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