$gmane$org@goop.org>
2008-10-01 20:03 ` Anthony Liguori
2008-10-01 20:08 ` Jeremy Fitzhardinge
[not found] ` <48E3D8A8.604__13396.6479487301$1222891831$gmane$org@goop.org>
2008-10-01 21:03 ` Anthony Liguori
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=48E42234.9000808@zytor.com \
--to=hpa@zytor.com \
--cc=akataria@vmware.com \
--cc=anthony@codemonkey.ws \
--cc=avi@redhat.com \
--cc=dhecht@vmware.com \
--cc=jeremy@goop.org \
--cc=jun.nakajima@intel.com \
--cc=kraxel@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=rusty@rustcorp.com.au \
--cc=virtualization@lists.linux-foundation.org \
--cc=x86@kernel.org \
--cc=zach@vmware.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).