[GitHub/mt8127/android_kernel_alcatel_ttab.git] / mm / sparse.c

/*
 * sparse memory mappings.
 */
#include <linux/mm.h>
#include <linux/mmzone.h>
#include <linux/bootmem.h>
#include <linux/highmem.h>
#include <linux/module.h>
#include <linux/spinlock.h>
#include <linux/vmalloc.h>
#include <asm/dma.h>
#include <asm/pgalloc.h>
#include <asm/pgtable.h>

/*
 * Permanent SPARSEMEM data:
 *
 * 1) mem_section	- memory sections, mem_map's for valid memory
 */
#ifdef CONFIG_SPARSEMEM_EXTREME
struct mem_section *mem_section[NR_SECTION_ROOTS]
	____cacheline_internodealigned_in_smp;
#else
struct mem_section mem_section[NR_SECTION_ROOTS][SECTIONS_PER_ROOT]
	____cacheline_internodealigned_in_smp;
#endif
EXPORT_SYMBOL(mem_section);

#ifdef NODE_NOT_IN_PAGE_FLAGS
/*
 * If we did not store the node number in the page then we have to
 * do a lookup in the section_to_node_table in order to find which
 * node the page belongs to.
 */
#if MAX_NUMNODES <= 256
static u8 section_to_node_table[NR_MEM_SECTIONS] __cacheline_aligned;
#else
static u16 section_to_node_table[NR_MEM_SECTIONS] __cacheline_aligned;
#endif

int page_to_nid(struct page *page)
{
	return section_to_node_table[page_to_section(page)];
}
EXPORT_SYMBOL(page_to_nid);

static void set_section_nid(unsigned long section_nr, int nid)
{
	section_to_node_table[section_nr] = nid;
}
#else /* !NODE_NOT_IN_PAGE_FLAGS */
static inline void set_section_nid(unsigned long section_nr, int nid)
{
}
#endif

#ifdef CONFIG_SPARSEMEM_EXTREME
static struct mem_section noinline __init_refok *sparse_index_alloc(int nid)
{
	struct mem_section *section = NULL;
	unsigned long array_size = SECTIONS_PER_ROOT *
				   sizeof(struct mem_section);

	if (slab_is_available())
		section = kmalloc_node(array_size, GFP_KERNEL, nid);
	else
		section = alloc_bootmem_node(NODE_DATA(nid), array_size);

	if (section)
		memset(section, 0, array_size);

	return section;
}

static int __meminit sparse_index_init(unsigned long section_nr, int nid)
{
	static DEFINE_SPINLOCK(index_init_lock);
	unsigned long root = SECTION_NR_TO_ROOT(section_nr);
	struct mem_section *section;
	int ret = 0;

	if (mem_section[root])
		return -EEXIST;

	section = sparse_index_alloc(nid);
	if (!section)
		return -ENOMEM;
	/*
	 * This lock keeps two different sections from
	 * reallocating for the same index
	 */
	spin_lock(&index_init_lock);

	if (mem_section[root]) {
		ret = -EEXIST;
		goto out;
	}

	mem_section[root] = section;
out:
	spin_unlock(&index_init_lock);
	return ret;
}
#else /* !SPARSEMEM_EXTREME */
static inline int sparse_index_init(unsigned long section_nr, int nid)
{
	return 0;
}
#endif

/*
 * Although written for the SPARSEMEM_EXTREME case, this happens
 * to also work for the flat array case because
 * NR_SECTION_ROOTS==NR_MEM_SECTIONS.
 */
int __section_nr(struct mem_section* ms)
{
	unsigned long root_nr;
	struct mem_section* root;

	for (root_nr = 0; root_nr < NR_SECTION_ROOTS; root_nr++) {
		root = __nr_to_section(root_nr * SECTIONS_PER_ROOT);
		if (!root)
			continue;

		if ((ms >= root) && (ms < (root + SECTIONS_PER_ROOT)))
		     break;
	}

	return (root_nr * SECTIONS_PER_ROOT) + (ms - root);
}

/*
 * During early boot, before section_mem_map is used for an actual
 * mem_map, we use section_mem_map to store the section's NUMA
 * node.  This keeps us from having to use another data structure.  The
 * node information is cleared just before we store the real mem_map.
 */
static inline unsigned long sparse_encode_early_nid(int nid)
{
	return (nid << SECTION_NID_SHIFT);
}

static inline int sparse_early_nid(struct mem_section *section)
{
	return (section->section_mem_map >> SECTION_NID_SHIFT);
}

/* Record a memory area against a node. */
void __init memory_present(int nid, unsigned long start, unsigned long end)
{
	unsigned long max_arch_pfn = 1UL << (MAX_PHYSMEM_BITS-PAGE_SHIFT);
	unsigned long pfn;

