[RFC] Drop Non-smp PowerPC Kernel Flavor

[Jeremy Kerr]

Hi Leann,

> Jeremy Kerr did some testing a while back and noted that the PowerPC smp
> flavor should run on non-smp hardware, but did acknowledge there could
> be a slight performance degradation.  Jeremy, do you still have the data
> available?


I haven't done any testing myself, so don't have any quantitative data 
on the performance impact. This was just information from checking out 
the kernel source and having a chat with Ben H, the powerpc kernel 
maintainer.

The issue is that some SMP-specific components will be present on UP 
builds, mainly spinlocks. Of course, there'll be no lock contention, but 
the instructions are still present. ISTR the estimate was around 1-2% 
performance difference.

Cheers,


Jeremy