4 #include <linux/sched.h>
5 #include <linux/errno.h>
9 #include <linux/config.h>
10 #include <linux/gfp.h>
11 #include <linux/list.h>
12 #include <linux/mmzone.h>
13 #include <linux/rbtree.h>
14 #include <linux/prio_tree.h>
16 #include <linux/ckrm_mem.h>
21 #ifndef CONFIG_DISCONTIGMEM /* Don't use mapnrs, do it properly */
22 extern unsigned long max_mapnr;
25 extern unsigned long num_physpages;
26 extern void * high_memory;
27 extern unsigned long vmalloc_earlyreserve;
28 extern int page_cluster;
31 extern int sysctl_legacy_va_layout;
33 #define sysctl_legacy_va_layout 0
37 #include <asm/pgtable.h>
38 #include <asm/processor.h>
39 #include <asm/atomic.h>
42 #define MM_VM_SIZE(mm) TASK_SIZE
45 #define nth_page(page,n) pfn_to_page(page_to_pfn((page)) + (n))
48 * Linux kernel virtual memory manager primitives.
49 * The idea being to have a "virtual" mm in the same way
50 * we have a virtual fs - giving a cleaner interface to the
51 * mm details, and allowing different kinds of memory mappings
52 * (from shared memory to executable loading to arbitrary
57 * This struct defines a memory VMM memory area. There is one of these
58 * per VM-area/task. A VM area is any part of the process virtual memory
59 * space that has a special rule for the page-fault handlers (ie a shared
60 * library, the executable area etc).
62 struct vm_area_struct {
63 struct mm_struct * vm_mm; /* The address space we belong to. */
64 unsigned long vm_start; /* Our start address within vm_mm. */
65 unsigned long vm_end; /* The first byte after our end address
68 /* linked list of VM areas per task, sorted by address */
69 struct vm_area_struct *vm_next;
71 pgprot_t vm_page_prot; /* Access permissions of this VMA. */
72 unsigned long vm_flags; /* Flags, listed below. */
77 * For areas with an address space and backing store,
78 * linkage into the address_space->i_mmap prio tree, or
79 * linkage to the list of like vmas hanging off its node, or
80 * linkage of vma in the address_space->i_mmap_nonlinear list.
84 struct list_head list;
85 void *parent; /* aligns with prio_tree_node parent */
86 struct vm_area_struct *head;
89 struct prio_tree_node prio_tree_node;
93 * A file's MAP_PRIVATE vma can be in both i_mmap tree and anon_vma
94 * list, after a COW of one of the file pages. A MAP_SHARED vma
95 * can only be in the i_mmap tree. An anonymous MAP_PRIVATE, stack
96 * or brk vma (with NULL file) can only be in an anon_vma list.
