Hi Juerg, On Sat, 27 Jun 2009 11:37:34 -0700, Juerg Haefliger wrote: > > On Thu, 18 Jun 2009 10:09:33 -0700, Kushal Koolwal wrote: > >> Any ideas why I need to give force_start? Is this a chip /BIOS/OS issues? > > > > BIOS issue. It's up to the BIOS to enable monitoring after the chip is > > properly set up. That being said, some of our drivers silently enable > > parts if they were disabled. I'm not so sure why the dme1737 driver > > asks for the user to pass an extra parameter. But I can't really > > complain, maybe I'm the one who asked for it, I just can't remember. > > Juerg? > > Just being conservative. I didn't want to flip any bits without the > user explicitly asking for it. I thought that was the general approach > for the hwmon subsystem. Not really. Looking at other drivers you'll see more than 20 examples where we do start monitoring if needed: adm1026, adm9240, adt7470, adt7473, asb100, dme1737, it87, lm78, lm80, lm85, lm93, smsc47m192, via686a, w83627ehf, w83627hf, w83781d, w83791d, w83792d, w83793, w83l786ng, adm1031 and f71805f, and maybe more. What we do _not_ forcibly enable by default is the I/O area of Super-I/O chips. But if the I/O area is setup and enabled, I have no objection to letting the driver enable monitoring itself if needed. Whether you want to update the dme1737 driver in that direction, is up to you. -- Jean Delvare
