[net-next-2.6.git] / mm / memory_hotplug.c

/*
 *  linux/mm/memory_hotplug.c
 *
 *  Copyright (C)
 */

#include <linux/stddef.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/interrupt.h>
#include <linux/pagemap.h>
#include <linux/bootmem.h>
#include <linux/compiler.h>
#include <linux/module.h>
#include <linux/pagevec.h>
#include <linux/writeback.h>
#include <linux/slab.h>
#include <linux/sysctl.h>
#include <linux/cpu.h>
#include <linux/memory.h>
#include <linux/memory_hotplug.h>
#include <linux/highmem.h>
#include <linux/vmalloc.h>
#include <linux/ioport.h>
#include <linux/cpuset.h>

#include <asm/tlbflush.h>

/* add this memory to iomem resource */
static struct resource *register_memory_resource(u64 start, u64 size)
{
	struct resource *res;
	res = kzalloc(sizeof(struct resource), GFP_KERNEL);
	BUG_ON(!res);

	res->name = "System RAM";
	res->start = start;
	res->end = start + size - 1;
	res->flags = IORESOURCE_MEM;
	if (request_resource(&iomem_resource, res) < 0) {
		printk("System RAM resource %llx - %llx cannot be added\n",
		(unsigned long long)res->start, (unsigned long long)res->end);
		kfree(res);
		res = NULL;
	}
	return res;
}

static void release_memory_resource(struct resource *res)
{
	if (!res)
		return;
	release_resource(res);
	kfree(res);
	return;
}


#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
static int __add_zone(struct zone *zone, unsigned long phys_start_pfn)
{
	struct pglist_data *pgdat = zone->zone_pgdat;
	int nr_pages = PAGES_PER_SECTION;
	int nid = pgdat->node_id;
	int zone_type;

	zone_type = zone - pgdat->node_zones;
	if (!zone->wait_table) {
		int ret = 0;
		ret = init_currently_empty_zone(zone, phys_start_pfn,
						nr_pages, MEMMAP_HOTPLUG);
		if (ret < 0)
			return ret;
	}
	memmap_init_zone(nr_pages, nid, zone_type,
			 phys_start_pfn, MEMMAP_HOTPLUG);
	return 0;
}

static int __add_section(struct zone *zone, unsigned long phys_start_pfn)
{
	int nr_pages = PAGES_PER_SECTION;
	int ret;

	if (pfn_valid(phys_start_pfn))
		return -EEXIST;

	ret = sparse_add_one_section(zone, phys_start_pfn, nr_pages);

	if (ret < 0)
		return ret;

	ret = __add_zone(zone, phys_start_pfn);

	if (ret < 0)
		return ret;

	return register_new_memory(__pfn_to_section(phys_start_pfn));
}

/*
 * Reasonably generic function for adding memory.  It is
 * expected that archs that support memory hotplug will
 * call this function after deciding the zone to which to
 * add the new pages.
 */
int __add_pages(struct zone *zone, unsigned long phys_start_pfn,
		 unsigned long nr_pages)
{
	unsigned long i;
	int err = 0;
	int start_sec, end_sec;
	/* during initialize mem_map, align hot-added range to section */
	start_sec = pfn_to_section_nr(phys_start_pfn);
	end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);

	for (i = start_sec; i <= end_sec; i++) {
		err = __add_section(zone, i << PFN_SECTION_SHIFT);

		/*
		 * EEXIST is finally dealed with by ioresource collision
		 * check. see add_memory() => register_memory_resource()
		 * Warning will be printed if there is collision.
		 */
		if (err && (err != -EEXIST))
			break;
		err = 0;
	}

	return err;
}
EXPORT_SYMBOL_GPL(__add_pages);

static void grow_zone_span(struct zone *zone,
		unsigned long start_pfn, unsigned long end_pfn)
{
	unsigned long old_zone_end_pfn;

	zone_span_writelock(zone);

	old_zone_end_pfn = zone->zone_start_pfn + zone->spanned_pages;
	if (start_pfn < zone->zone_start_pfn)
		zone->zone_start_pfn = start_pfn;

	zone->spanned_pages = max(old_zone_end_pfn, end_pfn) -
				zone->zone_start_pfn;

	zone_span_writeunlock(zone);
}

static void grow_pgdat_span(struct pglist_data *pgdat,
		unsigned long start_pfn, unsigned long end_pfn)
{
	unsigned long old_pgdat_end_pfn =
		pgdat->node_start_pfn + pgdat->node_spanned_pages;

	if (start_pfn < pgdat->node_start_pfn)
		pgdat->node_start_pfn = start_pfn;

	pgdat->node_spanned_pages = max(old_pgdat_end_pfn, end_pfn) -
					pgdat->node_start_pfn;
}

int online_pages(unsigned long pfn, unsigned long nr_pages)
{
	unsigned long i;
	unsigned long flags;
	unsigned long onlined_pages = 0;
	struct resource res;
	u64 section_end;
	unsigned long start_pfn;
	struct zone *zone;
	int need_zonelists_rebuild = 0;

