From owner-acpi-jp@jp.FreeBSD.org Wed Mar 31 05:14:21 2004
Received: (from daemon@localhost)
	by castle.jp.FreeBSD.org (8.11.6p2+3.4W/8.11.3) id i2UKELJ23504;
	Wed, 31 Mar 2004 05:14:21 +0900 (JST)
	(envelope-from owner-acpi-jp@jp.FreeBSD.org)
Received: from root.org (root.org [67.118.192.226])
	by castle.jp.FreeBSD.org (8.11.6p2+3.4W/8.11.3) with SMTP/inet id i2UKEKd23499
	for <acpi-jp@jp.FreeBSD.org>; Wed, 31 Mar 2004 05:14:20 +0900 (JST)
	(envelope-from nate@root.org)
Received: (qmail 82337 invoked by uid 1000); 30 Mar 2004 20:14:07 -0000
From: Nate Lawson <nate@root.org>
To: Bruce Cran <bruce@cran.org.uk>
cc: acpi-jp@jp.FreeBSD.org
In-Reply-To: <20040323230947.GA778@box1.cran>
Message-ID: <20040330120837.J81785@root.org>
References: <20040316143411.GA4468@portege.clkao.org> <20040316100826.S95770@root.org>
 <20040323230947.GA778@box1.cran>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: acpi-jp@jp.FreeBSD.org
Precedence: list
Date: Tue, 30 Mar 2004 12:14:07 -0800
X-Sequence: acpi-jp 3161
Subject: [acpi-jp 3161] Re: acpi FADT and throttle problems
Sender: owner-acpi-jp@jp.FreeBSD.org
X-Originator: nate@root.org
X-Distribute: distribute version 2.1 (Alpha) patchlevel 24e+040307

On Tue, 23 Mar 2004, Bruce Cran wrote:
> On Tue, Mar 16, 2004 at 10:10:36AM -0800, Nate Lawson wrote:
> > On Tue, 16 Mar 2004, Chia-liang Kao wrote:
> > > I have some problems with throttling my Toshiba 3490.
> > >
> > > acpidump -t shows me: DUTY_OFFSET=1, DUTY_WIDTH=0 in FADT, so i don't
> > > have the hw.acpi.throttle MIB at all. while my dmesg says:
> > > CPU: Intel Pentium III (500.02-MHz 686-class CPU)
> > >   Origin = "GenuineIntel"  Id = 0x686  Stepping = 6
> > >
> > > So it looks like it should be able to trottle. Is it possible to
> > > override the FADT or fix it somehow?
> >
> > Throttling is related to the chipset as well as CPU.  The system designer
> > specified 0 for the width for a reason.  You should try the acpi_toshiba
> > driver.  It provides throttling sysctls of its own.  If you absolutely
> > have to try ACPI throttling, you can just remove the check for cpu_width
> > == 0 in acpi_cpu.c.
>
> I'm seeing a regression in the new ACPI system on my Dell Inpiron 8500.  It
> has a 2GHz P4-M, which has the standard 8 steps from 100-8.5%.   It used
> to throttle, but now it stays at full speed whatever I set
> hw.acpi.cpu.throttle_state to.   I can confirm this by running a program which
> basically does:

I'm not sure why setting throttle_state doesn't work for you.  Perhaps
others that have throttling support can make sure manual throttling still
works for them so we can isolate this?

The "performance states changed" has nothing to do with throttling.  It is
for Px support which will be in the upcoming cpufreq driver.  I need to
fix cpu newbus attachments before I can finish that driver though.

You need to run devd to get automatic power profile changes now.  See
/etc/defaults/rc.conf for the values that /etc/rc.d/power_profile uses.
The default for throttling is not to change on AC line transitions since
it has a high performance cost.

-Nate
