[linux.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>

/*
 * 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 CONFIG_SPARSEMEM_EXTREME
static struct mem_section *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 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);
	/*
	 * 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 becase
 * 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 memory_present(int nid, unsigned long start, unsigned long end)
{
	unsigned long 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);

		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_valid(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)));
}

/*
 * We need this if we ever free the mem_maps.  While not implemented yet,
 * this function is included for parity with its sibling.
 */
static __attribute((unused))
struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum)
{
	return ((struct page *)coded_mem_map) + section_nr_to_pfn(pnum);
}

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

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

	return 1;
}

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

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

	printk(KERN_WARNING "%s: allocation failed\n", __FUNCTION__);
	ms->section_mem_map = 0;
	return NULL;
}

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, 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 int vaddr_in_vmalloc_area(void *addr)
{
	if (addr >= (void *)VMALLOC_START &&
	    addr < (void *)VMALLOC_END)
		return 1;
	return 0;
}

static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
{
	if (vaddr_in_vmalloc_area(memmap))
		vfree(memmap);
	else
		free_pages((unsigned long)memmap,
			   get_order(sizeof(struct page) * nr_pages));
}

/*
 * Allocate the accumulated non-linear sections, allocate a mem_map
 * for each and record the physical to section mapping.
 */
void sparse_init(void)
{
	unsigned long pnum;
	struct page *map;

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

		map = sparse_early_mem_map_alloc(pnum);
		if (!map)
			continue;
		sparse_init_one_section(__nr_to_section(pnum), pnum, map);
	}
}

/*
 * 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 flags;
	int ret;

	/*
	 * no locking for this, because it does its own
	 * plus, it does a kmalloc
	 */
	sparse_index_init(section_nr, pgdat->node_id);
	memmap = __kmalloc_section_memmap(nr_pages);

	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);

