4 #include <linux/spinlock.h>
5 #include <linux/linkage.h>
6 #include <linux/mmzone.h>
7 #include <linux/list.h>
8 #include <linux/memcontrol.h>
9 #include <linux/sched.h>
10 #include <linux/node.h>
12 #include <linux/atomic.h>
15 struct notifier_block
;
19 #define SWAP_FLAG_PREFER 0x8000 /* set if swap priority specified */
20 #define SWAP_FLAG_PRIO_MASK 0x7fff
21 #define SWAP_FLAG_PRIO_SHIFT 0
22 #define SWAP_FLAG_DISCARD 0x10000 /* discard swap cluster after use */
24 static inline int current_is_kswapd(void)
26 return current
->flags
& PF_KSWAPD
;
30 * MAX_SWAPFILES defines the maximum number of swaptypes: things which can
31 * be swapped to. The swap type and the offset into that swap type are
32 * encoded into pte's and into pgoff_t's in the swapcache. Using five bits
33 * for the type means that the maximum number of swapcache pages is 27 bits
34 * on 32-bit-pgoff_t architectures. And that assumes that the architecture packs
35 * the type/offset into the pte as 5/27 as well.
37 #define MAX_SWAPFILES_SHIFT 5
40 * Use some of the swap files numbers for other purposes. This
41 * is a convenient way to hook into the VM to trigger special
46 * NUMA node memory migration support
48 #ifdef CONFIG_MIGRATION
49 #define SWP_MIGRATION_NUM 2
50 #define SWP_MIGRATION_READ (MAX_SWAPFILES + SWP_HWPOISON_NUM)
51 #define SWP_MIGRATION_WRITE (MAX_SWAPFILES + SWP_HWPOISON_NUM + 1)
53 #define SWP_MIGRATION_NUM 0
57 * Handling of hardware poisoned pages with memory corruption.
59 #ifdef CONFIG_MEMORY_FAILURE
60 #define SWP_HWPOISON_NUM 1
61 #define SWP_HWPOISON MAX_SWAPFILES
63 #define SWP_HWPOISON_NUM 0
66 #define MAX_SWAPFILES \
67 ((1 << MAX_SWAPFILES_SHIFT) - SWP_MIGRATION_NUM - SWP_HWPOISON_NUM)
70 * Magic header for a swap area. The first part of the union is
71 * what the swap magic looks like for the old (limited to 128MB)
72 * swap area format, the second part of the union adds - in the
73 * old reserved area - some extra information. Note that the first
74 * kilobyte is reserved for boot loader or disk label stuff...
76 * Having the magic at the end of the PAGE_SIZE makes detecting swap
77 * areas somewhat tricky on machines that support multiple page sizes.
78 * For 2.5 we'll probably want to move the magic to just beyond the
83 char reserved
[PAGE_SIZE
- 10];
84 char magic
[10]; /* SWAP-SPACE or SWAPSPACE2 */
87 char bootbits
[1024]; /* Space for disklabel etc. */
91 unsigned char sws_uuid
[16];
92 unsigned char sws_volume
[16];
98 /* A swap entry has to fit into a "unsigned long", as
99 * the entry is hidden in the "index" field of the
100 * swapper address space.
107 * current->reclaim_state points to one of these when a task is running
110 struct reclaim_state
{
111 unsigned long reclaimed_slab
;
116 struct address_space
;
118 struct writeback_control
;
122 * A swap extent maps a range of a swapfile's PAGE_SIZE pages onto a range of
123 * disk blocks. A list of swap extents maps the entire swapfile. (Where the
124 * term `swapfile' refers to either a blockdevice or an IS_REG file. Apart
125 * from setup, they're handled identically.
127 * We always assume that blocks are of size PAGE_SIZE.
130 struct list_head list
;
133 sector_t start_block
;
137 * Max bad pages in the new format..
