Johan Danielsson wrote:
> > I just ported the it(4) driver [from OpenBSD].
> 
Of course, form OpenBSD.
> Looks reasonable, but it should probably be called iteenv or similar,
> and the manpage is missing.
Manpage is now ready. 
> 
> Why does it use 103/20 and 83/20 when 36/7 and 29/7 would be more
> accurate (the datasheet says 4 1/7 for resistor quotient for negative
> twelve)? Not that it matters much.
> 
These value is from oryginaly driver. You are right. I replace
it for value (and default sensors names) from ITE8705 datasheet. 

New version is at ftp://ftp.atom.eu.org/pub/it/it.tgz .

> For reference, the half-finished driver I wrote is appended.
Thanks for this.

Why envsys api not implemented ENVSYS_SNEGVOLTS_DC? If this unit exists
it very easy to replace envsys_basic_info.rfact and calculate right
value for negative volts. Actually you should resort to the methods used
by yourself (i thinking about tables iteenv_softc.sc_negative_input) and
those methods shouldnt be changed by ENVSYS_STREINFO.

And why envsys API must be replaced by sysctl? (Someone working on this?)

-- 
Bartosz Kuzma.