Linus,
	I believe Alan cleared up the only objection to this small patch.  Please 
apply.

Changelog:
	Cleans up a bit of a hack to get compressed (gziped) kernels booting on 
NUMA-Q.  Resulting code is more readable and understandable.

Cheers!

-Matt

Alan Cox wrote:
> On Tue, 2002-10-01 at 22:44, Matthew Dobson wrote:
> 
>>>STANDALONE seems to be a very namespace-polluting choice of define.
>>>MULTIQUAD_STANDALONE, MQ_STANDALONE... anything would be better imo.
>>
>>The #define is most definitely *not* NUMA/Multiquad specific.  In this
>>particular instance, it is guarding Multiquad specific code...  The 
>>STANDALONE option (please clarify if I'm wrong, Alan) is for code that 
>>is compiled along with the kernel, with the kernel headers, etc, but is 
>>not acually part of the kernel proper.
> 
> 
> Indeed
> 
> Its set by the boot loader code that wants to also use inb/outb etc but
> not get the kernel magic wonders of numa-q and other evil abuses of PC
> iomapping