Hi, > First of all, know that neither rivatv 0.8.5 nor CVS will work properly with > Linux 2.6.7 as far as hardware monitoring is concerned. The author missed a > define names change, and as a consequence, hardware monitoring chips will not > recognize the nVidia busses as usable. I will send a patch to the author. I applied your patch to rivatv and recompiled it, but it still doesn't work. > Second, not only no LM99 chip is detected by sensors-detect, but the script > doesn?t even see a chip at a possible address (0x4C or 0x4D). This can mean > either of two things: > 1* You simply don?t have a LM99 on that board. I guess that there are several > different brands and models of GeForce FX cards out there, not all may have > hardware monitoring features. Are you sure yours do? http://www.hothardware.com/viewarticle.cfm?articleid=421&catid=2 This page says "Advanced thermal management and thermal monitoring" so I suppose it has one. > 2* The bus errors you are experiencing (sendbytes: error - bailout.) are > preventing the chip from being detected (simply as being there, not even as > being a LM99). > > Nothing wrong with the client on bus 2 that cannot be probed though. It simply > means that an i2c client driver (most likely saa7108e) is already loaded for > this device. > > At any rate, the problem here is with the i2c bus, not the LM99 (if it exists). > You should contact the maintainer of the rivatv module (as an i2c bus driver), > maybe he?ll have ideas about how you can investigate the problem. > > You could also try sensors-detect as found in lm_sensors 2.8.6. Detection method > slightly changed since. New method is believed to be better, of course, and was > successfully tested, but your system might be special for some reason. > > You can contact us on the mailing-list if you have further questions or want to > tell us the end of the story.