98 struct list_head anon_vma_node; /* Serialized by anon_vma->lock */
99 struct anon_vma *anon_vma; /* Serialized by page_table_lock */
101 /* Function pointers to deal with this struct. */
102 struct vm_operations_struct * vm_ops;
104 /* Information about our backing store: */
105 unsigned long vm_pgoff; /* Offset (within vm_file) in PAGE_SIZE
106 units, *not* PAGE_CACHE_SIZE */
107 struct file * vm_file; /* File we map to (can be NULL). */
108 void * vm_private_data; /* was vm_pte (shared mem) */
111 struct mempolicy *vm_policy; /* NUMA policy for the VMA */
118 #define VM_READ 0x00000001 /* currently active flags */
119 #define VM_WRITE 0x00000002
120 #define VM_EXEC 0x00000004
121 #define VM_SHARED 0x00000008
123 #define VM_MAYREAD 0x00000010 /* limits for mprotect() etc */
124 #define VM_MAYWRITE 0x00000020
125 #define VM_MAYEXEC 0x00000040
126 #define VM_MAYSHARE 0x00000080
128 #define VM_GROWSDOWN 0x00000100 /* general info on the segment */
129 #define VM_GROWSUP 0x00000200
130 #define VM_SHM 0x00000400 /* shared memory area, don't swap out */
131 #define VM_DENYWRITE 0x00000800 /* ETXTBSY on write attempts.. */
133 #define VM_EXECUTABLE 0x00001000
134 #define VM_LOCKED 0x00002000
135 #define VM_IO 0x00004000 /* Memory mapped I/O or similar */
137 /* Used by sys_madvise() */
138 #define VM_SEQ_READ 0x00008000 /* App will access data sequentially */
139 #define VM_RAND_READ 0x00010000 /* App will not benefit from clustered reads */
141 #define VM_DONTCOPY 0x00020000 /* Do not copy this vma on fork */
142 #define VM_DONTEXPAND 0x00040000 /* Cannot expand with mremap() */
143 #define VM_RESERVED 0x00080000 /* Don't unmap it from swap_out */
144 #define VM_ACCOUNT 0x00100000 /* Is a VM accounted object */
145 #define VM_HUGETLB 0x00400000 /* Huge TLB Page VM */
146 #define VM_NONLINEAR 0x00800000 /* Is non-linear (remap_file_pages) */
148 #ifndef VM_STACK_DEFAULT_FLAGS /* arch can override this */
149 #define VM_STACK_DEFAULT_FLAGS VM_DATA_DEFAULT_FLAGS
152 #ifdef CONFIG_STACK_GROWSUP
153 #define VM_STACK_FLAGS (VM_GROWSUP | VM_STACK_DEFAULT_FLAGS | VM_ACCOUNT)
155 #define VM_STACK_FLAGS (VM_GROWSDOWN | VM_STACK_DEFAULT_FLAGS | VM_ACCOUNT)
158 #define VM_READHINTMASK (VM_SEQ_READ | VM_RAND_READ)
159 #define VM_ClearReadHint(v) (v)->vm_flags &= ~VM_READHINTMASK
160 #define VM_NormalReadHint(v) (!((v)->vm_flags & VM_READHINTMASK))
161 #define VM_SequentialReadHint(v) ((v)->vm_flags & VM_SEQ_READ)
162 #define VM_RandomReadHint(v) ((v)->vm_flags & VM_RAND_READ)
165 * mapping from the currently active vm_flags protection bits (the
166 * low four bits) to a page protection mask..
168 extern pgprot_t protection_map[16];
172 * These are the virtual MM functions - opening of an area, closing and
173 * unmapping it (needed to keep files on disk up-to-date etc), pointer
174 * to the functions called when a no-page or a wp-page exception occurs.
176 struct vm_operations_struct {
177 void (*open)(struct vm_area_struct * area);
178 void (*close)(struct vm_area_struct * area);
179 struct page * (*nopage)(struct vm_area_struct * area, unsigned long address, int *type);
180 int (*populate)(struct vm_area_struct * area, unsigned long address, unsigned long len, pgprot_t prot, unsigned long pgoff, int nonblock);
182 int (*set_policy)(struct vm_area_struct *vma, struct mempolicy *new);
183 struct mempolicy *(*get_policy)(struct vm_area_struct *vma,
191 #ifdef ARCH_HAS_ATOMIC_UNSIGNED
192 typedef unsigned page_flags_t;
194 typedef unsigned long page_flags_t;
198 * Each physical page in the system has a struct page associated with
199 * it to keep track of whatever it is we are using the page for at the
200 * moment. Note that we have no way to track which tasks are using
204 page_flags_t flags; /* Atomic flags, some possibly
205 * updated asynchronously */
206 atomic_t _count; /* Usage count, see below. */
207 atomic_t _mapcount; /* Count of ptes mapped in mms,
208 * to show when page is mapped
209 * & limit reverse map searches.