	/*
	 * This doesn't need a lock to do pfn_to_page().
	 * The section can't be removed here because of the
	 * memory_block->state_sem.
	 */
	zone = page_zone(pfn_to_page(pfn));
	pgdat_resize_lock(zone->zone_pgdat, &flags);
	grow_zone_span(zone, pfn, pfn + nr_pages);
	grow_pgdat_span(zone->zone_pgdat, pfn, pfn + nr_pages);
	pgdat_resize_unlock(zone->zone_pgdat, &flags);

	/*
	 * If this zone is not populated, then it is not in zonelist.
	 * This means the page allocator ignores this zone.
	 * So, zonelist must be updated after online.
	 */
	if (!populated_zone(zone))
		need_zonelists_rebuild = 1;

	res.start = (u64)pfn << PAGE_SHIFT;
	res.end = res.start + ((u64)nr_pages << PAGE_SHIFT) - 1;
	res.flags = IORESOURCE_MEM; /* we just need system ram */
	section_end = res.end;

	while ((res.start < res.end) && (find_next_system_ram(&res) >= 0)) {
		start_pfn = (unsigned long)(res.start >> PAGE_SHIFT);
		nr_pages = (unsigned long)
                           ((res.end + 1 - res.start) >> PAGE_SHIFT);

		if (PageReserved(pfn_to_page(start_pfn))) {
			/* this region's page is not onlined now */
			for (i = 0; i < nr_pages; i++) {
				struct page *page = pfn_to_page(start_pfn + i);
				online_page(page);
				onlined_pages++;
			}
		}

		res.start = res.end + 1;
		res.end = section_end;
	}
	zone->present_pages += onlined_pages;
	zone->zone_pgdat->node_present_pages += onlined_pages;

	setup_per_zone_pages_min();

	if (need_zonelists_rebuild)
		build_all_zonelists();
	vm_total_pages = nr_free_pagecache_pages();
	writeback_set_ratelimit();
	return 0;
}
#endif /* CONFIG_MEMORY_HOTPLUG_SPARSE */

static pg_data_t *hotadd_new_pgdat(int nid, u64 start)
{
	struct pglist_data *pgdat;
	unsigned long zones_size[MAX_NR_ZONES] = {0};
	unsigned long zholes_size[MAX_NR_ZONES] = {0};
	unsigned long start_pfn = start >> PAGE_SHIFT;

	pgdat = arch_alloc_nodedata(nid);
	if (!pgdat)
		return NULL;

	arch_refresh_nodedata(nid, pgdat);

	/* we can use NODE_DATA(nid) from here */

	/* init node's zones as empty zones, we don't have any present pages.*/
	free_area_init_node(nid, pgdat, zones_size, start_pfn, zholes_size);

	return pgdat;
}

static void rollback_node_hotadd(int nid, pg_data_t *pgdat)
{
	arch_refresh_nodedata(nid, NULL);
	arch_free_nodedata(pgdat);
	return;
}


int add_memory(int nid, u64 start, u64 size)
{
	pg_data_t *pgdat = NULL;
	int new_pgdat = 0;
	struct resource *res;
	int ret;

	res = register_memory_resource(start, size);
	if (!res)
		return -EEXIST;

	if (!node_online(nid)) {
		pgdat = hotadd_new_pgdat(nid, start);
		if (!pgdat)
			return -ENOMEM;
		new_pgdat = 1;
		ret = kswapd_run(nid);
		if (ret)
			goto error;
	}

	/* call arch's memory hotadd */
	ret = arch_add_memory(nid, start, size);

	if (ret < 0)
		goto error;

