> That approach is not going to fly for a large number of reasons. I understand, I'm fine with that, I've been quite happy with this current behavior so far.
> macppc also has the smu(4) way, and perhaps other ways. I see smu(4) does a similar thing to my diff in smu_attach through TEMP_TRG -if I'm not mistaken. So perhaps as you say, it's a matter of deciding to which machines this should apply. I will seek to find out if other macppc machines with adt(4) are happy with this. Thanks for pointing out that adt is used in a bunch of architectures
