include/linux/swap.h

   1 #ifndef _LINUX_SWAP_H
   2 #define _LINUX_SWAP_H
   3
   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>
  11 #include <linux/fs.h>
  12 #include <linux/atomic.h>
  13 #include <asm/page.h>
  14
  15 struct notifier_block;
  16
  17 struct bio;
  18
  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 */
  23
  24 #define SWAP_FLAGS_VALID        (SWAP_FLAG_PRIO_MASK | SWAP_FLAG_PREFER | \
  25                                  SWAP_FLAG_DISCARD)
  26
  27 static inline int current_is_kswapd(void)
  28 {
  29         return current->flags & PF_KSWAPD;
  30 }
  31
  32 /*
  33  * MAX_SWAPFILES defines the maximum number of swaptypes: things which can
  34  * be swapped to.  The swap type and the offset into that swap type are
  35  * encoded into pte's and into pgoff_t's in the swapcache.  Using five bits
  36  * for the type means that the maximum number of swapcache pages is 27 bits
  37  * on 32-bit-pgoff_t architectures.  And that assumes that the architecture packs
  38  * the type/offset into the pte as 5/27 as well.
  39  */
  40 #define MAX_SWAPFILES_SHIFT     5
  41
  42 /*
  43  * Use some of the swap files numbers for other purposes. This
  44  * is a convenient way to hook into the VM to trigger special
  45  * actions on faults.
  46  */
  47
  48 /*
  49  * NUMA node memory migration support
  50  */
  51 #ifdef CONFIG_MIGRATION
  52 #define SWP_MIGRATION_NUM 2
  53 #define SWP_MIGRATION_READ      (MAX_SWAPFILES + SWP_HWPOISON_NUM)
  54 #define SWP_MIGRATION_WRITE     (MAX_SWAPFILES + SWP_HWPOISON_NUM + 1)
  55 #else
  56 #define SWP_MIGRATION_NUM 0
  57 #endif
  58
  59 /*
  60  * Handling of hardware poisoned pages with memory corruption.
  61  */
  62 #ifdef CONFIG_MEMORY_FAILURE
  63 #define SWP_HWPOISON_NUM 1
  64 #define SWP_HWPOISON            MAX_SWAPFILES
  65 #else
  66 #define SWP_HWPOISON_NUM 0
  67 #endif
  68
  69 #define MAX_SWAPFILES \
  70         ((1 << MAX_SWAPFILES_SHIFT) - SWP_MIGRATION_NUM - SWP_HWPOISON_NUM)
  71
  72 /*
  73  * Magic header for a swap area. The first part of the union is
  74  * what the swap magic looks like for the old (limited to 128MB)
  75  * swap area format, the second part of the union adds - in the
  76  * old reserved area - some extra information. Note that the first
  77  * kilobyte is reserved for boot loader or disk label stuff...
  78  *
  79  * Having the magic at the end of the PAGE_SIZE makes detecting swap
  80  * areas somewhat tricky on machines that support multiple page sizes.
  81  * For 2.5 we'll probably want to move the magic to just beyond the
  82  * bootbits...
  83  */
  84 union swap_header {
  85         struct {
  86                 char reserved[PAGE_SIZE - 10];
  87                 char magic[10];                 /* SWAP-SPACE or SWAPSPACE2 */
  88         } magic;
  89         struct {
  90                 char            bootbits[1024]; /* Space for disklabel etc. */
  91                 __u32           version;
  92                 __u32           last_page;
  93                 __u32           nr_badpages;
  94                 unsigned char   sws_uuid[16];
  95                 unsigned char   sws_volume[16];
  96                 __u32           padding[117];
  97                 __u32           badpages[1];
  98         } info;
  99 };
 100
 101  /* A swap entry has to fit into a "unsigned long", as
 102   * the entry is hidden in the "index" field of the
 103   * swapper address space.