139 #define __swapoffset(x) ((unsigned long)&((union swap_header *)0)->x)
140 #define MAX_SWAP_BADPAGES \
141 ((__swapoffset(magic.magic) - __swapoffset(info.badpages)) / sizeof(int))
144 SWP_USED
= (1 << 0), /* is slot in swap_info[] used? */
145 SWP_WRITEOK
= (1 << 1), /* ok to write to this swap? */
146 SWP_DISCARDABLE
= (1 << 2), /* swapon+blkdev support discard */
147 SWP_DISCARDING
= (1 << 3), /* now discarding a free cluster */
148 SWP_SOLIDSTATE
= (1 << 4), /* blkdev seeks are cheap */
149 SWP_CONTINUED
= (1 << 5), /* swap_map has count continuation */
150 SWP_BLKDEV
= (1 << 6), /* its a block device */
151 /* add others here before... */
152 SWP_SCANNING
= (1 << 8), /* refcount in scan_swap_map */
155 #define SWAP_CLUSTER_MAX 32
156 #define COMPACT_CLUSTER_MAX SWAP_CLUSTER_MAX
159 * Ratio between the present memory in the zone and the "gap" that
160 * we're allowing kswapd to shrink in addition to the per-zone high
161 * wmark, even for zones that already have the high wmark satisfied,
162 * in order to provide better per-zone lru behavior. We are ok to
163 * spend not more than 1% of the memory for this zone balancing "gap".
165 #define KSWAPD_ZONE_BALANCE_GAP_RATIO 100
167 #define SWAP_MAP_MAX 0x3e /* Max duplication count, in first swap_map */
168 #define SWAP_MAP_BAD 0x3f /* Note pageblock is bad, in first swap_map */
169 #define SWAP_HAS_CACHE 0x40 /* Flag page is cached, in first swap_map */
170 #define SWAP_CONT_MAX 0x7f /* Max count, in each swap_map continuation */
171 #define COUNT_CONTINUED 0x80 /* See swap_map continuation for full count */
172 #define SWAP_MAP_SHMEM 0xbf /* Owned by shmem/tmpfs, in first swap_map */
175 * The in-memory structure used to track swap areas.
177 struct swap_info_struct
{
178 unsigned long flags
; /* SWP_USED etc: see above */
179 signed short prio
; /* swap priority of this type */
180 signed char type
; /* strange name for an index */
181 signed char next
; /* next type on the swap list */
182 unsigned int max
; /* extent of the swap_map */
183 unsigned char *swap_map
; /* vmalloc'ed array of usage counts */
184 unsigned int lowest_bit
; /* index of first free in swap_map */
185 unsigned int highest_bit
; /* index of last free in swap_map */
186 unsigned int pages
; /* total of usable pages of swap */
187 unsigned int inuse_pages
; /* number of those currently in use */
188 unsigned int cluster_next
; /* likely index for next allocation */
189 unsigned int cluster_nr
; /* countdown to next cluster search */
190 unsigned int lowest_alloc
; /* while preparing discard cluster */
191 unsigned int highest_alloc
; /* while preparing discard cluster */
192 struct swap_extent
*curr_swap_extent
;
193 struct swap_extent first_swap_extent
;
194 struct block_device
*bdev
; /* swap device or bdev of swap file */
195 struct file
*swap_file
; /* seldom referenced */
196 unsigned int old_block_size
; /* seldom referenced */
200 int head
; /* head of priority-ordered swapfile list */
201 int next
; /* swapfile to be used next */
204 /* Swap 50% full? Release swapcache more aggressively.. */
205 #define vm_swap_full() (nr_swap_pages*2 < total_swap_pages)
207 /* linux/mm/page_alloc.c */
208 extern unsigned long totalram_pages
;
209 extern unsigned long totalreserve_pages
;
210 extern unsigned int nr_free_buffer_pages(void);
211 extern unsigned int nr_free_pagecache_pages(void);
213 /* Definition of global_page_state not available yet */
214 #define nr_free_pages() global_page_state(NR_FREE_PAGES)
217 /* linux/mm/swap.c */
218 extern void __lru_cache_add(struct page
*, enum lru_list lru
);
219 extern void lru_cache_add_lru(struct page
*, enum lru_list lru
);
220 extern void lru_add_page_tail(struct zone
* zone
,
221 struct page
*page
, struct page
*page_tail
);
222 extern void activate_page(struct page
*);
223 extern void mark_page_accessed(struct page
*);