out:
	pgdat_resize_unlock(pgdat, &flags);
	if (ret <= 0)
		__kfree_section_memmap(memmap, nr_pages);
	return ret;
}
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 AW	11	#include <asm/dma.h>
	12
	13	/*
	14	* Permanent SPARSEMEM data:
	15	*
	16	* 1) mem_section - memory sections, mem_map's for valid memory
	17	*/
3e347261	18	#ifdef CONFIG_SPARSEMEM_EXTREME
802f192e	19	struct mem_section *mem_section[NR_SECTION_ROOTS]
22fc6ecc	20	____cacheline_internodealigned_in_smp;
3e347261 BP	21	#else
3e347261 BP	22	struct mem_section mem_section[NR_SECTION_ROOTS][SECTIONS_PER_ROOT]
22fc6ecc	23	____cacheline_internodealigned_in_smp;
3e347261 BP	24	#endif
	25	EXPORT_SYMBOL(mem_section);
	26
3e347261	27	#ifdef CONFIG_SPARSEMEM_EXTREME
28ae55c9 DH	28	static struct mem_section *sparse_index_alloc(int nid)
	29	{
	30	struct mem_section *section = NULL;
	31	unsigned long array_size = SECTIONS_PER_ROOT *
	32	sizeof(struct mem_section);
	33
39d24e64	34	if (slab_is_available())
46a66eec MK	35	section = kmalloc_node(array_size, GFP_KERNEL, nid);
	36	else
	37	section = alloc_bootmem_node(NODE_DATA(nid), array_size);
28ae55c9 DH	38
	39	if (section)
	40	memset(section, 0, array_size);
	41
	42	return section;
3e347261	43	}
802f192e	44
28ae55c9	45	static int sparse_index_init(unsigned long section_nr, int nid)
802f192e	46	{
34af946a	47	static DEFINE_SPINLOCK(index_init_lock);
28ae55c9 DH	48	unsigned long root = SECTION_NR_TO_ROOT(section_nr);
	49	struct mem_section *section;
	50	int ret = 0;
802f192e BP	51
802f192e BP	52	if (mem_section[root])
28ae55c9	53	return -EEXIST;
3e347261	54
28ae55c9 DH	55	section = sparse_index_alloc(nid);
	56	/*
	57	* This lock keeps two different sections from
	58	* reallocating for the same index
	59	*/
	60	spin_lock(&index_init_lock);
3e347261	61
28ae55c9 DH	62	if (mem_section[root]) {
	63	ret = -EEXIST;
	64	goto out;
	65	}
	66
	67	mem_section[root] = section;
	68	out:
	69	spin_unlock(&index_init_lock);
	70	return ret;
	71	}
	72	#else /* !SPARSEMEM_EXTREME */
	73	static inline int sparse_index_init(unsigned long section_nr, int nid)
	74	{
	75	return 0;
802f192e	76	}
28ae55c9 DH	77	#endif
28ae55c9 DH	78
4ca644d9 DH	79	/*
	80	* Although written for the SPARSEMEM_EXTREME case, this happens
	81	* to also work for the flat array case becase
	82	* NR_SECTION_ROOTS==NR_MEM_SECTIONS.
	83	*/
	84	int __section_nr(struct mem_section* ms)
	85	{
	86	unsigned long root_nr;
	87	struct mem_section* root;
	88
12783b00 MK	89	for (root_nr = 0; root_nr < NR_SECTION_ROOTS; root_nr++) {
12783b00 MK	90	root = __nr_to_section(root_nr * SECTIONS_PER_ROOT);
4ca644d9 DH	91	if (!root)
	92	continue;
	93
	94	if ((ms >= root) && (ms < (root + SECTIONS_PER_ROOT)))
	95	break;
	96	}
	97
	98	return (root_nr * SECTIONS_PER_ROOT) + (ms - root);
	99	}
	100
30c253e6 AW	101	/*
	102	* During early boot, before section_mem_map is used for an actual
	103	* mem_map, we use section_mem_map to store the section's NUMA
	104	* node. This keeps us from having to use another data structure. The
	105	* node information is cleared just before we store the real mem_map.
	106	*/
	107	static inline unsigned long sparse_encode_early_nid(int nid)
	108	{
	109	return (nid << SECTION_NID_SHIFT);
	110	}
	111
	112	static inline int sparse_early_nid(struct mem_section *section)
	113	{
	114	return (section->section_mem_map >> SECTION_NID_SHIFT);
	115	}
	116
d41dee36 AW	117	/* Record a memory area against a node. */
	118	void memory_present(int nid, unsigned long start, unsigned long end)
	119	{
	120	unsigned long pfn;
	121
	122	start &= PAGE_SECTION_MASK;
	123	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION) {
	124	unsigned long section = pfn_to_section_nr(pfn);
802f192e BP	125	struct mem_section *ms;
	126
	127	sparse_index_init(section, nid);
	128
	129	ms = __nr_to_section(section);
	130	if (!ms->section_mem_map)
30c253e6 AW	131	ms->section_mem_map = sparse_encode_early_nid(nid) \|
30c253e6 AW	132	SECTION_MARKED_PRESENT;
d41dee36 AW	133	}
	134	}
	135
	136	/*
	137	* Only used by the i386 NUMA architecures, but relatively
	138	* generic code.
	139	*/
	140	unsigned long __init node_memmap_size_bytes(int nid, unsigned long start_pfn,
	141	unsigned long end_pfn)
	142	{
	143	unsigned long pfn;
	144	unsigned long nr_pages = 0;
	145
	146	for (pfn = start_pfn; pfn < end_pfn; pfn += PAGES_PER_SECTION) {
	147	if (nid != early_pfn_to_nid(pfn))
	148	continue;
	149
	150	if (pfn_valid(pfn))
	151	nr_pages += PAGES_PER_SECTION;
	152	}
	153
	154	return nr_pages * sizeof(struct page);
	155	}
	156
29751f69 AW	157	/*
	158	* Subtle, we encode the real pfn into the mem_map such that
