Re: port-i386/18205 (lm driver temperature miscalculation)

[pgoyette%NetBSD.org@localhost]

Synopsis: lm driver temperature miscalculation

Responsible-Changed-From-To: port-i386-maintainer->pgoyette
Responsible-Changed-By: pgoyette%NetBSD.org@localhost
Responsible-Changed-When: Fri, 19 Feb 2010 03:06:00 +0000
Responsible-Changed-Why:
I picked it up.


State-Changed-From-To: open->closed
State-Changed-By: pgoyette%NetBSD.org@localhost
State-Changed-When: Fri, 19 Feb 2010 03:06:00 +0000
State-Changed-Why:
Issue is most like fixed by recent changes, but originator is not able to 
verify.  He has given OK to close the PR.

Date: Thu, 18 Feb 2010 20:15:38 -0600 (CST)
From: MLH <mlh%goathill.org@localhost>
To: paul%whooppee.com@localhost
Subject: Re: PR kern/18205

> 
> Would it be OK then if we close the PR?

Sure.

> On Thu, 18 Feb 2010, MLH wrote:
> 
> >>
> >> This is an old PR, but I think we've recently addressed it!
> >>
> >> Many of the Winbond chips can actually use different sensor types for
> >> the temp sensors.  You can now select between Pentium-II diode,
> >> thermistor, or 2N3904 BiPolar sensors using the config flags.  The
lm(4)
> >> man page was updated to indicate the appropriate values.
> >>
> >> Can you check and see if this fixes your temperature problem?
> >
> > That was for a compute farm with about 500 cpus that had the problem
> > (of several thousand) that I suspect have been replaced.  I haven't
> > had access to them for several years.  Sorry.