Hi Larry, On Wed, 4 Sep 2013 20:17:41 -0500, Larry Lade wrote: > Thank you Jean and Guenter for taking the time to reply. > > Since it was still new, I decided to work with the retailer to replace the > device. I think it was the best move, indeed. Not only for you, but it will also show the hardware vendor that they probably should go for a more robust design. > I have an suspicion there's a EEPROM in this device that is shipping > without write-protection set, and the sensors-detect probe rakes over it at > random before eventually eventually setting the chip into read-only mode. This is a possibility, yes. But there are many other possibilities and we are pretty clueless at the moment :( > If that's the case, then it's probably a lost cause. But EEPROM programming > is all very much over my head! I'm happy to not solve the mystery if I > don't have to. I understand. > Perhaps the thing to do is file a bug with Ubuntu to push out a more > up-to-date version of lm-sensors to their long-term support release. Yes, this should be fixed by distributions. I have added a note about that on the main page of lm-sensors.org: http://www.lm-sensors.org/wiki Distributions don't have to upgrade to lm-sensors 3.3.3+, but they should at least backport the two critical fixes that were added to 3.3.3 to avoid further hardware breakage: http://www.lm-sensors.org/changeset/6040 http://www.lm-sensors.org/changeset/6084 So if you can get Ubuntu to backport these to affected products still under maintenance, that would be great. I'll do the same for openSUSE. I also have plans to make sensors-detect even safer by default, I'll discuss these on this list in a separate thread. -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