 104   */
 105 typedef struct {
 106         unsigned long val;
 107 } swp_entry_t;
 108
 109 /*
 110  * current->reclaim_state points to one of these when a task is running
 111  * memory reclaim
 112  */
 113 struct reclaim_state {
 114         unsigned long reclaimed_slab;
 115 };
 116
 117 #ifdef __KERNEL__
 118
 119 struct address_space;
 120 struct sysinfo;
 121 struct writeback_control;
 122 struct zone;
 123
 124 /*
 125  * A swap extent maps a range of a swapfile's PAGE_SIZE pages onto a range of
 126  * disk blocks.  A list of swap extents maps the entire swapfile.  (Where the
 127  * term `swapfile' refers to either a blockdevice or an IS_REG file.  Apart
 128  * from setup, they're handled identically.
 129  *
 130  * We always assume that blocks are of size PAGE_SIZE.
 131  */
 132 struct swap_extent {
 133         struct list_head list;
 134         pgoff_t start_page;
 135         pgoff_t nr_pages;
 136         sector_t start_block;
 137 };
 138
 139 /*
 140  * Max bad pages in the new format..
 141  */
 142 #define __swapoffset(x) ((unsigned long)&((union swap_header *)0)->x)
 143 #define MAX_SWAP_BADPAGES \
 144         ((__swapoffset(magic.magic) - __swapoffset(info.badpages)) / sizeof(int))
 145
 146 enum {
 147         SWP_USED        = (1 << 0),     /* is slot in swap_info[] used? */
 148         SWP_WRITEOK     = (1 << 1),     /* ok to write to this swap?    */
 149         SWP_DISCARDABLE = (1 << 2),     /* swapon+blkdev support discard */
 150         SWP_DISCARDING  = (1 << 3),     /* now discarding a free cluster */
 151         SWP_SOLIDSTATE  = (1 << 4),     /* blkdev seeks are cheap */
 152         SWP_CONTINUED   = (1 << 5),     /* swap_map has count continuation */
 153         SWP_BLKDEV      = (1 << 6),     /* its a block device */
 154         SWP_FILE        = (1 << 7),     /* set after swap_activate success */
 155                                         /* add others here before... */
 156         SWP_SCANNING    = (1 << 8),     /* refcount in scan_swap_map */
 157 };
 158
 159 #define SWAP_CLUSTER_MAX 32UL
 160 #define COMPACT_CLUSTER_MAX SWAP_CLUSTER_MAX
 161
 162 /*
 163  * Ratio between the present memory in the zone and the "gap" that
 164  * we're allowing kswapd to shrink in addition to the per-zone high
 165  * wmark, even for zones that already have the high wmark satisfied,
 166  * in order to provide better per-zone lru behavior. We are ok to
 167  * spend not more than 1% of the memory for this zone balancing "gap".
 168  */
 169 #define KSWAPD_ZONE_BALANCE_GAP_RATIO 100
 170
 171 #define SWAP_MAP_MAX    0x3e    /* Max duplication count, in first swap_map */
 172 #define SWAP_MAP_BAD    0x3f    /* Note pageblock is bad, in first swap_map */
 173 #define SWAP_HAS_CACHE  0x40    /* Flag page is cached, in first swap_map */
 174 #define SWAP_CONT_MAX   0x7f    /* Max count, in each swap_map continuation */
 175 #define COUNT_CONTINUED 0x80    /* See swap_map continuation for full count */
 176 #define SWAP_MAP_SHMEM  0xbf    /* Owned by shmem/tmpfs, in first swap_map */
 177
 178 /*
 179  * The in-memory structure used to track swap areas.
 180  */
 181 struct swap_info_struct {
 182         unsigned long   flags;          /* SWP_USED etc: see above */
 183         signed short    prio;           /* swap priority of this type */
 184         signed char     type;           /* strange name for an index */
 185         signed char     next;           /* next type on the swap list */
 186         unsigned int    max;            /* extent of the swap_map */
 187         unsigned char *swap_map;        /* vmalloc'ed array of usage counts */
 188         unsigned int lowest_bit;        /* index of first free in swap_map */
 189         unsigned int highest_bit;       /* index of last free in swap_map */
 190         unsigned int pages;             /* total of usable pages of swap */
 191         unsigned int inuse_pages;       /* number of those currently in use */
 192         unsigned int cluster_next;      /* likely index for next allocation */
 193         unsigned int cluster_nr;        /* countdown to next cluster search */
 194         unsigned int lowest_alloc;      /* while preparing discard cluster */
 195         unsigned int highest_alloc;     /* while preparing discard cluster */
 196         struct swap_extent *curr_swap_extent;
 197         struct swap_extent first_swap_extent;
 198         struct block_device *bdev;      /* swap device or bdev of swap file */
 199         struct file *swap_file;         /* seldom referenced */
 200         unsigned int old_block_size;    /* seldom referenced */
 201 #ifdef CONFIG_FRONTSWAP
 202         unsigned long *frontswap_map;   /* frontswap in-use, one bit per page */
 203         atomic_t frontswap_pages;       /* frontswap pages in-use counter */
 204 #endif
 205         spinlock_t lock;                /*
 206                                          * protect map scan related fields like
 207                                          * swap_map, lowest_bit, highest_bit,
 208                                          * inuse_pages, cluster_next,
 209                                          * cluster_nr, lowest_alloc and
 210                                          * highest_alloc. other fields are only
 211                                          * changed at swapon/swapoff, so are
 212                                          * protected by swap_lock. changing
 213                                          * flags need hold this lock and
 214                                          * swap_lock. If both locks need hold,
 215                                          * hold swap_lock first.
