mmap.c 88.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5
/*
 * mm/mmap.c
 *
 * Written by obz.
 *
A
Alan Cox 已提交
6
 * Address space accounting code	<alan@lxorguk.ukuu.org.uk>
L
Linus Torvalds 已提交
7 8
 */

9 10
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

11
#include <linux/kernel.h>
L
Linus Torvalds 已提交
12
#include <linux/slab.h>
A
Alexey Dobriyan 已提交
13
#include <linux/backing-dev.h>
L
Linus Torvalds 已提交
14
#include <linux/mm.h>
D
Davidlohr Bueso 已提交
15
#include <linux/vmacache.h>
L
Linus Torvalds 已提交
16 17 18 19 20
#include <linux/shm.h>
#include <linux/mman.h>
#include <linux/pagemap.h>
#include <linux/swap.h>
#include <linux/syscalls.h>
21
#include <linux/capability.h>
L
Linus Torvalds 已提交
22 23 24 25 26 27 28
#include <linux/init.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/personality.h>
#include <linux/security.h>
#include <linux/hugetlb.h>
#include <linux/profile.h>
29
#include <linux/export.h>
L
Linus Torvalds 已提交
30 31 32
#include <linux/mount.h>
#include <linux/mempolicy.h>
#include <linux/rmap.h>
A
Andrea Arcangeli 已提交
33
#include <linux/mmu_notifier.h>
34
#include <linux/mmdebug.h>
35
#include <linux/perf_event.h>
A
Al Viro 已提交
36
#include <linux/audit.h>
A
Andrea Arcangeli 已提交
37
#include <linux/khugepaged.h>
38
#include <linux/uprobes.h>
39
#include <linux/rbtree_augmented.h>
40
#include <linux/sched/sysctl.h>
41 42
#include <linux/notifier.h>
#include <linux/memory.h>
43
#include <linux/printk.h>
L
Linus Torvalds 已提交
44 45 46 47

#include <asm/uaccess.h>
#include <asm/cacheflush.h>
#include <asm/tlb.h>
48
#include <asm/mmu_context.h>
L
Linus Torvalds 已提交
49

50 51
#include "internal.h"

52 53 54 55
#ifndef arch_mmap_check
#define arch_mmap_check(addr, len, flags)	(0)
#endif

56 57 58 59
#ifndef arch_rebalance_pgtables
#define arch_rebalance_pgtables(addr, len)		(addr)
#endif

60 61 62 63
static void unmap_region(struct mm_struct *mm,
		struct vm_area_struct *vma, struct vm_area_struct *prev,
		unsigned long start, unsigned long end);

L
Linus Torvalds 已提交
64 65 66 67 68 69 70 71 72
/* description of effects of mapping type and prot in current implementation.
 * this is due to the limited x86 page protection hardware.  The expected
 * behavior is in parens:
 *
 * map_type	prot
 *		PROT_NONE	PROT_READ	PROT_WRITE	PROT_EXEC
 * MAP_SHARED	r: (no) no	r: (yes) yes	r: (no) yes	r: (no) yes
 *		w: (no) no	w: (no) no	w: (yes) yes	w: (no) no
 *		x: (no) no	x: (no) yes	x: (no) yes	x: (yes) yes
V
vishnu.ps 已提交
73
 *
L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82 83
 * MAP_PRIVATE	r: (no) no	r: (yes) yes	r: (no) yes	r: (no) yes
 *		w: (no) no	w: (no) no	w: (copy) copy	w: (no) no
 *		x: (no) no	x: (no) yes	x: (no) yes	x: (yes) yes
 *
 */
pgprot_t protection_map[16] = {
	__P000, __P001, __P010, __P011, __P100, __P101, __P110, __P111,
	__S000, __S001, __S010, __S011, __S100, __S101, __S110, __S111
};

84 85
pgprot_t vm_get_page_prot(unsigned long vm_flags)
{
86 87 88
	return __pgprot(pgprot_val(protection_map[vm_flags &
				(VM_READ|VM_WRITE|VM_EXEC|VM_SHARED)]) |
			pgprot_val(arch_vm_get_page_prot(vm_flags)));
89 90 91
}
EXPORT_SYMBOL(vm_get_page_prot);