	/*
	 * Sanity checks - do not allow an architecture to pass
	 * in larger pfns than the maximum scope of sparsemem:
	 */
	if (start >= max_arch_pfn)
		return;
	if (end >= max_arch_pfn)
		end = max_arch_pfn;

	start &= PAGE_SECTION_MASK;
	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
		unsigned long section = pfn_to_section_nr(pfn);
		struct mem_section *ms;

		sparse_index_init(section, nid);
		set_section_nid(section, nid);

		ms = __nr_to_section(section);
		if (!ms->section_mem_map)
			ms->section_mem_map = sparse_encode_early_nid(nid) |
							SECTION_MARKED_PRESENT;
	}
}

/*
 * Only used by the i386 NUMA architecures, but relatively
 * generic code.
 */
unsigned long __init node_memmap_size_bytes(int nid, unsigned long start_pfn,
						     unsigned long end_pfn)
{
	unsigned long pfn;
	unsigned long nr_pages = 0;

	for (pfn = start_pfn; pfn < end_pfn; pfn += PAGES_PER_SECTION) {
		if (nid != early_pfn_to_nid(pfn))
			continue;

		if (pfn_present(pfn))
			nr_pages += PAGES_PER_SECTION;
	}

	return nr_pages * sizeof(struct page);
}

/*
 * Subtle, we encode the real pfn into the mem_map such that
 * the identity pfn - section_mem_map will return the actual
 * physical page frame number.
 */
static unsigned long sparse_encode_mem_map(struct page *mem_map, unsigned long pnum)
{
	return (unsigned long)(mem_map - (section_nr_to_pfn(pnum)));
}

/*
 * Decode mem_map from the coded memmap
 */
struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum)
{
	/* mask off the extra low bits of information */
	coded_mem_map &= SECTION_MAP_MASK;
	return ((struct page *)coded_mem_map) + section_nr_to_pfn(pnum);
}

static int __meminit sparse_init_one_section(struct mem_section *ms,
		unsigned long pnum, struct page *mem_map,
		unsigned long *pageblock_bitmap)
{
	if (!present_section(ms))
		return -EINVAL;

	ms->section_mem_map &= ~SECTION_MAP_MASK;
	ms->section_mem_map |= sparse_encode_mem_map(mem_map, pnum) |
							SECTION_HAS_MEM_MAP;
 	ms->pageblock_flags = pageblock_bitmap;

	return 1;
}

unsigned long usemap_size(void)
{
	unsigned long size_bytes;
	size_bytes = roundup(SECTION_BLOCKFLAGS_BITS, 8) / 8;
	size_bytes = roundup(size_bytes, sizeof(unsigned long));
	return size_bytes;
}

#ifdef CONFIG_MEMORY_HOTPLUG
static unsigned long *__kmalloc_section_usemap(void)
{
	return kmalloc(usemap_size(), GFP_KERNEL);
}
#endif /* CONFIG_MEMORY_HOTPLUG */

static unsigned long *__init sparse_early_usemap_alloc(unsigned long pnum)
{
	unsigned long *usemap, section_nr;
	struct mem_section *ms = __nr_to_section(pnum);
	int nid = sparse_early_nid(ms);
	struct pglist_data *pgdat = NODE_DATA(nid);

	/*
	 * Usemap's page can't be freed until freeing other sections
	 * which use it. And, Pgdat has same feature.
	 * If section A has pgdat and section B has usemap for other
	 * sections (includes section A), both sections can't be removed,
	 * because there is the dependency each other.
	 * To solve above issue, this collects all usemap on the same section
	 * which has pgdat.
	 */
	section_nr = pfn_to_section_nr(__pa(pgdat) >> PAGE_SHIFT);
	usemap = alloc_bootmem_section(usemap_size(), section_nr);
	if (usemap)
		return usemap;

	/* Stupid: suppress gcc warning for SPARSEMEM && !NUMA */
	nid = 0;

	printk(KERN_WARNING "%s: allocation failed\n", __FUNCTION__);
	return NULL;
}

#ifndef CONFIG_SPARSEMEM_VMEMMAP
struct page __init *sparse_mem_map_populate(unsigned long pnum, int nid)
{
	struct page *map;

	map = alloc_remap(nid, sizeof(struct page) * PAGES_PER_SECTION);
	if (map)
		return map;

	map = alloc_bootmem_pages_node(NODE_DATA(nid),
		       PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION));
	return map;
}
#endif /* !CONFIG_SPARSEMEM_VMEMMAP */

