1/* SPDX-License-Identifier: GPL-2.0 */
2#ifndef __LINUX_MEMORY_HOTPLUG_H
3#define __LINUX_MEMORY_HOTPLUG_H
4
5#include <linux/mmzone.h>
6#include <linux/spinlock.h>
7#include <linux/notifier.h>
8#include <linux/bug.h>
9
10struct page;
11struct zone;
12struct pglist_data;
13struct mem_section;
14struct memory_group;
15struct resource;
16struct vmem_altmap;
17struct dev_pagemap;
18
19#ifdef CONFIG_HAVE_ARCH_NODEDATA_EXTENSION
20/*
21 * For supporting node-hotadd, we have to allocate a new pgdat.
22 *
23 * If an arch has generic style NODE_DATA(),
24 * node_data[nid] = kzalloc() works well. But it depends on the architecture.
25 *
26 * In general, generic_alloc_nodedata() is used.
27 *
28 */
29extern pg_data_t *arch_alloc_nodedata(int nid);
30extern void arch_refresh_nodedata(int nid, pg_data_t *pgdat);
31
32#else /* CONFIG_HAVE_ARCH_NODEDATA_EXTENSION */
33
34#define arch_alloc_nodedata(nid)	generic_alloc_nodedata(nid)
35
36#ifdef CONFIG_NUMA
37/*
38 * XXX: node aware allocation can't work well to get new node's memory at this time.
39 *	Because, pgdat for the new node is not allocated/initialized yet itself.
40 *	To use new node's memory, more consideration will be necessary.
41 */
42#define generic_alloc_nodedata(nid)				\
43({								\
44	memblock_alloc(sizeof(*pgdat), SMP_CACHE_BYTES);	\
45})
46
47extern pg_data_t *node_data[];
48static inline void arch_refresh_nodedata(int nid, pg_data_t *pgdat)
49{
50	node_data[nid] = pgdat;
51}
52
53#else /* !CONFIG_NUMA */
54
55/* never called */
56static inline pg_data_t *generic_alloc_nodedata(int nid)
57{
58	BUG();
59	return NULL;
60}
61static inline void arch_refresh_nodedata(int nid, pg_data_t *pgdat)
62{
63}
64#endif /* CONFIG_NUMA */
65#endif /* CONFIG_HAVE_ARCH_NODEDATA_EXTENSION */
66
67#ifdef CONFIG_MEMORY_HOTPLUG
68struct page *pfn_to_online_page(unsigned long pfn);
69
70/* Types for control the zone type of onlined and offlined memory */
71enum {
72	/* Offline the memory. */
73	MMOP_OFFLINE = 0,
74	/* Online the memory. Zone depends, see default_zone_for_pfn(). */
75	MMOP_ONLINE,
76	/* Online the memory to ZONE_NORMAL. */
77	MMOP_ONLINE_KERNEL,
78	/* Online the memory to ZONE_MOVABLE. */
79	MMOP_ONLINE_MOVABLE,
80};
81
82/* Flags for add_memory() and friends to specify memory hotplug details. */
83typedef int __bitwise mhp_t;
84
85/* No special request */
86#define MHP_NONE		((__force mhp_t)0)
87/*
88 * Allow merging of the added System RAM resource with adjacent,
89 * mergeable resources. After a successful call to add_memory_resource()
90 * with this flag set, the resource pointer must no longer be used as it
91 * might be stale, or the resource might have changed.
92 */
93#define MHP_MERGE_RESOURCE	((__force mhp_t)BIT(0))
94
95/*
96 * We want memmap (struct page array) to be self contained.
97 * To do so, we will use the beginning of the hot-added range to build
98 * the page tables for the memmap array that describes the entire range.
99 * Only selected architectures support it with SPARSE_VMEMMAP.
100 * This is only a hint, the core kernel can decide to not do this based on
101 * different alignment checks.