92 93
int sysctl_overcommit_memory __read_mostly = OVERCOMMIT_GUESS;  /* heuristic overcommit */
int sysctl_overcommit_ratio __read_mostly = 50;	/* default is 50% */
94
unsigned long sysctl_overcommit_kbytes __read_mostly;
95
int sysctl_max_map_count __read_mostly = DEFAULT_MAX_MAP_COUNT;
96
unsigned long sysctl_user_reserve_kbytes __read_mostly = 1UL << 17; /* 128MB */
97
unsigned long sysctl_admin_reserve_kbytes __read_mostly = 1UL << 13; /* 8MB */
98 99 100 101 102
/*
 * Make sure vm_committed_as in one cacheline and not cacheline shared with
 * other variables. It can be updated by several CPUs frequently.
 */
struct percpu_counter vm_committed_as ____cacheline_aligned_in_smp;
L
Linus Torvalds 已提交
103

104 105 106 107 108 109 110 111 112 113 114 115 116 117
/*
 * The global memory commitment made in the system can be a metric
 * that can be used to drive ballooning decisions when Linux is hosted
 * as a guest. On Hyper-V, the host implements a policy engine for dynamically
 * balancing memory across competing virtual machines that are hosted.
 * Several metrics drive this policy engine including the guest reported
 * memory commitment.
 */
unsigned long vm_memory_committed(void)
{
	return percpu_counter_read_positive(&vm_committed_as);
}
EXPORT_SYMBOL_GPL(vm_memory_committed);

L
Linus Torvalds 已提交
118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
/*
 * Check that a process has enough memory to allocate a new virtual
 * mapping. 0 means there is enough memory for the allocation to
 * succeed and -ENOMEM implies there is not.
 *
 * We currently support three overcommit policies, which are set via the
 * vm.overcommit_memory sysctl.  See Documentation/vm/overcommit-accounting
 *
 * Strict overcommit modes added 2002 Feb 26 by Alan Cox.
 * Additional code 2002 Jul 20 by Robert Love.
 *
 * cap_sys_admin is 1 if the process has admin privileges, 0 otherwise.
 *
 * Note this is a helper function intended to be used by LSMs which
 * wish to use this logic.
 */
134
int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin)
L
Linus Torvalds 已提交
135
{
136
	unsigned long free, allowed, reserve;
L
Linus Torvalds 已提交
137

138 139 140 141
	VM_WARN_ONCE(percpu_counter_read(&vm_committed_as) <
			-(s64)vm_committed_as_batch * num_online_cpus(),
			"memory commitment underflow");

L
Linus Torvalds 已提交
142 143 144 145 146 147 148 149 150
	vm_acct_memory(pages);

	/*
	 * Sometimes we want to use more memory than we have
	 */
	if (sysctl_overcommit_memory == OVERCOMMIT_ALWAYS)
		return 0;

	if (sysctl_overcommit_memory == OVERCOMMIT_GUESS) {
151 152 153 154 155 156 157 158 159 160
		free = global_page_state(NR_FREE_PAGES);
		free += global_page_state(NR_FILE_PAGES);

		/*
		 * shmem pages shouldn't be counted as free in this
		 * case, they can't be purged, only swapped out, and
		 * that won't affect the overall amount of available
		 * memory in the system.
		 */
		free -= global_page_state(NR_SHMEM);
L
Linus Torvalds 已提交
161

162
		free += get_nr_swap_pages();
L
Linus Torvalds 已提交
163 164 165 166 167 168 169

		/*
		 * Any slabs which are created with the
		 * SLAB_RECLAIM_ACCOUNT flag claim to have contents
		 * which are reclaimable, under pressure.  The dentry
		 * cache and most inode caches should fall into this
		 */
170
		free += global_page_state(NR_SLAB_RECLAIMABLE);
L
Linus Torvalds 已提交
171

172 173 174
		/*
		 * Leave reserved pages. The pages are not for anonymous pages.
		 */
175
		if (free <= totalreserve_pages)
176 177
			goto error;
		else
178
			free -= totalreserve_pages;
179 180

		/*
181
		 * Reserve some for root
182
		 */
L
Linus Torvalds 已提交
183
		if (!cap_sys_admin)
184
			free -= sysctl_admin_reserve_kbytes >> (PAGE_SHIFT - 10);
L
Linus Torvalds 已提交
185 186 187

		if (free > pages)
			return 0;
188 189

		goto error;
L
Linus Torvalds 已提交
190 191
	}

192
	allowed = vm_commit_limit();
L
Linus Torvalds 已提交
193
	/*
194
	 * Reserve some for root
L
Linus Torvalds 已提交
195 196
	 */
	if (!cap_sys_admin)
197
		allowed -= sysctl_admin_reserve_kbytes >> (PAGE_SHIFT - 10);
L
Linus Torvalds 已提交
198