struct page __init *sparse_early_mem_map_alloc(unsigned long pnum)
{
	struct page *map;
	struct mem_section *ms = __nr_to_section(pnum);
	int nid = sparse_early_nid(ms);

	map = sparse_mem_map_populate(pnum, nid);
	if (map)
		return map;

	printk(KERN_ERR "%s: sparsemem memory map backing failed "
			"some memory will not be available.\n", __FUNCTION__);
	ms->section_mem_map = 0;
	return NULL;
}

void __attribute__((weak)) __meminit vmemmap_populate_print_last(void)
{
}
/*
 * Allocate the accumulated non-linear sections, allocate a mem_map
 * for each and record the physical to section mapping.
 */
void __init sparse_init(void)
{
	unsigned long pnum;
	struct page *map;
	unsigned long *usemap;
	unsigned long **usemap_map;
	int size;

	/*
	 * map is using big page (aka 2M in x86 64 bit)
	 * usemap is less one page (aka 24 bytes)
	 * so alloc 2M (with 2M align) and 24 bytes in turn will
	 * make next 2M slip to one more 2M later.
	 * then in big system, the memory will have a lot of holes...
	 * here try to allocate 2M pages continously.
	 *
	 * powerpc need to call sparse_init_one_section right after each
	 * sparse_early_mem_map_alloc, so allocate usemap_map at first.
	 */
	size = sizeof(unsigned long *) * NR_MEM_SECTIONS;
	usemap_map = alloc_bootmem(size);
	if (!usemap_map)
		panic("can not allocate usemap_map\n");

	for (pnum = 0; pnum < NR_MEM_SECTIONS; pnum++) {
		if (!present_section_nr(pnum))
			continue;
		usemap_map[pnum] = sparse_early_usemap_alloc(pnum);
	}

	for (pnum = 0; pnum < NR_MEM_SECTIONS; pnum++) {
		if (!present_section_nr(pnum))
			continue;

		usemap = usemap_map[pnum];
		if (!usemap)
			continue;

		map = sparse_early_mem_map_alloc(pnum);
		if (!map)
			continue;

		sparse_init_one_section(__nr_to_section(pnum), pnum, map,
								usemap);
	}

	vmemmap_populate_print_last();

	free_bootmem(__pa(usemap_map), size);
}

#ifdef CONFIG_MEMORY_HOTPLUG
#ifdef CONFIG_SPARSEMEM_VMEMMAP
static inline struct page *kmalloc_section_memmap(unsigned long pnum, int nid,
						 unsigned long nr_pages)
{
	/* This will make the necessary allocations eventually. */
	return sparse_mem_map_populate(pnum, nid);
}
static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
{
	return; /* XXX: Not implemented yet */
}
#else
static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
{
	struct page *page, *ret;
	unsigned long memmap_size = sizeof(struct page) * nr_pages;

	page = alloc_pages(GFP_KERNEL|__GFP_NOWARN, get_order(memmap_size));
	if (page)
		goto got_map_page;

	ret = vmalloc(memmap_size);
	if (ret)
		goto got_map_ptr;

	return NULL;
got_map_page:
	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
got_map_ptr:
	memset(ret, 0, memmap_size);

	return ret;
}

static inline struct page *kmalloc_section_memmap(unsigned long pnum, int nid,
						  unsigned long nr_pages)
{
	return __kmalloc_section_memmap(nr_pages);
}

static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
{
	if (is_vmalloc_addr(memmap))
		vfree(memmap);
	else
		free_pages((unsigned long)memmap,
			   get_order(sizeof(struct page) * nr_pages));
}
#endif /* CONFIG_SPARSEMEM_VMEMMAP */

static void free_section_usemap(struct page *memmap, unsigned long *usemap)
{
	if (!usemap)
		return;

	/*
	 * Check to see if allocation came from hot-plug-add
	 */
	if (PageSlab(virt_to_page(usemap))) {
		kfree(usemap);
		if (memmap)
			__kfree_section_memmap(memmap, PAGES_PER_SECTION);
		return;
	}

	/*
	 * TODO: Allocations came from bootmem - how do I free up ?
	 */
	printk(KERN_WARNING "Not freeing up allocations from bootmem "
			"- leaking memory\n");
}

/*
 * returns the number of sections whose mem_maps were properly
 * set.  If this is <=0, then that means that the passed-in
 * map was not consumed and must be freed.
 */
int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
			   int nr_pages)
{
	unsigned long section_nr = pfn_to_section_nr(start_pfn);
	struct pglist_data *pgdat = zone->zone_pgdat;
	struct mem_section *ms;
	struct page *memmap;
	unsigned long *usemap;
	unsigned long flags;
	int ret;