211 unsigned long private; /* Mapping-private opaque data:
212 * usually used for buffer_heads
213 * if PagePrivate set; used for
214 * swp_entry_t if PageSwapCache
216 struct address_space *mapping; /* If low bit clear, points to
217 * inode address_space, or NULL.
218 * If page mapped as anonymous
219 * memory, low bit is set, and
220 * it points to anon_vma object:
221 * see PAGE_MAPPING_ANON below.
223 pgoff_t index; /* Our offset within mapping. */
224 struct list_head lru; /* Pageout list, eg. active_list
225 * protected by zone->lru_lock !
228 * On machines where all RAM is mapped into kernel address space,
229 * we can simply calculate the virtual address. On machines with
230 * highmem some memory is mapped into kernel virtual memory
231 * dynamically, so we need a place to store that address.
232 * Note that this field could be 16 bits on x86 ... ;)
234 * Architectures with slow multiplication can define
235 * WANT_PAGE_VIRTUAL in asm/page.h
237 #if defined(WANT_PAGE_VIRTUAL)
238 void *virtual; /* Kernel virtual address (NULL if
239 not kmapped, ie. highmem) */
240 #endif /* WANT_PAGE_VIRTUAL */
241 #ifdef CONFIG_CKRM_RES_MEM
242 struct ckrm_zone *ckrm_zone;
243 #endif // CONFIG_CKRM_RES_MEM
247 * FIXME: take this include out, include page-flags.h in
248 * files which need it (119 of them)
250 #include <linux/page-flags.h>
253 * Methods to modify the page usage count.
255 * What counts for a page usage:
256 * - cache mapping (page->mapping)
257 * - private data (page->private)
258 * - page mapped in a task's page tables, each mapping
259 * is counted separately
261 * Also, many kernel routines increase the page count before a critical
262 * routine so they can be sure the page doesn't go away from under them.
264 * Since 2.6.6 (approx), a free page has ->_count = -1. This is so that we
265 * can use atomic_add_negative(-1, page->_count) to detect when the page
266 * becomes free and so that we can also use atomic_inc_and_test to atomically
267 * detect when we just tried to grab a ref on a page which some other CPU has
268 * already deemed to be freeable.
270 * NO code should make assumptions about this internal detail! Use the provided
271 * macros which retain the old rules: page_count(page) == 0 is a free page.
275 * Drop a ref, return true if the logical refcount fell to zero (the page has
278 #define put_page_testzero(p) \
280 BUG_ON(page_count(p) == 0); \
281 atomic_add_negative(-1, &(p)->_count); \
285 * Grab a ref, return true if the page previously had a logical refcount of
286 * zero. ie: returns true if we just grabbed an already-deemed-to-be-free page
288 #define get_page_testone(p) atomic_inc_and_test(&(p)->_count)
290 #define set_page_count(p,v) atomic_set(&(p)->_count, v - 1)
291 #define __put_page(p) atomic_dec(&(p)->_count)
293 extern void FASTCALL(__page_cache_release(struct page *));
295 #ifdef CONFIG_HUGETLB_PAGE
297 static inline int page_count(struct page *p)
300 p = (struct page *)p->private;
301 return atomic_read(&(p)->_count) + 1;
304 static inline void get_page(struct page *page)
306 if (unlikely(PageCompound(page)))
307 page = (struct page *)page->private;
308 atomic_inc(&page->_count);
311 void put_page(struct page *page);
313 #else /* CONFIG_HUGETLB_PAGE */
315 #define page_count(p) (atomic_read(&(p)->_count) + 1)
317 static inline void get_page(struct page *page)
319 atomic_inc(&page->_count);
322 static inline void put_page(struct page *page)
324 if (!PageReserved(page) && put_page_testzero(page))
325 __page_cache_release(page);
328 #endif /* CONFIG_HUGETLB_PAGE */
331 * Multiple processes may "see" the same page. E.g. for untouched
332 * mappings of /dev/null, all processes see the same page full of
333 * zeroes, and text pages of executables and shared libraries have
334 * only one copy in memory, at most, normally.