199 200 201 202 203 204 205
	/*
	 * Don't let a single process grow so big a user can't recover
	 */
	if (mm) {
		reserve = sysctl_user_reserve_kbytes >> (PAGE_SHIFT - 10);
		allowed -= min(mm->total_vm / 32, reserve);
	}
L
Linus Torvalds 已提交
206

207
	if (percpu_counter_read_positive(&vm_committed_as) < allowed)
L
Linus Torvalds 已提交
208
		return 0;
209
error:
L
Linus Torvalds 已提交
210 211 212 213 214 215
	vm_unacct_memory(pages);

	return -ENOMEM;
}

/*
216
 * Requires inode->i_mapping->i_mmap_mutex
L
Linus Torvalds 已提交
217 218 219 220 221
 */
static void __remove_shared_vm_struct(struct vm_area_struct *vma,
		struct file *file, struct address_space *mapping)
{
	if (vma->vm_flags & VM_DENYWRITE)
A
Al Viro 已提交
222
		atomic_inc(&file_inode(file)->i_writecount);
L
Linus Torvalds 已提交
223
	if (vma->vm_flags & VM_SHARED)
224
		mapping_unmap_writable(mapping);
L
Linus Torvalds 已提交
225 226 227

	flush_dcache_mmap_lock(mapping);
	if (unlikely(vma->vm_flags & VM_NONLINEAR))
228
		list_del_init(&vma->shared.nonlinear);
L
Linus Torvalds 已提交
229
	else
230
		vma_interval_tree_remove(vma, &mapping->i_mmap);
L
Linus Torvalds 已提交
231 232 233 234
	flush_dcache_mmap_unlock(mapping);
}

/*
235
 * Unlink a file-based vm structure from its interval tree, to hide
236
 * vma from rmap and vmtruncate before freeing its page tables.
L
Linus Torvalds 已提交
237
 */
238
void unlink_file_vma(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
239 240 241 242 243
{
	struct file *file = vma->vm_file;

	if (file) {
		struct address_space *mapping = file->f_mapping;
244
		mutex_lock(&mapping->i_mmap_mutex);
L
Linus Torvalds 已提交
245
		__remove_shared_vm_struct(vma, file, mapping);
246
		mutex_unlock(&mapping->i_mmap_mutex);
L
Linus Torvalds 已提交
247
	}
248 249 250 251 252 253 254 255 256 257
}

/*
 * Close a vm structure and free it, returning the next.
 */
static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
{
	struct vm_area_struct *next = vma->vm_next;

	might_sleep();
L
Linus Torvalds 已提交
258 259
	if (vma->vm_ops && vma->vm_ops->close)
		vma->vm_ops->close(vma);
260
	if (vma->vm_file)
261
		fput(vma->vm_file);
262
	mpol_put(vma_policy(vma));
L
Linus Torvalds 已提交
263
	kmem_cache_free(vm_area_cachep, vma);
264
	return next;
L
Linus Torvalds 已提交
265 266
}

267 268
static unsigned long do_brk(unsigned long addr, unsigned long len);

269
SYSCALL_DEFINE1(brk, unsigned long, brk)
L
Linus Torvalds 已提交
270 271 272 273
{
	unsigned long rlim, retval;
	unsigned long newbrk, oldbrk;
	struct mm_struct *mm = current->mm;
274
	unsigned long min_brk;
275
	bool populate;
L
Linus Torvalds 已提交
276 277 278

	down_write(&mm->mmap_sem);

279
#ifdef CONFIG_COMPAT_BRK
280 281 282 283 284
	/*
	 * CONFIG_COMPAT_BRK can still be overridden by setting
	 * randomize_va_space to 2, which will still cause mm->start_brk
	 * to be arbitrarily shifted
	 */
285
	if (current->brk_randomized)
286 287 288
		min_brk = mm->start_brk;
	else
		min_brk = mm->end_data;
289 290 291 292
#else
	min_brk = mm->start_brk;
#endif
	if (brk < min_brk)
L
Linus Torvalds 已提交
293
		goto out;
R
Ram Gupta 已提交
294 295 296 297 298 299 300

	/*
	 * Check against rlimit here. If this check is done later after the test
	 * of oldbrk with newbrk then it can escape the test and let the data
	 * segment grow beyond its set limit the in case where the limit is
	 * not page aligned -Ram Gupta
	 */
J
Jiri Slaby 已提交
301
	rlim = rlimit(RLIMIT_DATA);
J
Jiri Kosina 已提交
302 303
	if (rlim < RLIM_INFINITY && (brk - mm->start_brk) +
			(mm->end_data - mm->start_data) > rlim)
R
Ram Gupta 已提交
304 305
		goto out;