	159	* the identity pfn - section_mem_map will return the actual
	160	* physical page frame number.
	161	*/
	162	static unsigned long sparse_encode_mem_map(struct page *mem_map, unsigned long pnum)
	163	{
	164	return (unsigned long)(mem_map - (section_nr_to_pfn(pnum)));
	165	}
	166
	167	/*
	168	* We need this if we ever free the mem_maps. While not implemented yet,
	169	* this function is included for parity with its sibling.
	170	*/
	171	static __attribute((unused))
	172	struct page *sparse_decode_mem_map(unsigned long coded_mem_map, unsigned long pnum)
	173	{
	174	return ((struct page *)coded_mem_map) + section_nr_to_pfn(pnum);
	175	}
	176
	177	static int sparse_init_one_section(struct mem_section *ms,
	178	unsigned long pnum, struct page *mem_map)
	179	{
	180	if (!valid_section(ms))
	181	return -EINVAL;
	182
30c253e6	183	ms->section_mem_map &= ~SECTION_MAP_MASK;
29751f69 AW	184	ms->section_mem_map \|= sparse_encode_mem_map(mem_map, pnum);
	185
	186	return 1;
	187	}
	188
	189	static struct page *sparse_early_mem_map_alloc(unsigned long pnum)
	190	{
	191	struct page *map;
802f192e	192	struct mem_section *ms = __nr_to_section(pnum);
30c253e6	193	int nid = sparse_early_nid(ms);
29751f69 AW	194
	195	map = alloc_remap(nid, sizeof(struct page) * PAGES_PER_SECTION);
	196	if (map)
	197	return map;
	198
	199	map = alloc_bootmem_node(NODE_DATA(nid),
	200	sizeof(struct page) * PAGES_PER_SECTION);
	201	if (map)
	202	return map;
	203
	204	printk(KERN_WARNING "%s: allocation failed\n", __FUNCTION__);
802f192e	205	ms->section_mem_map = 0;
29751f69 AW	206	return NULL;
	207	}
	208
0b0acbec DH	209	static struct page *__kmalloc_section_memmap(unsigned long nr_pages)
	210	{
	211	struct page page, ret;
	212	unsigned long memmap_size = sizeof(struct page) * nr_pages;
	213
	214	page = alloc_pages(GFP_KERNEL, get_order(memmap_size));
	215	if (page)
	216	goto got_map_page;
	217
	218	ret = vmalloc(memmap_size);
	219	if (ret)
	220	goto got_map_ptr;
	221
	222	return NULL;
	223	got_map_page:
	224	ret = (struct page *)pfn_to_kaddr(page_to_pfn(page));
	225	got_map_ptr:
	226	memset(ret, 0, memmap_size);
	227
	228	return ret;
	229	}
	230
	231	static int vaddr_in_vmalloc_area(void *addr)
	232	{
	233	if (addr >= (void *)VMALLOC_START &&
	234	addr < (void *)VMALLOC_END)
	235	return 1;
	236	return 0;
	237	}
	238
	239	static void __kfree_section_memmap(struct page *memmap, unsigned long nr_pages)
	240	{
	241	if (vaddr_in_vmalloc_area(memmap))
	242	vfree(memmap);
	243	else
	244	free_pages((unsigned long)memmap,
	245	get_order(sizeof(struct page) * nr_pages));
	246	}
	247
d41dee36 AW	248	/*
	249	* Allocate the accumulated non-linear sections, allocate a mem_map
	250	* for each and record the physical to section mapping.
	251	*/
	252	void sparse_init(void)
	253	{
	254	unsigned long pnum;
	255	struct page *map;
d41dee36 AW	256
d41dee36 AW	257	for (pnum = 0; pnum < NR_MEM_SECTIONS; pnum++) {
29751f69	258	if (!valid_section_nr(pnum))
d41dee36 AW	259	continue;
d41dee36 AW	260
29751f69	261	map = sparse_early_mem_map_alloc(pnum);
802f192e BP	262	if (!map)
	263	continue;
	264	sparse_init_one_section(__nr_to_section(pnum), pnum, map);
d41dee36 AW	265	}
d41dee36 AW	266	}
29751f69 AW	267
	268	/*
	269	* returns the number of sections whose mem_maps were properly
	270	* set. If this is <=0, then that means that the passed-in
	271	* map was not consumed and must be freed.
	272	*/
0b0acbec DH	273	int sparse_add_one_section(struct zone *zone, unsigned long start_pfn,
0b0acbec DH	274	int nr_pages)
29751f69	275	{
0b0acbec DH	276	unsigned long section_nr = pfn_to_section_nr(start_pfn);
	277	struct pglist_data *pgdat = zone->zone_pgdat;
	278	struct mem_section *ms;
	279	struct page *memmap;
	280	unsigned long flags;
	281	int ret;
29751f69	282
0b0acbec DH	283	/*
	284	* no locking for this, because it does its own
	285	* plus, it does a kmalloc
	286	*/
	287	sparse_index_init(section_nr, pgdat->node_id);
	288	memmap = __kmalloc_section_memmap(nr_pages);
	289
	290	pgdat_resize_lock(pgdat, &flags);
29751f69	291
0b0acbec DH	292	ms = __pfn_to_section(start_pfn);
	293	if (ms->section_mem_map & SECTION_MARKED_PRESENT) {
	294	ret = -EEXIST;
	295	goto out;
	296	}
29751f69 AW	297	ms->section_mem_map \|= SECTION_MARKED_PRESENT;
29751f69 AW	298
0b0acbec DH	299	ret = sparse_init_one_section(ms, section_nr, memmap);
0b0acbec DH	300
0b0acbec DH	301	out:
0b0acbec DH	302	pgdat_resize_unlock(pgdat, &flags);
46a66eec MK	303	if (ret <= 0)
46a66eec MK	304	__kfree_section_memmap(memmap, nr_pages);
0b0acbec	305	return ret;
29751f69	306	}