	/* we online node here. we can't roll back from here. */
	node_set_online(nid);

	cpuset_track_online_nodes();

	if (new_pgdat) {
		ret = register_one_node(nid);
		/*
		 * If sysfs file of new node can't create, cpu on the node
		 * can't be hot-added. There is no rollback way now.
		 * So, check by BUG_ON() to catch it reluctantly..
		 */
		BUG_ON(ret);
	}

	return ret;
error:
	/* rollback pgdat allocation and others */
	if (new_pgdat)
		rollback_node_hotadd(nid, pgdat);
	if (res)
		release_memory_resource(res);

	return ret;
}
EXPORT_SYMBOL_GPL(add_memory);
Commit	Line	Data
3947be19 DH	1	/*
	2	* linux/mm/memory_hotplug.c
	3	*
	4	* Copyright (C)
	5	*/
	6
3947be19 DH	7	#include <linux/stddef.h>
	8	#include <linux/mm.h>
	9	#include <linux/swap.h>
	10	#include <linux/interrupt.h>
	11	#include <linux/pagemap.h>
	12	#include <linux/bootmem.h>
	13	#include <linux/compiler.h>
	14	#include <linux/module.h>
	15	#include <linux/pagevec.h>
2d1d43f6	16	#include <linux/writeback.h>
3947be19 DH	17	#include <linux/slab.h>
	18	#include <linux/sysctl.h>
	19	#include <linux/cpu.h>
	20	#include <linux/memory.h>
	21	#include <linux/memory_hotplug.h>
	22	#include <linux/highmem.h>
	23	#include <linux/vmalloc.h>
0a547039	24	#include <linux/ioport.h>
38837fc7	25	#include <linux/cpuset.h>
3947be19 DH	26
	27	#include <asm/tlbflush.h>
	28
45e0b78b KM	29	/* add this memory to iomem resource */
	30	static struct resource *register_memory_resource(u64 start, u64 size)
	31	{
	32	struct resource *res;
	33	res = kzalloc(sizeof(struct resource), GFP_KERNEL);
	34	BUG_ON(!res);
	35
	36	res->name = "System RAM";
	37	res->start = start;
	38	res->end = start + size - 1;
	39	res->flags = IORESOURCE_MEM;
	40	if (request_resource(&iomem_resource, res) < 0) {
	41	printk("System RAM resource %llx - %llx cannot be added\n",
	42	(unsigned long long)res->start, (unsigned long long)res->end);
	43	kfree(res);
	44	res = NULL;
	45	}
	46	return res;
	47	}
	48
	49	static void release_memory_resource(struct resource *res)
	50	{
	51	if (!res)
	52	return;
	53	release_resource(res);
	54	kfree(res);
	55	return;
	56	}
	57
	58
53947027	59	#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
718127cc	60	static int __add_zone(struct zone *zone, unsigned long phys_start_pfn)
3947be19 DH	61	{
	62	struct pglist_data *pgdat = zone->zone_pgdat;
	63	int nr_pages = PAGES_PER_SECTION;
	64	int nid = pgdat->node_id;
	65	int zone_type;
	66
	67	zone_type = zone - pgdat->node_zones;
13466c84	68	if (!zone->wait_table) {
718127cc	69	int ret = 0;
a2f3aa02 DH	70	ret = init_currently_empty_zone(zone, phys_start_pfn,
a2f3aa02 DH	71	nr_pages, MEMMAP_HOTPLUG);
718127cc YG	72	if (ret < 0)
	73	return ret;
	74	}
a2f3aa02 DH	75	memmap_init_zone(nr_pages, nid, zone_type,
a2f3aa02 DH	76	phys_start_pfn, MEMMAP_HOTPLUG);
718127cc	77	return 0;
3947be19 DH	78	}
3947be19 DH	79
3947be19 DH	80	static int __add_section(struct zone *zone, unsigned long phys_start_pfn)
3947be19 DH	81	{
3947be19	82	int nr_pages = PAGES_PER_SECTION;
3947be19 DH	83	int ret;
3947be19 DH	84
ebd15302 KH	85	if (pfn_valid(phys_start_pfn))
	86	return -EEXIST;
	87
0b0acbec	88	ret = sparse_add_one_section(zone, phys_start_pfn, nr_pages);
3947be19 DH	89
	90	if (ret < 0)
	91	return ret;
	92
718127cc YG	93	ret = __add_zone(zone, phys_start_pfn);
	94
	95	if (ret < 0)
	96	return ret;
	97
3947be19 DH	98	return register_new_memory(__pfn_to_section(phys_start_pfn));