	/*
	 * no locking for this, because it does its own
	 * plus, it does a kmalloc
	 */
	ret = sparse_index_init(section_nr, pgdat->node_id);
	if (ret < 0 && ret != -EEXIST)
		return ret;
	memmap = kmalloc_section_memmap(section_nr, pgdat->node_id, nr_pages);
	if (!memmap)
		return -ENOMEM;
	usemap = __kmalloc_section_usemap();
	if (!usemap) {
		__kfree_section_memmap(memmap, nr_pages);
		return -ENOMEM;
	}

	pgdat_resize_lock(pgdat, &flags);

	ms = __pfn_to_section(start_pfn);
	if (ms->section_mem_map & SECTION_MARKED_PRESENT) {
		ret = -EEXIST;
		goto out;
	}

	ms->section_mem_map |= SECTION_MARKED_PRESENT;

	ret = sparse_init_one_section(ms, section_nr, memmap, usemap);

out:
	pgdat_resize_unlock(pgdat, &flags);
	if (ret <= 0) {
		kfree(usemap);
		__kfree_section_memmap(memmap, nr_pages);
	}
	return ret;
}

void sparse_remove_one_section(struct zone *zone, struct mem_section *ms)
{
	struct page *memmap = NULL;
	unsigned long *usemap = NULL;

	if (ms->section_mem_map) {
		usemap = ms->pageblock_flags;
		memmap = sparse_decode_mem_map(ms->section_mem_map,
						__section_nr(ms));
		ms->section_mem_map = 0;
		ms->pageblock_flags = NULL;
	}

