Hi, > The above is not necessarily correct -- many of the chips supported by > the lm(4) driver, on NetBSD, OpenBSD and DragonFly BSD, certainly do, > in fact, implement a whole bunch of fan-controlling features; see > http://sensors.cnst.su/fanctl/ for a sample BSD implementation of the > fan controlling logic.
Ah - I looked at the LM manuals, as I had the references handy, and assumed the others were similar. They are almost different enough to have separate drivers. However, I think that it makes sense to add the changes in the patch to our driver. > Winbond W83627HF is supported by both http://mdoc.su/o,n,d/wbsio.4 and > http://mdoc.su/n,o,d/lm.4, and is one of the chips supported by the > fanctl patch above, too. Looks like wbsio handles the attachment glue (rather than lm at iic). > However, the above manual from Tyan doesn't appear to specify how > exactly are the fans actually wired up (e.g., to which pins on which > chip). That seems fairly standard. If the firmware can display the sensor values, it's possible to do some correlation, but otherwise I haven't found a good way, especially with the range of temperatures and voltage sources that can be measured. As an example where the firmware can show the values, I can translate sensors on a Sun fire V240: [adm1026hm0] F0.RS: 5720 RPM ... MB.P0.T_CORE: 47.000 degC ... MB.V_GBE_+2V5: 2.508 V ... [lmtemp0] MB.T_ENC: 11.000 degC which is nice. Regards, J PS. Approximate translations: F0.RS fan 0 rotational speed MB.P0.T_CORE main board processor 0 temperature (core) MB.V_GBE_+2V5 main board gigabit ethernet +2.5V MB.T_ENC main board temperature (enclosure) -- My other computer runs NetBSD too - http://www.netbsd.org/