Hi. On Sun, 2005-12-18 at 17:38 +0100, Jean Delvare wrote: > I just checked and doc/developers/proc doesn't say that, which is good > because that's not true. Hmm, indeed, seems that I mixed up something here. Maybe I shouldn't report things like that after working for 14+ hours. :( Sorry for the confusion that might have caused. > As the LM77 reports temperature values with a resolution of 0.5 degree > Celsius, your driver should declare a magnitude (number of digits after > the decimal point) of 1, and should report the temperature values with > a LSB representing 0.1 degree Celsius. After having a closer look how the LM87 does these things (and after enabling that sensor on a second box I have at hands for testing purposes) I got that working now. With this and your comments in mind it ended up to be an issue of "RTFM" - I somehow missed to check the meaning of SENSORS_PROC_REAL_INFO, which would have cleared up a lot. > lm75 driver for an example. The lm77 case should be very similar. If > the driver currently reports 4200.0 for 42 degrees, it means that the > magnitude is properly set to 1, but the temperature value has been > multiplied by 500 (from the register value) instead of 5. Yeah, that was exactly the kind of problem I was facing. Thanks a lot for your comments on that and for the patience to explain it rather than giving the "RTFM" I'd deserve :) Bye, Mike