	99	}
	100
	101	/*
	102	* Reasonably generic function for adding memory. It is
	103	* expected that archs that support memory hotplug will
	104	* call this function after deciding the zone to which to
	105	* add the new pages.
	106	*/
	107	int __add_pages(struct zone *zone, unsigned long phys_start_pfn,
	108	unsigned long nr_pages)
	109	{
	110	unsigned long i;
	111	int err = 0;
6f712711 KH	112	int start_sec, end_sec;
	113	/* during initialize mem_map, align hot-added range to section */
	114	start_sec = pfn_to_section_nr(phys_start_pfn);
	115	end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
3947be19	116
6f712711 KH	117	for (i = start_sec; i <= end_sec; i++) {
6f712711 KH	118	err = __add_section(zone, i << PFN_SECTION_SHIFT);
3947be19	119
6f712711 KH	120	/*
	121	* EEXIST is finally dealed with by ioresource collision
	122	* check. see add_memory() => register_memory_resource()
	123	* Warning will be printed if there is collision.
bed120c6 JS	124	*/
bed120c6 JS	125	if (err && (err != -EEXIST))
3947be19	126	break;
6f712711	127	err = 0;
3947be19 DH	128	}
	129
	130	return err;
	131	}
bed120c6	132	EXPORT_SYMBOL_GPL(__add_pages);
3947be19 DH	133
	134	static void grow_zone_span(struct zone *zone,
	135	unsigned long start_pfn, unsigned long end_pfn)
	136	{
	137	unsigned long old_zone_end_pfn;
	138
	139	zone_span_writelock(zone);
	140
	141	old_zone_end_pfn = zone->zone_start_pfn + zone->spanned_pages;
	142	if (start_pfn < zone->zone_start_pfn)
	143	zone->zone_start_pfn = start_pfn;
	144
25a6df95 YG	145	zone->spanned_pages = max(old_zone_end_pfn, end_pfn) -
25a6df95 YG	146	zone->zone_start_pfn;
3947be19 DH	147
	148	zone_span_writeunlock(zone);
	149	}
	150
	151	static void grow_pgdat_span(struct pglist_data *pgdat,
	152	unsigned long start_pfn, unsigned long end_pfn)
	153	{
	154	unsigned long old_pgdat_end_pfn =
	155	pgdat->node_start_pfn + pgdat->node_spanned_pages;
	156
	157	if (start_pfn < pgdat->node_start_pfn)
	158	pgdat->node_start_pfn = start_pfn;
	159
25a6df95 YG	160	pgdat->node_spanned_pages = max(old_pgdat_end_pfn, end_pfn) -
25a6df95 YG	161	pgdat->node_start_pfn;
3947be19 DH	162	}
	163
	164	int online_pages(unsigned long pfn, unsigned long nr_pages)
	165	{
	166	unsigned long i;
	167	unsigned long flags;
	168	unsigned long onlined_pages = 0;
2842f114 KH	169	struct resource res;
	170	u64 section_end;
	171	unsigned long start_pfn;
3947be19	172	struct zone *zone;
6811378e	173	int need_zonelists_rebuild = 0;
3947be19 DH	174
	175	/*
	176	* This doesn't need a lock to do pfn_to_page().
	177	* The section can't be removed here because of the
	178	* memory_block->state_sem.
	179	*/
	180	zone = page_zone(pfn_to_page(pfn));
	181	pgdat_resize_lock(zone->zone_pgdat, &flags);
	182	grow_zone_span(zone, pfn, pfn + nr_pages);
	183	grow_pgdat_span(zone->zone_pgdat, pfn, pfn + nr_pages);
	184	pgdat_resize_unlock(zone->zone_pgdat, &flags);
	185
6811378e YG	186	/*
	187	* If this zone is not populated, then it is not in zonelist.
	188	* This means the page allocator ignores this zone.
	189	* So, zonelist must be updated after online.
	190	*/
	191	if (!populated_zone(zone))
	192	need_zonelists_rebuild = 1;
	193
2842f114 KH	194	res.start = (u64)pfn << PAGE_SHIFT;
	195	res.end = res.start + ((u64)nr_pages << PAGE_SHIFT) - 1;
	196	res.flags = IORESOURCE_MEM; /* we just need system ram */
