On Wed, 05 Mar 2008 11:31:30 +0100, Stefan Sassenberg wrote: > Jean Delvare wrote: > > Any message in the kernel logs when you modprobe w83627hf? Does the > > modprobe itself fail or succeed? If it fails, what's the error message? > > When I modprobe, a line is added to /var/log/dmesg: > w83627hf: Found W83627THF chip at 0x290 > There's no message on stdout or stderr. So the kernel itself seems to be happy... > (...) > > Please compare /proc/ioports between a system where the w83627hf driver > > works for you and a system where it doesn't. I'd suspect a resource > > conflict. > Using the 2.6.21.5 kernel: > 0295-0296 : pnp 00:09 > 0295-0296 : w83627hf > > Using the 2.6.23.14 kernel: > Before modprobing: > ... > 0295-0296 : pnp 00:09 > ... > > After modprobing: > ... > 0295-0296 : w83627hf > 0295-0296 : pnp 00:09 > 0295-0296 : w83627hf > ... And this looks OK as well. My first guess was wrong, let me suggest something else... I didn't notice at first but you wrote you were using lm-sensors 2.9.1. That's pretty old. In kernel 2.6.22 the w83627hf driver was converted from an i2c-isa driver to a platform driver (together with many other hwmon driver, as we were finally getting rid of i2c-isa.) Support for platform drivers was added to lm-sensors only in 2.10.0. So, no surprise it doesn't work for you. For a kernel >= 2.6.22, I would even recommend lm-sensors >= 2.10.3. -- Jean Delvare