Re: GSoC2007: cnst-sensors.2007-08-20.patch

On 11/09/2007 17:47, Harald Schmalzbauer wrote:

> Am Sonntag, 26. August 2007 01:44:47 schrieb Constantine A. Murenin:
> [schnip]
> 
>>P.S. BTW, please do report back how the tests went. Don't forget to
>>include the information on which exact chip you tested (for example, you
>>might have tried this code on Winbond W83627DHG and Core 2 Duo E4300).
> 
> 
> Hello,
> 
> I had only tested some newer hardware with core2 CPU (E6750) and coretemp 
> doesn't report correct values for that CPU (25°C, this is at least 10° 

Thanks, but this must be a bug in the default coretemp(4), not really 
related to the framework, so you might want to report to rpaulo_at_, cc'ing 
des_at_ and cnst_at_.

> difference) and the boards doesn't neccessarily have a HW monitor chip, so I 
> haven't reported anything yet.
> 
> But,
> I'm about to test some VIA C7 boards and accidentaly found that if I disable 
> ACPI (hint.acpi.0.disabled=1), dmesg shows a lm sensor:
> lm0: <Winbond W83697HF Hardware Monitor> at port 0x290-0x297 on isa0
> 
> Also systat -sensors shows adaequat values corresponding to what the BIOS HW 
> monitor reports.
> 
> But as soon as I boot default with ACPI enabled, lm0 vanishes.

It's probably something to do with ACPI (duh!), I'm not certain what it 
is or why it happens.

I have the following chip:

lm0: <Winbond W83627DHG Hardware Monitor> at port 0x290-0x297 on isa0

And it works fine with acpi turned on, so I tend to think that you're 
experiencing a bug that is somewhere else in the tree.

> Thanks for any hints.
> 
> -Harry
> 
> P.S. The patch doesn't apply cleanly for the sysctl man page, haven't had time 
> to investigate yet.

That's a good thing, because it means that some fixes were already 
integrated into CVS. ;)

P.S. I'll have a new patch soon, which will improve sysctl(3) support 
and compatibility with OpenBSD, will feature sysctl(8) support, and will 
include a new driver for ITE Tech Super I/O Hardware Monitors, too. ;)

C.