336 * For the non-reserved pages, page_count(page) denotes a reference count.
337 * page_count() == 0 means the page is free.
338 * page_count() == 1 means the page is used for exactly one purpose
339 * (e.g. a private data page of one process).
341 * A page may be used for kmalloc() or anyone else who does a
342 * __get_free_page(). In this case the page_count() is at least 1, and
343 * all other fields are unused but should be 0 or NULL. The
344 * management of this page is the responsibility of the one who uses
347 * The other pages (we may call them "process pages") are completely
348 * managed by the Linux memory manager: I/O, buffers, swapping etc.
349 * The following discussion applies only to them.
351 * A page may belong to an inode's memory mapping. In this case,
352 * page->mapping is the pointer to the inode, and page->index is the
353 * file offset of the page, in units of PAGE_CACHE_SIZE.
355 * A page contains an opaque `private' member, which belongs to the
356 * page's address_space. Usually, this is the address of a circular
357 * list of the page's disk buffers.
359 * For pages belonging to inodes, the page_count() is the number of
360 * attaches, plus 1 if `private' contains something, plus one for
361 * the page cache itself.
363 * All pages belonging to an inode are in these doubly linked lists:
364 * mapping->clean_pages, mapping->dirty_pages and mapping->locked_pages;
365 * using the page->list list_head. These fields are also used for
366 * freelist managemet (when page_count()==0).
368 * There is also a per-mapping radix tree mapping index to the page
369 * in memory if present. The tree is rooted at mapping->root.
371 * All process pages can do I/O:
372 * - inode pages may need to be read from disk,
373 * - inode pages which have been modified and are MAP_SHARED may need
374 * to be written to disk,
375 * - private pages which have been modified may need to be swapped out
376 * to swap space and (later) to be read back into memory.
380 * The zone field is never updated after free_area_init_core()
381 * sets it, so none of the operations on it need to be atomic.
382 * We'll have up to (MAX_NUMNODES * MAX_NR_ZONES) zones total,
383 * so we use (MAX_NODES_SHIFT + MAX_ZONES_SHIFT) here to get enough bits.
385 #define NODEZONE_SHIFT (sizeof(page_flags_t)*8 - MAX_NODES_SHIFT - MAX_ZONES_SHIFT)
386 #define NODEZONE(node, zone) ((node << ZONES_SHIFT) | zone)
388 static inline unsigned long page_zonenum(struct page *page)
390 return (page->flags >> NODEZONE_SHIFT) & (~(~0UL << ZONES_SHIFT));
392 static inline unsigned long page_to_nid(struct page *page)
394 return (page->flags >> (NODEZONE_SHIFT + ZONES_SHIFT));
398 extern struct zone *zone_table[];
400 static inline struct zone *page_zone(struct page *page)
402 return zone_table[page->flags >> NODEZONE_SHIFT];
405 static inline void set_page_zone(struct page *page, unsigned long nodezone_num)
407 page->flags &= ~(~0UL << NODEZONE_SHIFT);
408 page->flags |= nodezone_num << NODEZONE_SHIFT;
411 #ifndef CONFIG_DISCONTIGMEM
412 /* The array of struct pages - for discontigmem use pgdat->lmem_map */
413 extern struct page *mem_map;
416 static inline void *lowmem_page_address(struct page *page)
418 return __va(page_to_pfn(page) << PAGE_SHIFT);
421 #if defined(CONFIG_HIGHMEM) && !defined(WANT_PAGE_VIRTUAL)
422 #define HASHED_PAGE_VIRTUAL
425 #if defined(WANT_PAGE_VIRTUAL)
426 #define page_address(page) ((page)->virtual)
427 #define set_page_address(page, address) \
429 (page)->virtual = (address); \
431 #define page_address_init() do { } while(0)
434 #if defined(HASHED_PAGE_VIRTUAL)
435 void *page_address(struct page *page);
436 void set_page_address(struct page *page, void *virtual);
437 void page_address_init(void);
440 #if !defined(HASHED_PAGE_VIRTUAL) && !defined(WANT_PAGE_VIRTUAL)
441 #define page_address(page) lowmem_page_address(page)
442 #define set_page_address(page, address) do { } while(0)
443 #define page_address_init() do { } while(0)
447 * On an anonymous page mapped into a user virtual memory area,
448 * page->mapping points to its anon_vma, not to a struct address_space;
449 * with the PAGE_MAPPING_ANON bit set to distinguish it.