L
Linus Torvalds 已提交
306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
	newbrk = PAGE_ALIGN(brk);
	oldbrk = PAGE_ALIGN(mm->brk);
	if (oldbrk == newbrk)
		goto set_brk;

	/* Always allow shrinking brk. */
	if (brk <= mm->brk) {
		if (!do_munmap(mm, newbrk, oldbrk-newbrk))
			goto set_brk;
		goto out;
	}

	/* Check against existing mmap mappings. */
	if (find_vma_intersection(mm, oldbrk, newbrk+PAGE_SIZE))
		goto out;

	/* Ok, looks good - let it rip. */
	if (do_brk(oldbrk, newbrk-oldbrk) != oldbrk)
		goto out;
325

L
Linus Torvalds 已提交
326 327
set_brk:
	mm->brk = brk;
328 329 330 331 332 333
	populate = newbrk > oldbrk && (mm->def_flags & VM_LOCKED) != 0;
	up_write(&mm->mmap_sem);
	if (populate)
		mm_populate(oldbrk, newbrk - oldbrk);
	return brk;

L
Linus Torvalds 已提交
334 335 336 337 338 339
out:
	retval = mm->brk;
	up_write(&mm->mmap_sem);
	return retval;
}

340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
static long vma_compute_subtree_gap(struct vm_area_struct *vma)
{
	unsigned long max, subtree_gap;
	max = vma->vm_start;
	if (vma->vm_prev)
		max -= vma->vm_prev->vm_end;
	if (vma->vm_rb.rb_left) {
		subtree_gap = rb_entry(vma->vm_rb.rb_left,
				struct vm_area_struct, vm_rb)->rb_subtree_gap;
		if (subtree_gap > max)
			max = subtree_gap;
	}
	if (vma->vm_rb.rb_right) {
		subtree_gap = rb_entry(vma->vm_rb.rb_right,
				struct vm_area_struct, vm_rb)->rb_subtree_gap;
		if (subtree_gap > max)
			max = subtree_gap;
	}
	return max;
}

361
#ifdef CONFIG_DEBUG_VM_RB
L
Linus Torvalds 已提交
362 363
static int browse_rb(struct rb_root *root)
{
364
	int i = 0, j, bug = 0;
L
Linus Torvalds 已提交
365 366 367 368 369 370
	struct rb_node *nd, *pn = NULL;
	unsigned long prev = 0, pend = 0;

	for (nd = rb_first(root); nd; nd = rb_next(nd)) {
		struct vm_area_struct *vma;
		vma = rb_entry(nd, struct vm_area_struct, vm_rb);
371
		if (vma->vm_start < prev) {
372
			pr_emerg("vm_start %lx prev %lx\n", vma->vm_start, prev);
373 374 375
			bug = 1;
		}
		if (vma->vm_start < pend) {
376
			pr_emerg("vm_start %lx pend %lx\n", vma->vm_start, pend);
377 378 379
			bug = 1;
		}
		if (vma->vm_start > vma->vm_end) {
380
			pr_emerg("vm_end %lx < vm_start %lx\n",
381 382 383 384
				vma->vm_end, vma->vm_start);
			bug = 1;
		}
		if (vma->rb_subtree_gap != vma_compute_subtree_gap(vma)) {
385
			pr_emerg("free gap %lx, correct %lx\n",
386 387 388 389
			       vma->rb_subtree_gap,
			       vma_compute_subtree_gap(vma));
			bug = 1;
		}
L
Linus Torvalds 已提交
390 391
		i++;
		pn = nd;
D
David Miller 已提交
392 393
		prev = vma->vm_start;
		pend = vma->vm_end;
L
Linus Torvalds 已提交
394 395
	}
	j = 0;
396
	for (nd = pn; nd; nd = rb_prev(nd))
L
Linus Torvalds 已提交
397
		j++;
398
	if (i != j) {
399
		pr_emerg("backwards %d, forwards %d\n", j, i);
400
		bug = 1;
L
Linus Torvalds 已提交
401
	}
402
	return bug ? -1 : i;
L
Linus Torvalds 已提交
403 404
}

405 406 407 408 409 410 411 412 413
static void validate_mm_rb(struct rb_root *root, struct vm_area_struct *ignore)
{
	struct rb_node *nd;

	for (nd = rb_first(root); nd; nd = rb_next(nd)) {
		struct vm_area_struct *vma;
		vma = rb_entry(nd, struct vm_area_struct, vm_rb);
		BUG_ON(vma != ignore &&
		       vma->rb_subtree_gap != vma_compute_subtree_gap(vma));
L
Linus Torvalds 已提交
414 415 416
	}
}