	free_section_usemap(memmap, usemap);
}
#endif
Commit	Line	Data
d41dee36 AW	1	/*
	2	* sparse memory mappings.
	3	*/
d41dee36 AW	4	#include <linux/mm.h>
	5	#include <linux/mmzone.h>
	6	#include <linux/bootmem.h>
0b0acbec	7	#include <linux/highmem.h>
d41dee36	8	#include <linux/module.h>
28ae55c9	9	#include <linux/spinlock.h>
0b0acbec	10	#include <linux/vmalloc.h>
d41dee36	11	#include <asm/dma.h>
8f6aac41 CL	12	#include <asm/pgalloc.h>
8f6aac41 CL	13	#include <asm/pgtable.h>
d41dee36 AW	14
	15	/*
	16	* Permanent SPARSEMEM data:
	17	*
	18	* 1) mem_section - memory sections, mem_map's for valid memory
	19	*/
3e347261	20	#ifdef CONFIG_SPARSEMEM_EXTREME
802f192e	21	struct mem_section *mem_section[NR_SECTION_ROOTS]
22fc6ecc	22	____cacheline_internodealigned_in_smp;
3e347261 BP	23	#else
3e347261 BP	24	struct mem_section mem_section[NR_SECTION_ROOTS][SECTIONS_PER_ROOT]
22fc6ecc	25	____cacheline_internodealigned_in_smp;
3e347261 BP	26	#endif
	27	EXPORT_SYMBOL(mem_section);
	28
89689ae7 CL	29	#ifdef NODE_NOT_IN_PAGE_FLAGS
	30	/*
	31	* If we did not store the node number in the page then we have to
	32	* do a lookup in the section_to_node_table in order to find which
	33	* node the page belongs to.
	34	*/
	35	#if MAX_NUMNODES <= 256
	36	static u8 section_to_node_table[NR_MEM_SECTIONS] __cacheline_aligned;
	37	#else
	38	static u16 section_to_node_table[NR_MEM_SECTIONS] __cacheline_aligned;
	39	#endif
	40
25ba77c1	41	int page_to_nid(struct page *page)
89689ae7 CL	42	{
	43	return section_to_node_table[page_to_section(page)];
	44	}
	45	EXPORT_SYMBOL(page_to_nid);
85770ffe AW	46
	47	static void set_section_nid(unsigned long section_nr, int nid)
	48	{
	49	section_to_node_table[section_nr] = nid;
	50	}
	51	#else /* !NODE_NOT_IN_PAGE_FLAGS */
	52	static inline void set_section_nid(unsigned long section_nr, int nid)
	53	{
	54	}
89689ae7 CL	55	#endif
89689ae7 CL	56
3e347261	57	#ifdef CONFIG_SPARSEMEM_EXTREME
577a32f6	58	static struct mem_section noinline __init_refok *sparse_index_alloc(int nid)
28ae55c9 DH	59	{
	60	struct mem_section *section = NULL;
	61	unsigned long array_size = SECTIONS_PER_ROOT *
	62	sizeof(struct mem_section);
	63
39d24e64	64	if (slab_is_available())
46a66eec MK	65	section = kmalloc_node(array_size, GFP_KERNEL, nid);
	66	else
	67	section = alloc_bootmem_node(NODE_DATA(nid), array_size);
28ae55c9 DH	68
	69	if (section)
	70	memset(section, 0, array_size);
	71
	72	return section;
3e347261	73	}
802f192e	74
a3142c8e	75	static int __meminit sparse_index_init(unsigned long section_nr, int nid)
802f192e	76	{
34af946a	77	static DEFINE_SPINLOCK(index_init_lock);
28ae55c9 DH	78	unsigned long root = SECTION_NR_TO_ROOT(section_nr);
	79	struct mem_section *section;
	80	int ret = 0;
802f192e BP	81
802f192e BP	82	if (mem_section[root])
28ae55c9	83	return -EEXIST;
3e347261	84
28ae55c9	85	section = sparse_index_alloc(nid);
af0cd5a7 WC	86	if (!section)
af0cd5a7 WC	87	return -ENOMEM;
28ae55c9 DH	88	/*
	89	* This lock keeps two different sections from
	90	* reallocating for the same index
	91	*/
	92	spin_lock(&index_init_lock);
3e347261	93
28ae55c9 DH	94	if (mem_section[root]) {
	95	ret = -EEXIST;
	96	goto out;
	97	}
	98
	99	mem_section[root] = section;
	100	out:
	101	spin_unlock(&index_init_lock);
	102	return ret;
	103	}
	104	#else /* !SPARSEMEM_EXTREME */
	105	static inline int sparse_index_init(unsigned long section_nr, int nid)
	106	{
	107	return 0;
802f192e	108	}
28ae55c9 DH	109	#endif
28ae55c9 DH	110
4ca644d9 DH	111	/*
4ca644d9 DH	112	* Although written for the SPARSEMEM_EXTREME case, this happens
cd881a6b	113	* to also work for the flat array case because
4ca644d9 DH	114	* NR_SECTION_ROOTS==NR_MEM_SECTIONS.
	115	*/
	116	int __section_nr(struct mem_section* ms)
	117	{
	118	unsigned long root_nr;
	119	struct mem_section* root;
	120
12783b00 MK	121	for (root_nr = 0; root_nr < NR_SECTION_ROOTS; root_nr++) {
12783b00 MK	122	root = __nr_to_section(root_nr * SECTIONS_PER_ROOT);
4ca644d9 DH	123	if (!root)
	124	continue;
	125
	126	if ((ms >= root) && (ms < (root + SECTIONS_PER_ROOT)))
	127	break;
	128	}
	129
	130	return (root_nr * SECTIONS_PER_ROOT) + (ms - root);
	131	}
	132
30c253e6 AW	133	/*