451 * Please note that, confusingly, "page_mapping" refers to the inode
452 * address_space which maps the page from disk; whereas "page_mapped"
453 * refers to user virtual address space into which the page is mapped.
455 #define PAGE_MAPPING_ANON 1
457 extern struct address_space swapper_space;
458 static inline struct address_space *page_mapping(struct page *page)
460 struct address_space *mapping = page->mapping;
462 if (unlikely(PageSwapCache(page)))
463 mapping = &swapper_space;
464 else if (unlikely((unsigned long)mapping & PAGE_MAPPING_ANON))
469 static inline int PageAnon(struct page *page)
471 return ((unsigned long)page->mapping & PAGE_MAPPING_ANON) != 0;
475 * Return the pagecache index of the passed page. Regular pagecache pages
476 * use ->index whereas swapcache pages use ->private
478 static inline pgoff_t page_index(struct page *page)
480 if (unlikely(PageSwapCache(page)))
481 return page->private;
486 * The atomic page->_mapcount, like _count, starts from -1:
487 * so that transitions both from it and to it can be tracked,
488 * using atomic_inc_and_test and atomic_add_negative(-1).
490 static inline void reset_page_mapcount(struct page *page)
492 atomic_set(&(page)->_mapcount, -1);
495 static inline int page_mapcount(struct page *page)
497 return atomic_read(&(page)->_mapcount) + 1;
501 * Return true if this page is mapped into pagetables.
503 static inline int page_mapped(struct page *page)
505 return atomic_read(&(page)->_mapcount) >= 0;
509 * Error return values for the *_nopage functions
511 #define NOPAGE_SIGBUS (NULL)
512 #define NOPAGE_OOM ((struct page *) (-1))
515 * Different kinds of faults, as returned by handle_mm_fault().
516 * Used to decide whether a process gets delivered SIGBUS or
517 * just gets major/minor fault counters bumped up.
519 #define VM_FAULT_OOM (-1)
520 #define VM_FAULT_SIGBUS 0
521 #define VM_FAULT_MINOR 1
522 #define VM_FAULT_MAJOR 2
524 #define offset_in_page(p) ((unsigned long)(p) & ~PAGE_MASK)
526 extern void show_free_areas(void);
529 struct page *shmem_nopage(struct vm_area_struct *vma,
530 unsigned long address, int *type);
531 int shmem_set_policy(struct vm_area_struct *vma, struct mempolicy *new);
532 struct mempolicy *shmem_get_policy(struct vm_area_struct *vma,
534 int shmem_lock(struct file *file, int lock, struct user_struct *user);
536 #define shmem_nopage filemap_nopage
537 #define shmem_lock(a, b, c) ({0;}) /* always in memory, no need to lock */
538 #define shmem_set_policy(a, b) (0)
539 #define shmem_get_policy(a, b) (NULL)
541 struct file *shmem_file_setup(char *name, loff_t size, unsigned long flags);
543 int shmem_zero_setup(struct vm_area_struct *);
545 static inline int can_do_mlock(void)
547 if (capable(CAP_IPC_LOCK))
549 if (current->signal->rlim[RLIMIT_MEMLOCK].rlim_cur != 0)
553 extern int user_shm_lock(size_t, struct user_struct *);
554 extern void user_shm_unlock(size_t, struct user_struct *);
557 * Parameter block passed down to zap_pte_range in exceptional cases.