224 extern void lru_add_drain(void);
225 extern int lru_add_drain_all(void);
226 extern void rotate_reclaimable_page(struct page
*page
);
227 extern void deactivate_page(struct page
*page
);
228 extern void swap_setup(void);
230 extern void add_page_to_unevictable_list(struct page
*page
);
233 * lru_cache_add: add a page to the page lists
234 * @page: the page to add
236 static inline void lru_cache_add_anon(struct page
*page
)
238 __lru_cache_add(page
, LRU_INACTIVE_ANON
);
241 static inline void lru_cache_add_file(struct page
*page
)
243 __lru_cache_add(page
, LRU_INACTIVE_FILE
);
246 /* LRU Isolation modes. */
247 #define ISOLATE_INACTIVE 0 /* Isolate inactive pages. */
248 #define ISOLATE_ACTIVE 1 /* Isolate active pages. */
249 #define ISOLATE_BOTH 2 /* Isolate both active and inactive pages. */
251 /* linux/mm/vmscan.c */
252 extern unsigned long try_to_free_pages(struct zonelist
*zonelist
, int order
,
253 gfp_t gfp_mask
, nodemask_t
*mask
);
254 extern int __isolate_lru_page(struct page
*page
, int mode
, int file
);
255 extern unsigned long shrink_all_memory(unsigned long nr_pages
);
256 extern int vm_swappiness
;
257 extern int remove_mapping(struct address_space
*mapping
, struct page
*page
);
258 extern long vm_total_pages
;
261 extern int zone_reclaim_mode
;
262 extern int sysctl_min_unmapped_ratio
;
263 extern int sysctl_min_slab_ratio
;
264 extern int zone_reclaim(struct zone
*, gfp_t
, unsigned int);
266 #define zone_reclaim_mode 0
267 static inline int zone_reclaim(struct zone
*z
, gfp_t mask
, unsigned int order
)
273 extern int page_evictable(struct page
*page
, struct vm_area_struct
*vma
);
274 extern void scan_mapping_unevictable_pages(struct address_space
*);
276 extern unsigned long scan_unevictable_pages
;
277 extern int scan_unevictable_handler(struct ctl_table
*, int,
278 void __user
*, size_t *, loff_t
*);
280 extern int scan_unevictable_register_node(struct node
*node
);
281 extern void scan_unevictable_unregister_node(struct node
*node
);
283 static inline int scan_unevictable_register_node(struct node
*node
)
287 static inline void scan_unevictable_unregister_node(struct node
*node
)
292 extern int kswapd_run(int nid
);
293 extern void kswapd_stop(int nid
);
294 #ifdef CONFIG_CGROUP_MEM_RES_CTLR
295 extern int mem_cgroup_swappiness(struct mem_cgroup
*mem
);
297 static inline int mem_cgroup_swappiness(struct mem_cgroup
*mem
)
299 return vm_swappiness
;
303 /* linux/mm/page_io.c */
304 extern int swap_readpage(struct page
*);
305 extern int swap_writepage(struct page
*page
, struct writeback_control
*wbc
);
306 extern void end_swap_bio_read(struct bio
*bio
, int err
);
308 /* linux/mm/swap_state.c */
309 extern struct address_space swapper_space
;
310 #define total_swapcache_pages swapper_space.nrpages
311 extern void show_swap_cache_info(void);
312 extern int add_to_swap(struct page
*);
313 extern int add_to_swap_cache(struct page
*, swp_entry_t
, gfp_t
);
314 extern void __delete_from_swap_cache(struct page
*);
315 extern void delete_from_swap_cache(struct page
*);
316 extern void free_page_and_swap_cache(struct page
*);
317 extern void free_pages_and_swap_cache(struct page
**, int);
318 extern struct page
*lookup_swap_cache(swp_entry_t
);
319 extern struct page
*read_swap_cache_async(swp_entry_t
, gfp_t
,
320 struct vm_area_struct
*vma
, unsigned long addr
);
321 extern struct page
*swapin_readahead(swp_entry_t
, gfp_t
,
322 struct vm_area_struct
*vma
, unsigned long addr
);
324 /* linux/mm/swapfile.c */
325 extern long nr_swap_pages
;
326 extern long total_swap_pages
;
327 extern void si_swapinfo(struct sysinfo
*);
328 extern swp_entry_t
get_swap_page(void);
329 extern swp_entry_t
get_swap_page_of_type(int);
330 extern int valid_swaphandles(swp_entry_t
, unsigned long *);
331 extern int add_swap_count_continuation(swp_entry_t
, gfp_t
);
332 extern void swap_shmem_alloc(swp_entry_t
