On Tue, Jul 11, 2006 at 11:14:48PM +0000, Eduardo Horvath wrote:
> When I looked at the kernel logs it seemed that they had different 
> interrupt vector numbers for the same device.

I'm not sure what you are refering to, I checked the interrupt routing
relevant OF properties on another machine some time ago when this came
first up and they did not change between both OF versions, and the 
calculation the kernel does looked correct to me. In the logs Julian
posted it seems both have:

gem0: interrupting at ivec 3006
aceride0: using ivec 180c for native-PCI interrupt

Are you looking at some other device or the PCI headers?

Martin