560 struct vm_area_struct *nonlinear_vma; /* Check page->index if set */
561 struct address_space *check_mapping; /* Check page->mapping if set */
562 pgoff_t first_index; /* Lowest page->index to unmap */
563 pgoff_t last_index; /* Highest page->index to unmap */
564 int atomic; /* May not schedule() */
567 void zap_page_range(struct vm_area_struct *vma, unsigned long address,
568 unsigned long size, struct zap_details *);
569 int unmap_vmas(struct mmu_gather **tlbp, struct mm_struct *mm,
570 struct vm_area_struct *start_vma, unsigned long start_addr,
571 unsigned long end_addr, unsigned long *nr_accounted,
572 struct zap_details *);
573 void clear_page_tables(struct mmu_gather *tlb, unsigned long first, int nr);
574 int copy_page_range(struct mm_struct *dst, struct mm_struct *src,
575 struct vm_area_struct *vma);
576 int zeromap_page_range(struct vm_area_struct *vma, unsigned long from,
577 unsigned long size, pgprot_t prot);
578 void unmap_mapping_range(struct address_space *mapping,
579 loff_t const holebegin, loff_t const holelen, int even_cows);
581 static inline void unmap_shared_mapping_range(struct address_space *mapping,
582 loff_t const holebegin, loff_t const holelen)
584 unmap_mapping_range(mapping, holebegin, holelen, 0);
587 extern int vmtruncate(struct inode * inode, loff_t offset);
588 extern pmd_t *FASTCALL(__pmd_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address));
589 extern pte_t *FASTCALL(pte_alloc_kernel(struct mm_struct *mm, pmd_t *pmd, unsigned long address));
590 extern pte_t *FASTCALL(pte_alloc_map(struct mm_struct *mm, pmd_t *pmd, unsigned long address));
591 extern int install_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, struct page *page, pgprot_t prot);
592 extern int install_file_pte(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, unsigned long pgoff, pgprot_t prot);
593 extern int handle_mm_fault(struct mm_struct *mm,struct vm_area_struct *vma, unsigned long address, int write_access);
594 extern int make_pages_present(unsigned long addr, unsigned long end);
595 extern int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, int len, int write);
596 void install_arg_page(struct vm_area_struct *, struct page *, unsigned long);
598 int get_user_pages(struct task_struct *tsk, struct mm_struct *mm, unsigned long start,
599 int len, int write, int force, struct page **pages, struct vm_area_struct **vmas);
601 int __set_page_dirty_buffers(struct page *page);
602 int __set_page_dirty_nobuffers(struct page *page);
603 int redirty_page_for_writepage(struct writeback_control *wbc,
605 int FASTCALL(set_page_dirty(struct page *page));
606 int set_page_dirty_lock(struct page *page);
607 int clear_page_dirty_for_io(struct page *page);
610 * Prototype to add a shrinker callback for ageable caches.
612 * These functions are passed a count `nr_to_scan' and a gfpmask. They should
613 * scan `nr_to_scan' objects, attempting to free them.
615 * The callback must the number of objects which remain in the cache.
617 * The callback will be passes nr_to_scan == 0 when the VM is querying the
618 * cache size, so a fastpath for that case is appropriate.
620 typedef int (*shrinker_t)(int nr_to_scan, unsigned int gfp_mask);
623 * Add an aging callback. The int is the number of 'seeks' it takes
624 * to recreate one of the objects that these functions age.
627 #define DEFAULT_SEEKS 2
629 extern struct shrinker *set_shrinker(int, shrinker_t);
630 extern void remove_shrinker(struct shrinker *shrinker);
633 * On a two-level page table, this ends up being trivial. Thus the
634 * inlining and the symmetry break with pte_alloc_map() that does all
635 * of this out-of-line.