);
333 extern int swap_duplicate(swp_entry_t
);
334 extern int swapcache_prepare(swp_entry_t
);
335 extern void swap_free(swp_entry_t
);
336 extern void swapcache_free(swp_entry_t
, struct page
*page
);
337 extern int free_swap_and_cache(swp_entry_t
);
338 extern int swap_type_of(dev_t
, sector_t
, struct block_device
**);
339 extern unsigned int count_swap_pages(int, int);
340 extern sector_t
map_swap_page(struct page
*, struct block_device
**);
341 extern sector_t
swapdev_block(int, pgoff_t
);
342 extern int reuse_swap_page(struct page
*);
343 extern int try_to_free_swap(struct page
*);
344 struct backing_dev_info
;
346 /* linux/mm/thrash.c */
347 extern struct mm_struct
*swap_token_mm
;
348 extern void grab_swap_token(struct mm_struct
*);
349 extern void __put_swap_token(struct mm_struct
*);
350 extern void disable_swap_token(struct mem_cgroup
*memcg
);
352 static inline int has_swap_token(struct mm_struct
*mm
)
354 return (mm
== swap_token_mm
);
357 static inline void put_swap_token(struct mm_struct
*mm
)
359 if (has_swap_token(mm
))
360 __put_swap_token(mm
);
363 #ifdef CONFIG_CGROUP_MEM_RES_CTLR
365 mem_cgroup_uncharge_swapcache(struct page
*page
, swp_entry_t ent
, bool swapout
);
366 extern int mem_cgroup_count_swap_user(swp_entry_t ent
, struct page
**pagep
);
369 mem_cgroup_uncharge_swapcache(struct page
*page
, swp_entry_t ent
, bool swapout
)
373 #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
374 extern void mem_cgroup_uncharge_swap(swp_entry_t ent
);
376 static inline void mem_cgroup_uncharge_swap(swp_entry_t ent
)
381 #else /* CONFIG_SWAP */
383 #define nr_swap_pages 0L
384 #define total_swap_pages 0L
385 #define total_swapcache_pages 0UL
387 #define si_swapinfo(val) \
388 do { (val)->freeswap = (val)->totalswap = 0; } while (0)
389 /* only sparc can not include linux/pagemap.h in this file
390 * so leave page_cache_release and release_pages undeclared... */
391 #define free_page_and_swap_cache(page) \
392 page_cache_release(page)
393 #define free_pages_and_swap_cache(pages, nr) \
394 release_pages((pages), (nr), 0);
396 static inline void show_swap_cache_info(void)
400 #define free_swap_and_cache(swp) is_migration_entry(swp)
401 #define swapcache_prepare(swp) is_migration_entry(swp)
403 static inline int add_swap_count_continuation(swp_entry_t swp
, gfp_t gfp_mask
)
408 static inline void swap_shmem_alloc(swp_entry_t swp
)
412 static inline int swap_duplicate(swp_entry_t swp
)
417 static inline void swap_free(swp_entry_t swp
)
421 static inline void swapcache_free(swp_entry_t swp
, struct page
*page
)
425 static inline struct page
*swapin_readahead(swp_entry_t swp
, gfp_t gfp_mask
,
426 struct vm_area_struct
*vma
, unsigned long addr
)
431 static inline int swap_writepage(struct page
*p
, struct writeback_control
*wbc
)
436 static inline struct page
*lookup_swap_cache(swp_entry_t swp
)
441 static inline int add_to_swap(struct page
*page
)
446 static inline int add_to_swap_cache(struct page
*page
, swp_entry_t entry
,
452 static inline void __delete_from_swap_cache(struct page
*page
)
456 static inline void delete_from_swap_cache(struct page
*page
)
460 #define reuse_swap_page(page) (page_mapcount(page) == 1)
462 static inline int try_to_free_swap(struct page
*page
)
467 static inline swp_entry_t
get_swap_page(void)
474 /* linux/mm/thrash.c */
475 static inline void put_swap_token(struct mm_struct
*mm
)
479 static inline void grab_swap_token(struct mm_struct
*mm
)
483 static inline int has_swap_token(struct mm_struct
*mm
)
488 static inline void disable_swap_token(struct mem_cgroup
*memcg
)
493 mem_cgroup_uncharge_swapcache(struct page
*page
, swp_entry_t ent
)
497 #ifdef CONFIG_CGROUP_MEM_RES_CTLR
499 mem_cgroup_count_swap_user(swp_entry_t ent
, struct page
**pagep
)
505 #endif /* CONFIG_SWAP */
506 #endif /* __KERNEL__*/
507 #endif /* _LINUX_SWAP_H */