	134	* During early boot, before section_mem_map is used for an actual
	135	* mem_map, we use section_mem_map to store the section's NUMA
	136	* node. This keeps us from having to use another data structure. The
	137	* node information is cleared just before we store the real mem_map.
	138	*/
	139	static inline unsigned long sparse_encode_early_nid(int nid)
	140	{
	141	return (nid << SECTION_NID_SHIFT);
	142	}
	143
	144	static inline int sparse_early_nid(struct mem_section *section)
	145	{
	146	return (section->section_mem_map >> SECTION_NID_SHIFT);
	147	}
	148
d41dee36	149	/* Record a memory area against a node. */
a3142c8e	150	void __init memory_present(int nid, unsigned long start, unsigned long end)
d41dee36	151	{
bead9a3a	152	unsigned long max_arch_pfn = 1UL << (MAX_PHYSMEM_BITS-PAGE_SHIFT);
d41dee36 AW	153	unsigned long pfn;
d41dee36 AW	154
bead9a3a IM	155	/*
	156	* Sanity checks - do not allow an architecture to pass
	157	* in larger pfns than the maximum scope of sparsemem:
	158	*/
	159	if (start >= max_arch_pfn)
	160	return;
	161	if (end >= max_arch_pfn)
	162	end = max_arch_pfn;
	163
d41dee36 AW	164	start &= PAGE_SECTION_MASK;
	165	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
	166	unsigned long section = pfn_to_section_nr(pfn);
802f192e BP	167	struct mem_section *ms;
	168
	169	sparse_index_init(section, nid);
85770ffe	170	set_section_nid(section, nid);
802f192e BP	171
	172	ms = __nr_to_section(section);
	173	if (!ms->section_mem_map)
30c253e6 AW	174	ms->section_mem_map = sparse_encode_early_nid(nid) \|
30c253e6 AW	175	SECTION_MARKED_PRESENT;
d41dee36 AW	176	}
	177	}
	178
	179	/*
	180	* Only used by the i386 NUMA architecures, but relatively
	181	* generic code.
	182	*/
	183	unsigned long __init node_memmap_size_bytes(int nid, unsigned long start_pfn,
	184	unsigned long end_pfn)
	185	{
	186	unsigned long pfn;
	187	unsigned long nr_pages = 0;
	188
	189	for (pfn = start_pfn; pfn < end_pfn; pfn += PAGES_PER_SECTION) {
	190	if (nid != early_pfn_to_nid(pfn))
	191	continue;
	192
540557b9	193	if (pfn_present(pfn))
d41dee36 AW	194	nr_pages += PAGES_PER_SECTION;
	195	}
	196
	197	return nr_pages * sizeof(struct page);
	198	}
	199
29751f69 AW	200	/*
	201	* Subtle, we encode the real pfn into the mem_map such that
	202	* the identity pfn - section_mem_map will return the actual
	203	* physical page frame number.
	204	*/
	205	static unsigned long sparse_encode_mem_map(struct page *mem_map, unsigned long pnum)
	206	{
	207	return (unsigned long)(mem_map - (section_nr_to_pfn(pnum)));
	208	}
	209
	210	/*
ea01ea93	211	* Decode mem_map from the coded memmap
29751f69	212	*/
29751f69 AW	213	struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum)
29751f69 AW	214	{
ea01ea93 BP	215	/* mask off the extra low bits of information */
ea01ea93 BP	216	coded_mem_map &= SECTION_MAP_MASK;
29751f69 AW	217	return ((struct page *)coded_mem_map) + section_nr_to_pfn(pnum);
	218	}
	219
a3142c8e	220	static int __meminit sparse_init_one_section(struct mem_section *ms,
5c0e3066 MG	221	unsigned long pnum, struct page *mem_map,
5c0e3066 MG	222	unsigned long *pageblock_bitmap)
29751f69	223	{
540557b9	224	if (!present_section(ms))
29751f69 AW	225	return -EINVAL;
29751f69 AW	226
30c253e6	227	ms->section_mem_map &= ~SECTION_MAP_MASK;
540557b9 AW	228	ms->section_mem_map \|= sparse_encode_mem_map(mem_map, pnum) \|
540557b9 AW	229	SECTION_HAS_MEM_MAP;
5c0e3066	230	ms->pageblock_flags = pageblock_bitmap;
29751f69 AW	231
	232	return 1;
	233	}
	234
04753278	235	unsigned long usemap_size(void)
5c0e3066 MG	236	{
	237	unsigned long size_bytes;
	238	size_bytes = roundup(SECTION_BLOCKFLAGS_BITS, 8) / 8;
	239	size_bytes = roundup(size_bytes, sizeof(unsigned long));
	240	return size_bytes;
	241	}
	242
	243	#ifdef CONFIG_MEMORY_HOTPLUG
	244	static unsigned long *__kmalloc_section_usemap(void)
	245	{
	246	return kmalloc(usemap_size(), GFP_KERNEL);
	247	}
	248	#endif /* CONFIG_MEMORY_HOTPLUG */
	249
a322f8ab	250	static unsigned long *__init sparse_early_usemap_alloc(unsigned long pnum)
5c0e3066	251	{
86f6dae1	252	unsigned long *usemap, section_nr;
5c0e3066 MG	253	struct mem_section *ms = __nr_to_section(pnum);