102 */
103#define MHP_MEMMAP_ON_MEMORY   ((__force mhp_t)BIT(1))
104/*
105 * The nid field specifies a memory group id (mgid) instead. The memory group
106 * implies the node id (nid).
107 */
108#define MHP_NID_IS_MGID		((__force mhp_t)BIT(2))
109/*
110 * The hotplugged memory is completely inaccessible while the memory is
111 * offline. The memory provider will handle MEM_PREPARE_ONLINE /
112 * MEM_FINISH_OFFLINE notifications and make the memory accessible.
113 *
114 * This flag is only relevant when used along with MHP_MEMMAP_ON_MEMORY,
115 * because the altmap cannot be written (e.g., poisoned) when adding
116 * memory -- before it is set online.
117 *
118 * This allows for adding memory with an altmap that is not currently
119 * made available by a hypervisor. When onlining that memory, the
120 * hypervisor can be instructed to make that memory available, and
121 * the onlining phase will not require any memory allocations, which is
122 * helpful in low-memory situations.
123 */
124#define MHP_OFFLINE_INACCESSIBLE	((__force mhp_t)BIT(3))
125
126/*
127 * Extended parameters for memory hotplug:
128 * altmap: alternative allocator for memmap array (optional)
129 * pgprot: page protection flags to apply to newly created page tables
130 *	(required)
131 */
132struct mhp_params {
133	struct vmem_altmap *altmap;
134	pgprot_t pgprot;
135	struct dev_pagemap *pgmap;
136};
137
138bool mhp_range_allowed(u64 start, u64 size, bool need_mapping);
139struct range mhp_get_pluggable_range(bool need_mapping);
140bool mhp_supports_memmap_on_memory(void);
141
142/*
143 * Zone resizing functions
144 *
145 * Note: any attempt to resize a zone should has pgdat_resize_lock()
146 * zone_span_writelock() both held. This ensure the size of a zone
147 * can't be changed while pgdat_resize_lock() held.
148 */
149static inline unsigned zone_span_seqbegin(struct zone *zone)
150{
151	return read_seqbegin(&zone->span_seqlock);
152}
153static inline int zone_span_seqretry(struct zone *zone, unsigned iv)
154{
155	return read_seqretry(&zone->span_seqlock, iv);
156}
157static inline void zone_span_writelock(struct zone *zone)
158{
159	write_seqlock(&zone->span_seqlock);
160}
161static inline void zone_span_writeunlock(struct zone *zone)
162{
163	write_sequnlock(&zone->span_seqlock);
164}
165static inline void zone_seqlock_init(struct zone *zone)
166{
167	seqlock_init(&zone->span_seqlock);
168}
169extern void adjust_present_page_count(struct page *page,
170				      struct memory_group *group,
171				      long nr_pages);
172/* VM interface that may be used by firmware interface */
173extern int mhp_init_memmap_on_memory(unsigned long pfn, unsigned long nr_pages,
174				     struct zone *zone, bool mhp_off_inaccessible);
175extern void mhp_deinit_memmap_on_memory(unsigned long pfn, unsigned long nr_pages);
176extern int online_pages(unsigned long pfn, unsigned long nr_pages,
177			struct zone *zone, struct memory_group *group);
178extern void __offline_isolated_pages(unsigned long start_pfn,
179				     unsigned long end_pfn);
180
181typedef void (*online_page_callback_t)(struct page *page, unsigned int order);
182
183extern void generic_online_page(struct page *page, unsigned int order);
184extern int set_online_page_callback(online_page_callback_t callback);
185extern int restore_online_page_callback(online_page_callback_t callback);
