mm/memory.c:72:2: warning: #warning Unfortunate NUMA config, growing page-frame f

Hi Peter,

FYI, there are new compile warnings show up in

tree:   git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git master
head:   2ef756a2e79149dc84c88f7ccc42ec38cbb7722e
commit: 835e3cf29f4eeba6c5c8bcc188b2138029157938 [30/35] mm: Fold page::nid_last into page::flags when possible
config: i386-allmodconfig

All warnings:

mm/memory.c:72:2: warning: #warning Unfortunate NUMA config, growing page-frame for last_nid. [-Wcpp]

vim +72 mm/memory.c

42b77728 (Jan Beulich     2008-07-23  69) #include "internal.h"
42b77728 (Jan Beulich     2008-07-23  70) 
835e3cf2 (Peter Zijlstra  2012-08-13  71) #ifdef LAST_NID_NOT_IN_PAGE_FLAGS
835e3cf2 (Peter Zijlstra  2012-08-13 @72) #warning Unfortunate NUMA config, growing page-frame for last_nid.
835e3cf2 (Peter Zijlstra  2012-08-13  73) #endif
835e3cf2 (Peter Zijlstra  2012-08-13  74) 
d41dee36 (Andy Whitcroft  2005-06-23  75) #ifndef CONFIG_NEED_MULTIPLE_NODES
^1da177e (Linus Torvalds  2005-04-16  76) /* use the per-pgdat data instead for discontigmem - mbligh */
^1da177e (Linus Torvalds  2005-04-16  77) unsigned long max_mapnr;
^1da177e (Linus Torvalds  2005-04-16  78) struct page *mem_map;
^1da177e (Linus Torvalds  2005-04-16  79) 
^1da177e (Linus Torvalds  2005-04-16  80) EXPORT_SYMBOL(max_mapnr);

---
0-DAY kernel build testing backend Open Source Technology Centre
Fengguang Wu, Yuanhan Liu Intel Corporation

Yeah, 32bit NUMA config.. nobody sane should be doing or using that.

The other option was to hard-disable 32bit support for SCHED_NUMA, which I wouldn't mind either.. who gives about 32 bits computing anyway ;-)