Hi again Paul, > > vid: -3.425 V (VRM Version 9.1) > > Wow. Not good. What value does /sys/bus/i2c/devices/0-002e/cpu0_vid > contain? That and the missing +12V reading make me suspicious about the fact that you actually have an ADM1027 chip. I suspect that you actually have an ADT7463 chip, misdetected (which is plausible when I look at the driver code in 2.6.9). Do you confirm this? If you do not know which chip you do have, please provide the output of "i2cdump 0 0x2e b", this should answer my question. According to the datasheet, the ADT7463 can use VRM 10 (i.e. 6-bit VID values), while other chips in the family couldn't (5-bit VID values). This is done by connecting VID5 to the +12V input, and changing a configuration bit in the chip. This would explain the lack of +12V reading. This also explains the bogus VID reading, as the driver code didn't expect the new configuration bits in the VID register. We'll need to update the lm85 driver to support that case. Thanks, -- Jean Delvare