637 static inline pmd_t *pmd_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
640 return __pmd_alloc(mm, pgd, address);
641 return pmd_offset(pgd, address);
644 extern void free_area_init(unsigned long * zones_size);
645 extern void free_area_init_node(int nid, pg_data_t *pgdat,
646 unsigned long * zones_size, unsigned long zone_start_pfn,
647 unsigned long *zholes_size);
648 extern void memmap_init_zone(unsigned long, int, unsigned long, unsigned long);
649 extern void mem_init(void);
650 extern void show_mem(void);
651 extern void si_meminfo(struct sysinfo * val);
652 extern void si_meminfo_node(struct sysinfo *val, int nid);
655 void vma_prio_tree_add(struct vm_area_struct *, struct vm_area_struct *old);
656 void vma_prio_tree_insert(struct vm_area_struct *, struct prio_tree_root *);
657 void vma_prio_tree_remove(struct vm_area_struct *, struct prio_tree_root *);
658 struct vm_area_struct *vma_prio_tree_next(struct vm_area_struct *vma,
659 struct prio_tree_iter *iter);
661 #define vma_prio_tree_foreach(vma, iter, root, begin, end) \
662 for (prio_tree_iter_init(iter, root, begin, end), vma = NULL; \
663 (vma = vma_prio_tree_next(vma, iter)); )
665 static inline void vma_nonlinear_insert(struct vm_area_struct *vma,
666 struct list_head *list)
668 vma->shared.vm_set.parent = NULL;
669 list_add_tail(&vma->shared.vm_set.list, list);
673 extern void vma_adjust(struct vm_area_struct *vma, unsigned long start,
674 unsigned long end, pgoff_t pgoff, struct vm_area_struct *insert);
675 extern struct vm_area_struct *vma_merge(struct mm_struct *,
676 struct vm_area_struct *prev, unsigned long addr, unsigned long end,
677 unsigned long vm_flags, struct anon_vma *, struct file *, pgoff_t,
679 extern struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *);
680 extern int split_vma(struct mm_struct *,
681 struct vm_area_struct *, unsigned long addr, int new_below);
682 extern int insert_vm_struct(struct mm_struct *, struct vm_area_struct *);
683 extern void __vma_link_rb(struct mm_struct *, struct vm_area_struct *,
684 struct rb_node **, struct rb_node *);
685 extern struct vm_area_struct *copy_vma(struct vm_area_struct **,
686 unsigned long addr, unsigned long len, pgoff_t pgoff);
687 extern void exit_mmap(struct mm_struct *);
689 extern unsigned long get_unmapped_area_prot(struct file *, unsigned long, unsigned long, unsigned long, unsigned long, int);
692 static inline unsigned long get_unmapped_area(struct file * file, unsigned long addr,
693 unsigned long len, unsigned long pgoff, unsigned long flags)
695 return get_unmapped_area_prot(file, addr, len, pgoff, flags, 0);
698 extern unsigned long do_mmap_pgoff(struct file *file, unsigned long addr,
699 unsigned long len, unsigned long prot,
700 unsigned long flag, unsigned long pgoff);
702 static inline unsigned long do_mmap(struct file *file, unsigned long addr,
703 unsigned long len, unsigned long prot,
704 unsigned long flag, unsigned long offset)
706 unsigned long ret = -EINVAL;
707 if ((offset + PAGE_ALIGN(len)) < offset)
709 if (!(offset & ~PAGE_MASK))
710 ret = do_mmap_pgoff(file, addr, len, prot, flag, offset >> PAGE_SHIFT);
715 extern int do_munmap(struct mm_struct *, unsigned long, size_t);
717 extern unsigned long __do_brk(unsigned long, unsigned long);
718 extern unsigned long do_brk(unsigned long, unsigned long);
721 extern unsigned long page_unuse(struct page *);