186
187extern int try_online_node(int nid);
188
189extern int arch_add_memory(int nid, u64 start, u64 size,
190			   struct mhp_params *params);
191extern u64 max_mem_size;
192
193extern int mhp_online_type_from_str(const char *str);
194
195/* Default online_type (MMOP_*) when new memory blocks are added. */
196extern int mhp_default_online_type;
197/* If movable_node boot option specified */
198extern bool movable_node_enabled;
199static inline bool movable_node_is_enabled(void)
200{
201	return movable_node_enabled;
202}
203
204extern void arch_remove_memory(u64 start, u64 size, struct vmem_altmap *altmap);
205extern void __remove_pages(unsigned long start_pfn, unsigned long nr_pages,
206			   struct vmem_altmap *altmap);
207
208/* reasonably generic interface to expand the physical pages */
209extern int __add_pages(int nid, unsigned long start_pfn, unsigned long nr_pages,
210		       struct mhp_params *params);
211
212#ifndef CONFIG_ARCH_HAS_ADD_PAGES
213static inline int add_pages(int nid, unsigned long start_pfn,
214		unsigned long nr_pages, struct mhp_params *params)
215{
216	return __add_pages(nid, start_pfn, nr_pages, params);
217}
218#else /* ARCH_HAS_ADD_PAGES */
219int add_pages(int nid, unsigned long start_pfn, unsigned long nr_pages,
220	      struct mhp_params *params);
221#endif /* ARCH_HAS_ADD_PAGES */
222
223void get_online_mems(void);
224void put_online_mems(void);
225
226void mem_hotplug_begin(void);
227void mem_hotplug_done(void);
228
229/* See kswapd_is_running() */
230static inline void pgdat_kswapd_lock(pg_data_t *pgdat)
231{
232	mutex_lock(&pgdat->kswapd_lock);
233}
234
235static inline void pgdat_kswapd_unlock(pg_data_t *pgdat)
236{
237	mutex_unlock(&pgdat->kswapd_lock);
238}
239
240static inline void pgdat_kswapd_lock_init(pg_data_t *pgdat)
241{
242	mutex_init(&pgdat->kswapd_lock);
243}
244
245#else /* ! CONFIG_MEMORY_HOTPLUG */
246#define pfn_to_online_page(pfn)			\
247({						\
248	struct page *___page = NULL;		\
249	if (pfn_valid(pfn))			\
250		___page = pfn_to_page(pfn);	\
251	___page;				\
252 })
253
254static inline unsigned zone_span_seqbegin(struct zone *zone)
255{
256	return 0;
257}
258static inline int zone_span_seqretry(struct zone *zone, unsigned iv)
259{
260	return 0;
261}
262static inline void zone_span_writelock(struct zone *zone) {}
263static inline void zone_span_writeunlock(struct zone *zone) {}
264static inline void zone_seqlock_init(struct zone *zone) {}
265
266static inline int try_online_node(int nid)
267{
268	return 0;
269}
270
271static inline void get_online_mems(void) {}
272static inline void put_online_mems(void) {}
273
274static inline void mem_hotplug_begin(void) {}
275static inline void mem_hotplug_done(void) {}
276
277static inline bool movable_node_is_enabled(void)
278{
279	return false;
280}
281
282static inline bool mhp_supports_memmap_on_memory(void)
283{
284	return false;
285}
286
287static inline void pgdat_kswapd_lock(pg_data_t *pgdat) {}
288static inline void pgdat_kswapd_unlock(pg_data_t *pgdat) {}
289static inline void pgdat_kswapd_lock_init(pg_data_t *pgdat) {}
290#endif /* ! CONFIG_MEMORY_HOTPLUG */
291
292/*
293 * Keep this declaration outside CONFIG_MEMORY_HOTPLUG as some
294 * platforms might override and use arch_get_mappable_range()
295 * for internal non memory hotplug purposes.
