x86: Use HAVE_MEMBLOCK_NODE_MAP
authorTejun Heo <tj@kernel.org>
Thu, 14 Jul 2011 09:44:23 +0000 (11:44 +0200)
committerH. Peter Anvin <hpa@linux.intel.com>
Thu, 14 Jul 2011 18:47:43 +0000 (11:47 -0700)
commit0608f70c78a384c2f225f2de226ca057a196f108
treea5ffb473cc55ca8bb538b9869cb61d2982d99f5e
parent7c0caeb866b0f648d91bb75b8bc6f86af95bb033
x86: Use HAVE_MEMBLOCK_NODE_MAP

From 5732e1247898d67cbf837585150fe9f68974671d Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Thu, 14 Jul 2011 11:22:16 +0200

Convert x86 to HAVE_MEMBLOCK_NODE_MAP.  The only difference in memory
handling is that allocations can't no longer cross node boundaries
whether they're node affine or not, which shouldn't matter at all.

This conversion will enable further simplification of boot memory
handling.

-v2: Fix build failure on !NUMA configurations discovered by hpa.

Signed-off-by: Tejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/r/20110714094423.GG3455@htj.dyndns.org
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
arch/x86/Kconfig
arch/x86/include/asm/memblock.h
arch/x86/mm/init_32.c
arch/x86/mm/init_64.c
arch/x86/mm/memblock.c
arch/x86/mm/numa.c