 216                                          */
 217 };
 218
 219 struct swap_list_t {
 220         int head;       /* head of priority-ordered swapfile list */
 221         int next;       /* swapfile to be used next */
 222 };
 223
 224 /* linux/mm/page_alloc.c */
 225 extern unsigned long totalram_pages;
 226 extern unsigned long totalreserve_pages;
 227 extern unsigned long dirty_balance_reserve;
 228 extern unsigned long nr_free_buffer_pages(void);
 229 extern unsigned long nr_free_pagecache_pages(void);
 230
 231 /* Definition of global_page_state not available yet */
 232 #define nr_free_pages() global_page_state(NR_FREE_PAGES)
 233
 234
 235 /* linux/mm/swap.c */
 236 extern void __lru_cache_add(struct page *, enum lru_list lru);
 237 extern void lru_cache_add_lru(struct page *, enum lru_list lru);
 238 extern void lru_add_page_tail(struct page *page, struct page *page_tail,
 239                          struct lruvec *lruvec, struct list_head *head);
 240 extern void activate_page(struct page *);
 241 extern void mark_page_accessed(struct page *);
 242 extern void lru_add_drain(void);
 243 extern void lru_add_drain_cpu(int cpu);
 244 extern int lru_add_drain_all(void);
 245 extern void rotate_reclaimable_page(struct page *page);
 246 extern void deactivate_page(struct page *page);
 247 extern void swap_setup(void);
 248
 249 extern void add_page_to_unevictable_list(struct page *page);
 250
 251 /**
 252  * lru_cache_add: add a page to the page lists
 253  * @page: the page to add
 254  */
 255 static inline void lru_cache_add_anon(struct page *page)
 256 {
 257         __lru_cache_add(page, LRU_INACTIVE_ANON);
 258 }
 259
 260 static inline void lru_cache_add_file(struct page *page)
 261 {
 262         __lru_cache_add(page, LRU_INACTIVE_FILE);
 263 }
 264
 265 /* linux/mm/vmscan.c */
 266 extern unsigned long try_to_free_pages(struct zonelist *zonelist, int order,
 267                                         gfp_t gfp_mask, nodemask_t *mask);
 268 extern int __isolate_lru_page(struct page *page, isolate_mode_t mode);
 269 extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem,
 270                                                   gfp_t gfp_mask, bool noswap);
 271 extern unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *mem,
 272                                                 gfp_t gfp_mask, bool noswap,
 273                                                 struct zone *zone,
 274                                                 unsigned long *nr_scanned);
 275 extern unsigned long shrink_all_memory(unsigned long nr_pages);
 276 extern int vm_swappiness;
 277 extern int remove_mapping(struct address_space *mapping, struct page *page);
 278 extern unsigned long vm_total_pages;
 279
 280 #ifdef CONFIG_NUMA
 281 extern int zone_reclaim_mode;
 282 extern int sysctl_min_unmapped_ratio;
 283 extern int sysctl_min_slab_ratio;
 284 extern int zone_reclaim(struct zone *, gfp_t, unsigned int);
 285 #else
 286 #define zone_reclaim_mode 0
 287 static inline int zone_reclaim(struct zone *z, gfp_t mask, unsigned int order)
 288 {
 289         return 0;
 290 }
 291 #endif
 292
 293 extern int page_evictable(struct page *page);
 294 extern void check_move_unevictable_pages(struct page **, int nr_pages);
 295
 296 extern unsigned long scan_unevictable_pages;
 297 extern int scan_unevictable_handler(struct ctl_table *, int,
 298                                         void __user *, size_t *, loff_t *);
 299 #ifdef CONFIG_NUMA
 300 extern int scan_unevictable_register_node(struct node *node);
 301 extern void scan_unevictable_unregister_node(struct node *node);
