acpi_thermal on nforce 4 board

[Kevin Oberman]

> From: Pav Lucistnik <pav at FreeBSD.org>
> Date: Wed, 03 May 2006 16:53:31 +0200
> Sender: owner-freebsd-acpi at freebsd.org
> 
> Alex Zbyslaw pí¨e v st 03. 05. 2006 v 15:50 +0100:
> > Pav Lucistnik wrote:
> > 
> > >Alex Zbyslaw pí¹e v st 03. 05. 2006 v 15:04 +0100:
> > >
> > >  
> > >
> > >>>BTW I don't see why mbmon failed.  Maybe you should
> > >>>try something like that:
> > >>>
> > >>>mbmon -P winbond
> > >>> 
> > >>>
> > >>>      
> > >>>
> > >>Also for some winbond chips healthd can work where mbmon doesn't.
> > >>    
> > >>
> > >
> > >Riight, healthd works and reports sane values. Funny, that it reports
> > >different temperatures on the cores of my dual core CPU.
> > >  
> > >
> > I would think that it was reporting the temps for your "second" 
> > (non-existent) CPU, not per core, but I could be wrong.  The 2nd CPU 
> > temps and Vcore are just noise if you don't have 2 CPUs.
> 
> Could be, but CPU0 is 38C and CPU1 is 42C so it's not completely
> unplausible...

mbmon should be reporting three temperatures for the MSI mobo. The first
is the "system" temperature. (I have no idea exactly where this is
measured, but I suspect the northbridge), the second is CPU0 and the
third is CPU1. By CPU0, I mean the package, not the core.

I typically see a "temperature" way below room temp for CPU1, which I
don't have. Since modern CPUs all have internal temperature sensors and
they connect to pins on the package. If the chip is not there, the
circuit on the mobo that "translates" this analog value tends to drift
around with the temperature of its components, but not in an obviously
meaningful way.

Right now I show a system temp of 30C, a CPU temperature of 34C and a
phantom of 15C for the missing CPU. (The room it's in is at about 22C.)

The second Vcore is similarly silly.

Now, why does my mbmon work out of the box and you need a small patch?
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634