5c0e3066 MG	254	int nid = sparse_early_nid(ms);
86f6dae1	255	struct pglist_data *pgdat = NODE_DATA(nid);
5c0e3066	256
86f6dae1 YG	257	/*
	258	* Usemap's page can't be freed until freeing other sections
	259	* which use it. And, Pgdat has same feature.
	260	* If section A has pgdat and section B has usemap for other
	261	* sections (includes section A), both sections can't be removed,
	262	* because there is the dependency each other.
	263	* To solve above issue, this collects all usemap on the same section
	264	* which has pgdat.
	265	*/
	266	section_nr = pfn_to_section_nr(__pa(pgdat) >> PAGE_SHIFT);
	267	usemap = alloc_bootmem_section(usemap_size(), section_nr);
5c0e3066 MG	268	if (usemap)
	269	return usemap;
	270
	271	/* Stupid: suppress gcc warning for SPARSEMEM && !NUMA */
	272	nid = 0;
	273
	274	printk(KERN_WARNING "%s: allocation failed\n", __FUNCTION__);
	275	return NULL;
	276	}
	277
8f6aac41	278	#ifndef CONFIG_SPARSEMEM_VMEMMAP
98f3cfc1	279	struct page __init *sparse_mem_map_populate(unsigned long pnum, int nid)
29751f69 AW	280	{
29751f69 AW	281	struct page *map;
29751f69 AW	282
	283	map = alloc_remap(nid, sizeof(struct page) * PAGES_PER_SECTION);
	284	if (map)
	285	return map;
	286
9d99217a YG	287	map = alloc_bootmem_pages_node(NODE_DATA(nid),
9d99217a YG	288	PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION));
8f6aac41 CL	289	return map;
	290	}
	291	#endif /* !CONFIG_SPARSEMEM_VMEMMAP */
	292
	293	struct page __init *sparse_early_mem_map_alloc(unsigned long pnum)
	294	{
	295	struct page *map;
	296	struct mem_section *ms = __nr_to_section(pnum);
	297	int nid = sparse_early_nid(ms);
	298
98f3cfc1	299	map = sparse_mem_map_populate(pnum, nid);
29751f69 AW	300	if (map)
	301	return map;
	302
8f6aac41 CL	303	printk(KERN_ERR "%s: sparsemem memory map backing failed "
8f6aac41 CL	304	"some memory will not be available.\n", __FUNCTION__);
802f192e	305	ms->section_mem_map = 0;
29751f69 AW	306	return NULL;
	307	}
	308
c2b91e2e YL	309	void __attribute__((weak)) __meminit vmemmap_populate_print_last(void)
	310	{
	311	}
193faea9 SR	312	/*
	313	* Allocate the accumulated non-linear sections, allocate a mem_map
	314	* for each and record the physical to section mapping.
	315	*/
	316	void __init sparse_init(void)
	317	{
	318	unsigned long pnum;
	319	struct page *map;
5c0e3066	320	unsigned long *usemap;
e123dd3f YL	321	unsigned long **usemap_map;
	322	int size;
	323
	324	/*
	325	* map is using big page (aka 2M in x86 64 bit)
	326	* usemap is less one page (aka 24 bytes)
	327	* so alloc 2M (with 2M align) and 24 bytes in turn will
	328	* make next 2M slip to one more 2M later.
	329	* then in big system, the memory will have a lot of holes...
	330	* here try to allocate 2M pages continously.
	331	*
	332	* powerpc need to call sparse_init_one_section right after each
	333	* sparse_early_mem_map_alloc, so allocate usemap_map at first.
	334	*/
	335	size = sizeof(unsigned long ) NR_MEM_SECTIONS;
	336	usemap_map = alloc_bootmem(size);
	337	if (!usemap_map)
	338	panic("can not allocate usemap_map\n");
193faea9 SR	339
193faea9 SR	340	for (pnum = 0; pnum < NR_MEM_SECTIONS; pnum++) {
540557b9	341	if (!present_section_nr(pnum))
193faea9	342	continue;
e123dd3f YL	343	usemap_map[pnum] = sparse_early_usemap_alloc(pnum);
e123dd3f YL	344	}
193faea9	345
e123dd3f YL	346	for (pnum = 0; pnum < NR_MEM_SECTIONS; pnum++) {
e123dd3f YL	347	if (!present_section_nr(pnum))
193faea9	348	continue;
5c0e3066	349
e123dd3f	350	usemap = usemap_map[pnum];
5c0e3066 MG	351	if (!usemap)
	352	continue;
	353
e123dd3f YL	354	map = sparse_early_mem_map_alloc(pnum);
	355	if (!map)
	356	continue;
	357
5c0e3066 MG	358	sparse_init_one_section(__nr_to_section(pnum), pnum, map,
5c0e3066 MG	359	usemap);
193faea9	360	}
e123dd3f	361
c2b91e2e YL	362	vmemmap_populate_print_last();
c2b91e2e YL	363
e123dd3f	364	free_bootmem(__pa(usemap_map), size);
193faea9 SR	365	}
	366
	367	#ifdef CONFIG_MEMORY_HOTPLUG
98f3cfc1 YG	368	#ifdef CONFIG_SPARSEMEM_VMEMMAP
	369	static inline struct page *kmalloc_section_memmap(unsigned long pnum, int nid,
	370	unsigned long nr_pages)
	371	{
	372	/* This will make the necessary allocations eventually. */