722 extern void truncate_inode_pages(struct address_space *, loff_t);
724 /* generic vm_area_ops exported for stackable file systems */
725 struct page *filemap_nopage(struct vm_area_struct *, unsigned long, int *);
727 /* mm/page-writeback.c */
728 int write_one_page(struct page *page, int wait);
731 #define VM_MAX_READAHEAD 128 /* kbytes */
732 #define VM_MIN_READAHEAD 16 /* kbytes (includes current page) */
734 int do_page_cache_readahead(struct address_space *mapping, struct file *filp,
735 unsigned long offset, unsigned long nr_to_read);
736 int force_page_cache_readahead(struct address_space *mapping, struct file *filp,
737 unsigned long offset, unsigned long nr_to_read);
738 void page_cache_readahead(struct address_space *mapping,
739 struct file_ra_state *ra,
741 unsigned long offset);
742 void handle_ra_miss(struct address_space *mapping,
743 struct file_ra_state *ra, pgoff_t offset);
744 unsigned long max_sane_readahead(unsigned long nr);
746 /* Do stack extension */
747 extern int expand_stack(struct vm_area_struct * vma, unsigned long address);
749 /* Look up the first VMA which satisfies addr < vm_end, NULL if none. */
750 extern struct vm_area_struct * find_vma(struct mm_struct * mm, unsigned long addr);
751 extern struct vm_area_struct * find_vma_prev(struct mm_struct * mm, unsigned long addr,
752 struct vm_area_struct **pprev);
754 /* Look up the first VMA which intersects the interval start_addr..end_addr-1,
755 NULL if none. Assume start_addr < end_addr. */
756 static inline struct vm_area_struct * find_vma_intersection(struct mm_struct * mm, unsigned long start_addr, unsigned long end_addr)
758 struct vm_area_struct * vma = find_vma(mm,start_addr);
760 if (vma && end_addr <= vma->vm_start)
765 static inline unsigned long vma_pages(struct vm_area_struct *vma)
767 return (vma->vm_end - vma->vm_start) >> PAGE_SHIFT;
770 extern struct vm_area_struct *find_extend_vma(struct mm_struct *mm, unsigned long addr);
772 extern struct page * vmalloc_to_page(void *addr);
773 extern unsigned long vmalloc_to_pfn(void *addr);
774 extern struct page * follow_page(struct mm_struct *mm, unsigned long address,
776 int remap_pfn_range(struct vm_area_struct *, unsigned long,
777 unsigned long, unsigned long, pgprot_t);
779 static inline __deprecated /* since 25 Sept 2004 -- wli */
780 int remap_page_range(struct vm_area_struct *vma, unsigned long uvaddr,
781 unsigned long paddr, unsigned long size, pgprot_t prot)
783 return remap_pfn_range(vma, uvaddr, paddr >> PAGE_SHIFT, size, prot);
786 #ifdef CONFIG_PROC_FS
787 void __vm_stat_account(struct mm_struct *, unsigned long, struct file *, long);
789 static inline void __vm_stat_account(struct mm_struct *mm,
790 unsigned long flags, struct file *file, long pages)
793 #endif /* CONFIG_PROC_FS */
795 static inline void vm_stat_account(struct vm_area_struct *vma)
797 __vm_stat_account(vma->vm_mm, vma->vm_flags, vma->vm_file,
801 static inline void vm_stat_unaccount(struct vm_area_struct *vma)
803 __vm_stat_account(vma->vm_mm, vma->vm_flags, vma->vm_file,
807 #ifndef CONFIG_DEBUG_PAGEALLOC
809 kernel_map_pages(struct page *page, int numpages, int enable)
814 #ifndef CONFIG_ARCH_GATE_AREA
815 extern struct vm_area_struct *get_gate_vma(struct task_struct *tsk);
816 int in_gate_area(struct task_struct *task, unsigned long addr);
819 #endif /* __KERNEL__ */
820 #endif /* _LINUX_MM_H */