Thanks for the feedback, here is a simplifed pair of patches.I also like the x2apic optimization, since it addresses additional delays. -- though we may be able to optimize those in the non-x2apic case too.thanks,Len Brown, Intel Open Source Technology Center