Hi Jean, On Tue, 2011-05-31 at 13:15 -0400, Jean Delvare wrote: > Hi Guenter, > > On Tue, 31 May 2011 07:15:20 -0700, Guenter Roeck wrote: > > Commit a321cedb12904114e2ba5041a3673ca24deb09c9 excludes CPU models 0xe, 0xf, > > and 0x16 from TjMax temperature adjustment, even though several of those CPUs > > are known to have TiMax other than 100 degrees C, and even though the code in > > adjust_tjmax() explicitly handles those CPUs and points to a Web document > > listing several of the affected CPU IDs. > > Good catch. Tested on a Core Duo T2600 (model ID 0xe), your patch gets > rid of the "TjMax is assumed as 100 C!" warning message. > > > Reinstate TjMax adjustment for CPUs with model ID 0xe, 0xf, and 0x16. > > But then why not 0x1a as well? The documentation lists a number of > Nehalem-based processors with TjMax 90 or 105ÂC. It seems that > adjust_tjmax() would figure it out properly, so we should just call it? > Only reason is that I wasn't sure about those. > What about 0x1e (Lynnfield)? It was already supported prior to > Carsten's changes. I don't know when MSR_IA32_TEMPERATURE_TARGET was > introduced, but again I see no reason why adjust_tjmax() wouldn't be > good for Lynnfield - at least as good as blindly assuming TjMax at > 100ÂC. > Same here ... > BTW, I am curious why we read MSR_IA32_TEMPERATURE_TARGET on CPU models, > which we know do _not_ have this MSR? This leads to a warning message > in the logs. It would seem more logical to call adjust_tjmax() directly > if MSR_IA32_TEMPERATURE_TARGET isn't supported. > I think you are right. The warning doesn't really make sense. Maybe I should rearrange the code to just return the value reported by adjust_tjmax() if reading TjMax through MSR_IA32_TEMPERATURE_TARGET fails - what do you think ? Thanks, Guenter _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