	373	return sparse_mem_map_populate(pnum, nid);
	374	}
	375	static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
	376	{
	377	return; /* XXX: Not implemented yet */
	378	}
	379	#else
0b0acbec DH	380	static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
	381	{
	382	struct page page, ret;
	383	unsigned long memmap_size = sizeof(struct page) * nr_pages;
	384
f2d0aa5b	385	page = alloc_pages(GFP_KERNEL\|__GFP_NOWARN, get_order(memmap_size));
0b0acbec DH	386	if (page)
	387	goto got_map_page;
	388
	389	ret = vmalloc(memmap_size);
	390	if (ret)
	391	goto got_map_ptr;
	392
	393	return NULL;
	394	got_map_page:
	395	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
	396	got_map_ptr:
	397	memset(ret, 0, memmap_size);
	398
	399	return ret;
	400	}
	401
98f3cfc1 YG	402	static inline struct page *kmalloc_section_memmap(unsigned long pnum, int nid,
	403	unsigned long nr_pages)
	404	{
	405	return __kmalloc_section_memmap(nr_pages);
	406	}
	407
0b0acbec DH	408	static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
0b0acbec DH	409	{
9e2779fa	410	if (is_vmalloc_addr(memmap))
0b0acbec DH	411	vfree(memmap);
	412	else
	413	free_pages((unsigned long)memmap,
	414	get_order(sizeof(struct page) * nr_pages));
	415	}
98f3cfc1	416	#endif /* CONFIG_SPARSEMEM_VMEMMAP */
0b0acbec	417
ea01ea93 BP	418	static void free_section_usemap(struct page memmap, unsigned long usemap)
	419	{
	420	if (!usemap)
	421	return;
	422
	423	/*
	424	* Check to see if allocation came from hot-plug-add
	425	*/
	426	if (PageSlab(virt_to_page(usemap))) {
	427	kfree(usemap);
	428	if (memmap)
	429	__kfree_section_memmap(memmap, PAGES_PER_SECTION);
	430	return;
	431	}
	432
	433	/*
	434	* TODO: Allocations came from bootmem - how do I free up ?
	435	*/
	436	printk(KERN_WARNING "Not freeing up allocations from bootmem "
	437	"- leaking memory\n");
	438	}
	439
29751f69 AW	440	/*
	441	* returns the number of sections whose mem_maps were properly
	442	* set. If this is <=0, then that means that the passed-in
	443	* map was not consumed and must be freed.
	444	*/
0b0acbec DH	445	int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
0b0acbec DH	446	int nr_pages)
29751f69	447	{
0b0acbec DH	448	unsigned long section_nr = pfn_to_section_nr(start_pfn);
	449	struct pglist_data *pgdat = zone->zone_pgdat;
	450	struct mem_section *ms;
	451	struct page *memmap;
5c0e3066	452	unsigned long *usemap;
0b0acbec DH	453	unsigned long flags;
0b0acbec DH	454	int ret;
29751f69	455
0b0acbec DH	456	/*
	457	* no locking for this, because it does its own
	458	* plus, it does a kmalloc
	459	*/
bbd06825 WC	460	ret = sparse_index_init(section_nr, pgdat->node_id);
	461	if (ret < 0 && ret != -EEXIST)
	462	return ret;
98f3cfc1	463	memmap = kmalloc_section_memmap(section_nr, pgdat->node_id, nr_pages);
bbd06825 WC	464	if (!memmap)
bbd06825 WC	465	return -ENOMEM;
5c0e3066	466	usemap = __kmalloc_section_usemap();
bbd06825 WC	467	if (!usemap) {
	468	__kfree_section_memmap(memmap, nr_pages);
	469	return -ENOMEM;
	470	}
0b0acbec DH	471
0b0acbec DH	472	pgdat_resize_lock(pgdat, &flags);
29751f69	473
0b0acbec DH	474	ms = __pfn_to_section(start_pfn);
	475	if (ms->section_mem_map & SECTION_MARKED_PRESENT) {
	476	ret = -EEXIST;
	477	goto out;
	478	}
5c0e3066	479
29751f69 AW	480	ms->section_mem_map \|= SECTION_MARKED_PRESENT;
29751f69 AW	481
5c0e3066	482	ret = sparse_init_one_section(ms, section_nr, memmap, usemap);
0b0acbec	483
0b0acbec DH	484	out:
0b0acbec DH	485	pgdat_resize_unlock(pgdat, &flags);
bbd06825 WC	486	if (ret <= 0) {
bbd06825 WC	487	kfree(usemap);
46a66eec	488	__kfree_section_memmap(memmap, nr_pages);
bbd06825	489	}
0b0acbec	490	return ret;
29751f69	491	}
ea01ea93 BP	492
	493	void sparse_remove_one_section(struct zone zone, struct mem_section ms)
	494	{
	495	struct page *memmap = NULL;
	496	unsigned long *usemap = NULL;
	497
	498	if (ms->section_mem_map) {
	499	usemap = ms->pageblock_flags;
	500	memmap = sparse_decode_mem_map(ms->section_mem_map,
	501	__section_nr(ms));
	502	ms->section_mem_map = 0;
	503	ms->pageblock_flags = NULL;
	504	}
	505
	506	free_section_usemap(memmap, usemap);
	507	}
a3142c8e	508	#endif