296 */
297struct range arch_get_mappable_range(void);
298
299#if defined(CONFIG_MEMORY_HOTPLUG) || defined(CONFIG_DEFERRED_STRUCT_PAGE_INIT)
300/*
301 * pgdat resizing functions
302 */
303static inline
304void pgdat_resize_lock(struct pglist_data *pgdat, unsigned long *flags)
305{
306	spin_lock_irqsave(&pgdat->node_size_lock, *flags);
307}
308static inline
309void pgdat_resize_unlock(struct pglist_data *pgdat, unsigned long *flags)
310{
311	spin_unlock_irqrestore(&pgdat->node_size_lock, *flags);
312}
313static inline
314void pgdat_resize_init(struct pglist_data *pgdat)
315{
316	spin_lock_init(&pgdat->node_size_lock);
317}
318#else /* !(CONFIG_MEMORY_HOTPLUG || CONFIG_DEFERRED_STRUCT_PAGE_INIT) */
319/*
320 * Stub functions for when hotplug is off
321 */
322static inline void pgdat_resize_lock(struct pglist_data *p, unsigned long *f) {}
323static inline void pgdat_resize_unlock(struct pglist_data *p, unsigned long *f) {}
324static inline void pgdat_resize_init(struct pglist_data *pgdat) {}
325#endif /* !(CONFIG_MEMORY_HOTPLUG || CONFIG_DEFERRED_STRUCT_PAGE_INIT) */
326
327#ifdef CONFIG_MEMORY_HOTREMOVE
328
329extern void try_offline_node(int nid);
330extern int offline_pages(unsigned long start_pfn, unsigned long nr_pages,
331			 struct zone *zone, struct memory_group *group);
332extern int remove_memory(u64 start, u64 size);
333extern void __remove_memory(u64 start, u64 size);
334extern int offline_and_remove_memory(u64 start, u64 size);
335
336#else
337static inline void try_offline_node(int nid) {}
338
339static inline int offline_pages(unsigned long start_pfn, unsigned long nr_pages,
340				struct zone *zone, struct memory_group *group)
341{
342	return -EINVAL;
343}
344
345static inline int remove_memory(u64 start, u64 size)
346{
347	return -EBUSY;
348}
349
350static inline void __remove_memory(u64 start, u64 size) {}
351#endif /* CONFIG_MEMORY_HOTREMOVE */
352
353#ifdef CONFIG_MEMORY_HOTPLUG
354extern void __ref free_area_init_core_hotplug(struct pglist_data *pgdat);
355extern int __add_memory(int nid, u64 start, u64 size, mhp_t mhp_flags);
356extern int add_memory(int nid, u64 start, u64 size, mhp_t mhp_flags);
357extern int add_memory_resource(int nid, struct resource *resource,
358			       mhp_t mhp_flags);
359extern int add_memory_driver_managed(int nid, u64 start, u64 size,
360				     const char *resource_name,
361				     mhp_t mhp_flags);
362extern void move_pfn_range_to_zone(struct zone *zone, unsigned long start_pfn,
363				   unsigned long nr_pages,
364				   struct vmem_altmap *altmap, int migratetype);
365extern void remove_pfn_range_from_zone(struct zone *zone,
366				       unsigned long start_pfn,
367				       unsigned long nr_pages);
368extern int sparse_add_section(int nid, unsigned long pfn,
369		unsigned long nr_pages, struct vmem_altmap *altmap,
370		struct dev_pagemap *pgmap);
371extern void sparse_remove_section(unsigned long pfn, unsigned long nr_pages,
372				  struct vmem_altmap *altmap);
373extern struct page *sparse_decode_mem_map(unsigned long coded_mem_map,
374					  unsigned long pnum);
375extern struct zone *zone_for_pfn_range(int online_type, int nid,
376		struct memory_group *group, unsigned long start_pfn,
377		unsigned long nr_pages);
378extern int arch_create_linear_mapping(int nid, u64 start, u64 size,
379				      struct mhp_params *params);
380void arch_remove_linear_mapping(u64 start, u64 size);
381#endif /* CONFIG_MEMORY_HOTPLUG */
382
383#endif /* __LINUX_MEMORY_HOTPLUG_H */
384