[PATCH] SPARSEMEM EXTREME

A new option for SPARSEMEM is ARCH_SPARSEMEM_EXTREME.  Architecture
platforms with a very sparse physical address space would likely want to
select this option.  For those architecture platforms that don't select the
option, the code generated is equivalent to SPARSEMEM currently in -mm.
I'll be posting a patch on ia64 ml which uses this new SPARSEMEM feature.

ARCH_SPARSEMEM_EXTREME makes mem_section a one dimensional array of
pointers to mem_sections.  This two level layout scheme is able to achieve
smaller memory requirements for SPARSEMEM with the tradeoff of an
additional shift and load when fetching the memory section.  The current
SPARSEMEM -mm implementation is a one dimensional array of mem_sections
which is the default SPARSEMEM configuration.  The patch attempts isolates
the implementation details of the physical layout of the sparsemem section
array.

ARCH_SPARSEMEM_EXTREME depends on 64BIT and is by default boolean false.

I've boot tested under aim load ia64 configured for ARCH_SPARSEMEM_EXTREME.
 I've also boot tested a 4 way Opteron machine with !ARCH_SPARSEMEM_EXTREME
and tested with aim.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Bob Picco <bob.picco@hp.com>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/mm/Kconfig b/mm/Kconfig
index cd37993..fc644c5 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -89,3 +89,12 @@
 config HAVE_MEMORY_PRESENT
 	def_bool y
 	depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM
+
+#
+# Architectecture platforms which require a two level mem_section in SPARSEMEM
+# must select this option. This is usually for architecture platforms with
+# an extremely sparse physical address space.
+#
+config ARCH_SPARSEMEM_EXTREME
+	def_bool n
+	depends on SPARSEMEM && 64BIT
diff --git a/mm/sparse.c b/mm/sparse.c
index b54e304..b2b456b 100644
--- a/mm/sparse.c
+++ b/mm/sparse.c
@@ -13,7 +13,26 @@
  *
  * 1) mem_section	- memory sections, mem_map's for valid memory
  */
-struct mem_section mem_section[NR_MEM_SECTIONS];
+#ifdef CONFIG_ARCH_SPARSEMEM_EXTREME
+struct mem_section *mem_section[NR_SECTION_ROOTS]
+	____cacheline_maxaligned_in_smp;
+
+static void sparse_index_init(unsigned long section, int nid)
+{
+	unsigned long root = SECTION_TO_ROOT(section);
+
+	if (mem_section[root])
+		return;
+	mem_section[root] = alloc_bootmem_node(NODE_DATA(nid), PAGE_SIZE);
+	if (mem_section[root])
+		memset(mem_section[root], 0, PAGE_SIZE);
+	else
+		panic("memory_present: NO MEMORY\n");
+}
+#else
+struct mem_section mem_section[NR_MEM_SECTIONS]
+	____cacheline_maxaligned_in_smp;
+#endif
 EXPORT_SYMBOL(mem_section);
 
 /* Record a memory area against a node. */
@@ -24,8 +43,13 @@
 	start &= PAGE_SECTION_MASK;
 	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
 		unsigned long section = pfn_to_section_nr(pfn);
-		if (!mem_section[section].section_mem_map)
-			mem_section[section].section_mem_map = SECTION_MARKED_PRESENT;
+		struct mem_section *ms;
+
+		sparse_index_init(section, nid);
+
+		ms = __nr_to_section(section);
+		if (!ms->section_mem_map)
+			ms->section_mem_map = SECTION_MARKED_PRESENT;
 	}
 }
 
@@ -85,6 +109,7 @@
 {
 	struct page *map;
 	int nid = early_pfn_to_nid(section_nr_to_pfn(pnum));
+	struct mem_section *ms = __nr_to_section(pnum);
 
 	map = alloc_remap(nid, sizeof(struct page) * PAGES_PER_SECTION);
 	if (map)
@@ -96,7 +121,7 @@
 		return map;
 
 	printk(KERN_WARNING "%s: allocation failed\n", __FUNCTION__);
-	mem_section[pnum].section_mem_map = 0;
+	ms->section_mem_map = 0;
 	return NULL;
 }
 
@@ -114,8 +139,9 @@
 			continue;
 
 		map = sparse_early_mem_map_alloc(pnum);
-		if (map)
-			sparse_init_one_section(&mem_section[pnum], pnum, map);
+		if (!map)
+			continue;
+		sparse_init_one_section(__nr_to_section(pnum), pnum, map);
 	}
 }