rhn_register claims cpu model 0, which certainly looks odd. Basically, we should dig around in the kernel cpuinfo code to see if it says anything different for McKinley, if so, have the parser key off whatever differences are there.
discussed this with bill via email. The model number of the cpu on our ia64's is indeed "0". so..?
Finally looked at the kernel code. The model tag is pulled directly from the processor cpuinfo. No information on if or how it will change. However, the 'family' tag (currently set to 'Itanium') will say 'McKinley' if a McKinley is detected. So that might be something better to key off of.
ok. I'll populate the model info with the value pulled out of the "family" info.