	197	section_end = res.end;
	198
58c1b5b0	199	while ((res.start < res.end) && (find_next_system_ram(&res) >= 0)) {
2842f114 KH	200	start_pfn = (unsigned long)(res.start >> PAGE_SHIFT);
	201	nr_pages = (unsigned long)
	202	((res.end + 1 - res.start) >> PAGE_SHIFT);
	203
	204	if (PageReserved(pfn_to_page(start_pfn))) {
	205	/* this region's page is not onlined now */
	206	for (i = 0; i < nr_pages; i++) {
	207	struct page *page = pfn_to_page(start_pfn + i);
	208	online_page(page);
	209	onlined_pages++;
	210	}
	211	}
	212
	213	res.start = res.end + 1;
	214	res.end = section_end;
3947be19 DH	215	}
3947be19 DH	216	zone->present_pages += onlined_pages;
f2937be5	217	zone->zone_pgdat->node_present_pages += onlined_pages;
3947be19	218
61b13993 DH	219	setup_per_zone_pages_min();
61b13993 DH	220
6811378e YG	221	if (need_zonelists_rebuild)
6811378e YG	222	build_all_zonelists();
5a4d4361	223	vm_total_pages = nr_free_pagecache_pages();
2d1d43f6	224	writeback_set_ratelimit();
3947be19 DH	225	return 0;
3947be19 DH	226	}
53947027	227	#endif /* CONFIG_MEMORY_HOTPLUG_SPARSE */
bc02af93	228
9af3c2de YG	229	static pg_data_t *hotadd_new_pgdat(int nid, u64 start)
	230	{
	231	struct pglist_data *pgdat;
	232	unsigned long zones_size[MAX_NR_ZONES] = {0};
	233	unsigned long zholes_size[MAX_NR_ZONES] = {0};
	234	unsigned long start_pfn = start >> PAGE_SHIFT;
	235
	236	pgdat = arch_alloc_nodedata(nid);
	237	if (!pgdat)
	238	return NULL;
	239
	240	arch_refresh_nodedata(nid, pgdat);
	241
	242	/* we can use NODE_DATA(nid) from here */
	243
	244	/* init node's zones as empty zones, we don't have any present pages.*/
	245	free_area_init_node(nid, pgdat, zones_size, start_pfn, zholes_size);
	246
	247	return pgdat;
	248	}
	249
	250	static void rollback_node_hotadd(int nid, pg_data_t *pgdat)
	251	{
	252	arch_refresh_nodedata(nid, NULL);
	253	arch_free_nodedata(pgdat);
	254	return;
	255	}
	256
0a547039	257
bc02af93 YG	258	int add_memory(int nid, u64 start, u64 size)
bc02af93 YG	259	{
9af3c2de YG	260	pg_data_t *pgdat = NULL;
9af3c2de YG	261	int new_pgdat = 0;
ebd15302	262	struct resource *res;
bc02af93 YG	263	int ret;
bc02af93 YG	264
ebd15302 KH	265	res = register_memory_resource(start, size);
	266	if (!res)
	267	return -EEXIST;
	268
9af3c2de YG	269	if (!node_online(nid)) {
	270	pgdat = hotadd_new_pgdat(nid, start);
	271	if (!pgdat)
	272	return -ENOMEM;
	273	new_pgdat = 1;
	274	ret = kswapd_run(nid);
	275	if (ret)
	276	goto error;
	277	}
	278
bc02af93 YG	279	/* call arch's memory hotadd */
	280	ret = arch_add_memory(nid, start, size);
	281
9af3c2de YG	282	if (ret < 0)
	283	goto error;
	284
0fc44159	285	/* we online node here. we can't roll back from here. */
9af3c2de YG	286	node_set_online(nid);
9af3c2de YG	287
38837fc7 PJ	288	cpuset_track_online_nodes();
38837fc7 PJ	289
0fc44159 YG	290	if (new_pgdat) {
	291	ret = register_one_node(nid);
	292	/*
	293	* If sysfs file of new node can't create, cpu on the node
	294	* can't be hot-added. There is no rollback way now.
	295	* So, check by BUG_ON() to catch it reluctantly..
	296	*/
	297	BUG_ON(ret);
	298	}
	299
9af3c2de YG	300	return ret;
	301	error:
	302	/* rollback pgdat allocation and others */
	303	if (new_pgdat)
	304	rollback_node_hotadd(nid, pgdat);
ebd15302 KH	305	if (res)
ebd15302 KH	306	release_memory_resource(res);
9af3c2de	307
bc02af93 YG	308	return ret;
	309	}
	310	EXPORT_SYMBOL_GPL(add_memory);