 302 #else
 303 static inline int scan_unevictable_register_node(struct node *node)
 304 {
 305         return 0;
 306 }
 307 static inline void scan_unevictable_unregister_node(struct node *node)
 308 {
 309 }
 310 #endif
 311
 312 extern int kswapd_run(int nid);
 313 extern void kswapd_stop(int nid);
 314 #ifdef CONFIG_MEMCG
 315 extern int mem_cgroup_swappiness(struct mem_cgroup *mem);
 316 #else
 317 static inline int mem_cgroup_swappiness(struct mem_cgroup *mem)
 318 {
 319         return vm_swappiness;
 320 }
 321 #endif
 322 #ifdef CONFIG_MEMCG_SWAP
 323 extern void mem_cgroup_uncharge_swap(swp_entry_t ent);
 324 #else
 325 static inline void mem_cgroup_uncharge_swap(swp_entry_t ent)
 326 {
 327 }
 328 #endif
 329 #ifdef CONFIG_SWAP
 330 /* linux/mm/page_io.c */
 331 extern int swap_readpage(struct page *);
 332 extern int swap_writepage(struct page *page, struct writeback_control *wbc);
 333 extern void end_swap_bio_write(struct bio *bio, int err);
 334 extern int __swap_writepage(struct page *page, struct writeback_control *wbc,
 335         void (*end_write_func)(struct bio *, int));
 336 extern int swap_set_page_dirty(struct page *page);
 337 extern void end_swap_bio_read(struct bio *bio, int err);
 338
 339 int add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
 340                 unsigned long nr_pages, sector_t start_block);
 341 int generic_swapfile_activate(struct swap_info_struct *, struct file *,
 342                 sector_t *);
 343
 344 /* linux/mm/swap_state.c */
 345 extern struct address_space swapper_spaces[];
 346 #define swap_address_space(entry) (&swapper_spaces[swp_type(entry)])
 347 extern unsigned long total_swapcache_pages(void);
 348 extern void show_swap_cache_info(void);
 349 extern int add_to_swap(struct page *, struct list_head *list);
 350 extern int add_to_swap_cache(struct page *, swp_entry_t, gfp_t);
 351 extern int __add_to_swap_cache(struct page *page, swp_entry_t entry);
 352 extern void __delete_from_swap_cache(struct page *);
 353 extern void delete_from_swap_cache(struct page *);
 354 extern void free_page_and_swap_cache(struct page *);
 355 extern void free_pages_and_swap_cache(struct page **, int);
 356 extern struct page *lookup_swap_cache(swp_entry_t);
 357 extern struct page *read_swap_cache_async(swp_entry_t, gfp_t,
 358                         struct vm_area_struct *vma, unsigned long addr);
 359 extern struct page *swapin_readahead(swp_entry_t, gfp_t,
 360                         struct vm_area_struct *vma, unsigned long addr);
 361
 362 /* linux/mm/swapfile.c */
 363 extern atomic_long_t nr_swap_pages;
 364 extern long total_swap_pages;
 365
 366 /* Swap 50% full? Release swapcache more aggressively.. */
 367 static inline bool vm_swap_full(void)
 368 {
 369         return atomic_long_read(&nr_swap_pages) * 2 < total_swap_pages;
 370 }
 371
 372 static inline long get_nr_swap_pages(void)
 373 {
 374         return atomic_long_read(&nr_swap_pages);
 375 }
 376
 377 extern void si_swapinfo(struct sysinfo *);
 378 extern swp_entry_t get_swap_page(void);
 379 extern swp_entry_t get_swap_page_of_type(int);
 380 extern int add_swap_count_continuation(swp_entry_t, gfp_t);
 381 extern void swap_shmem_alloc(swp_entry_t);
 382 extern int swap_duplicate(swp_entry_t);
 383 extern int swapcache_prepare(swp_entry_t);
 384 extern void swap_free(swp_entry_t);
 385 extern void swapcache_free(swp_entry_t, struct page *page);
 386 extern int free_swap_and_cache(swp_entry_t);
 387 extern int swap_type_of(dev_t, sector_t, struct block_device **);
