Hi Grant, > This trivial patch changes vid reporting for adm1025 to the > standard cpu0_vid filename. It leaves the older in1_ref file > in, marked as deprecated for later removal so we do not pull > the rug out from anybody. I like it much. Especially keeping the old file in place is very smart. > My query is do we need to give notification of feature change > or removal as some other kernel sections are doing? Trying to > tread very gently... I don't think we have to. Here you are simply fixing an interface which was not following the standard. You are not changing a standard. > Compile tested, patch tested, but no run test :( The patch is quite simple, looks all OK to me. Could you please provide a patch to libsensors, such that cpu0_vid will be accessed instead of in1_ref for the adm1025/ne1619 chips? We might not apply it immediately so as to not break compatibility, but at least let's have it ready. As a side note, I noticed that the adm9240 support in libsensors uses in1_ref as well. Since your driver obviously won't use it, but cpu0_vid instead, this should be fixed as well (and without a delay). Thanks, -- Jean Delvare
