[linux-2.6-block.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>

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 (!populated_zone(zone)) {
		int ret = 0;
		ret = init_currently_empty_zone(zone, phys_start_pfn, nr_pages);
		if (ret < 0)
			return ret;
	}
	memmap_init_zone(nr_pages, nid, zone_type, phys_start_pfn);
	zonetable_add(zone, nid, zone_type, phys_start_pfn, nr_pages);
	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;
}

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;
}

/* 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;
}


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