 388 extern unsigned int count_swap_pages(int, int);
 389 extern sector_t map_swap_page(struct page *, struct block_device **);
 390 extern sector_t swapdev_block(int, pgoff_t);
 391 extern int page_swapcount(struct page *);
 392 extern struct swap_info_struct *page_swap_info(struct page *);
 393 extern int reuse_swap_page(struct page *);
 394 extern int try_to_free_swap(struct page *);
 395 struct backing_dev_info;
 396
 397 #ifdef CONFIG_MEMCG
 398 extern void
 399 mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent, bool swapout);
 400 #else
 401 static inline void
 402 mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent, bool swapout)
 403 {
 404 }
 405 #endif
 406
 407 #else /* CONFIG_SWAP */
 408
 409 #define get_nr_swap_pages()                     0L
 410 #define total_swap_pages                        0L
 411 #define total_swapcache_pages()                 0UL
 412 #define vm_swap_full()                          0
 413
 414 #define si_swapinfo(val) \
 415         do { (val)->freeswap = (val)->totalswap = 0; } while (0)
 416 /* only sparc can not include linux/pagemap.h in this file
 417  * so leave page_cache_release and release_pages undeclared... */
 418 #define free_page_and_swap_cache(page) \
 419         page_cache_release(page)
 420 #define free_pages_and_swap_cache(pages, nr) \
 421         release_pages((pages), (nr), 0);
 422
 423 static inline void show_swap_cache_info(void)
 424 {
 425 }
 426
 427 #define free_swap_and_cache(swp)        is_migration_entry(swp)
 428 #define swapcache_prepare(swp)          is_migration_entry(swp)
 429
 430 static inline int add_swap_count_continuation(swp_entry_t swp, gfp_t gfp_mask)
 431 {
 432         return 0;
 433 }
 434
 435 static inline void swap_shmem_alloc(swp_entry_t swp)
 436 {
 437 }
 438
 439 static inline int swap_duplicate(swp_entry_t swp)
 440 {
 441         return 0;
 442 }
 443
 444 static inline void swap_free(swp_entry_t swp)
 445 {
 446 }
 447
 448 static inline void swapcache_free(swp_entry_t swp, struct page *page)
 449 {
 450 }
 451
 452 static inline struct page *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
 453                         struct vm_area_struct *vma, unsigned long addr)
 454 {
 455         return NULL;
 456 }
 457
 458 static inline int swap_writepage(struct page *p, struct writeback_control *wbc)
 459 {
 460         return 0;
 461 }
 462
 463 static inline struct page *lookup_swap_cache(swp_entry_t swp)
 464 {
 465         return NULL;
 466 }
 467
 468 static inline int add_to_swap(struct page *page, struct list_head *list)
 469 {
 470         return 0;
 471 }
 472
 473 static inline int add_to_swap_cache(struct page *page, swp_entry_t entry,
 474                                                         gfp_t gfp_mask)
 475 {
 476         return -1;
 477 }
 478
 479 static inline void __delete_from_swap_cache(struct page *page)
 480 {
 481 }
 482
 483 static inline void delete_from_swap_cache(struct page *page)
 484 {
 485 }
 486
 487 static inline int page_swapcount(struct page *page)
 488 {
 489         return 0;
 490 }
 491
 492 #define reuse_swap_page(page)   (page_mapcount(page) == 1)
 493
 494 static inline int try_to_free_swap(struct page *page)
 495 {
 496         return 0;
 497 }
 498
 499 static inline swp_entry_t get_swap_page(void)
 500 {
 501         swp_entry_t entry;
 502         entry.val = 0;
 503         return entry;
 504 }
 505
 506 static inline void
 507 mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent)
 508 {
 509 }
 510
 511 #endif /* CONFIG_SWAP */
 512 #endif /* __KERNEL__*/
 513 #endif /* _LINUX_SWAP_H */