417
static void validate_mm(struct mm_struct *mm)
L
Linus Torvalds 已提交
418 419 420
{
	int bug = 0;
	int i = 0;
421
	unsigned long highest_address = 0;
422 423 424
	struct vm_area_struct *vma = mm->mmap;
	while (vma) {
		struct anon_vma_chain *avc;
425
		vma_lock_anon_vma(vma);
426 427
		list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
			anon_vma_interval_tree_verify(avc);
428
		vma_unlock_anon_vma(vma);
429
		highest_address = vma->vm_end;
430
		vma = vma->vm_next;
L
Linus Torvalds 已提交
431 432
		i++;
	}
433
	if (i != mm->map_count) {
434
		pr_emerg("map_count %d vm_next %d\n", mm->map_count, i);
435 436 437
		bug = 1;
	}
	if (highest_address != mm->highest_vm_end) {
438
		pr_emerg("mm->highest_vm_end %lx, found %lx\n",
439 440 441
		       mm->highest_vm_end, highest_address);
		bug = 1;
	}
L
Linus Torvalds 已提交
442
	i = browse_rb(&mm->mm_rb);
443
	if (i != mm->map_count) {
444
		pr_emerg("map_count %d rb %d\n", mm->map_count, i);
445 446
		bug = 1;
	}
447
	BUG_ON(bug);
L
Linus Torvalds 已提交
448 449
}
#else
450
#define validate_mm_rb(root, ignore) do { } while (0)
L
Linus Torvalds 已提交
451 452 453
#define validate_mm(mm) do { } while (0)
#endif

454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
RB_DECLARE_CALLBACKS(static, vma_gap_callbacks, struct vm_area_struct, vm_rb,
		     unsigned long, rb_subtree_gap, vma_compute_subtree_gap)

/*
 * Update augmented rbtree rb_subtree_gap values after vma->vm_start or
 * vma->vm_prev->vm_end values changed, without modifying the vma's position
 * in the rbtree.
 */
static void vma_gap_update(struct vm_area_struct *vma)
{
	/*
	 * As it turns out, RB_DECLARE_CALLBACKS() already created a callback
	 * function that does exacltly what we want.
	 */
	vma_gap_callbacks_propagate(&vma->vm_rb, NULL);
}

static inline void vma_rb_insert(struct vm_area_struct *vma,
				 struct rb_root *root)
{
	/* All rb_subtree_gap values must be consistent prior to insertion */
	validate_mm_rb(root, NULL);

	rb_insert_augmented(&vma->vm_rb, root, &vma_gap_callbacks);
}

static void vma_rb_erase(struct vm_area_struct *vma, struct rb_root *root)
{
	/*
	 * All rb_subtree_gap values must be consistent prior to erase,
	 * with the possible exception of the vma being erased.
	 */
	validate_mm_rb(root, vma);

	/*
	 * Note rb_erase_augmented is a fairly large inline function,
	 * so make sure we instantiate it only once with our desired
	 * augmented rbtree callbacks.
	 */
	rb_erase_augmented(&vma->vm_rb, root, &vma_gap_callbacks);
}

496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527
/*
 * vma has some anon_vma assigned, and is already inserted on that
 * anon_vma's interval trees.
 *
 * Before updating the vma's vm_start / vm_end / vm_pgoff fields, the
 * vma must be removed from the anon_vma's interval trees using
 * anon_vma_interval_tree_pre_update_vma().
 *
 * After the update, the vma will be reinserted using
 * anon_vma_interval_tree_post_update_vma().
 *
 * The entire update must be protected by exclusive mmap_sem and by
 * the root anon_vma's mutex.
 */
static inline void
anon_vma_interval_tree_pre_update_vma(struct vm_area_struct *vma)
{
	struct anon_vma_chain *avc;

	list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
		anon_vma_interval_tree_remove(avc, &avc->anon_vma->rb_root);
}

static inline void
anon_vma_interval_tree_post_update_vma(struct vm_area_struct *vma)
{
	struct anon_vma_chain *avc;

	list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
		anon_vma_interval_tree_insert(avc, &avc->anon_vma->rb_root);
}

528 529 530
static int find_vma_links(struct mm_struct *mm, unsigned long addr,
		unsigned long end, struct vm_area_struct **pprev,
		struct rb_node ***rb_link, struct rb_node **rb_parent)
L
Linus Torvalds 已提交
531
{
532
	struct rb_node **__rb_link, *__rb_parent, *rb_prev;
L
Linus Torvalds 已提交
533 534 535 536 537 538 539 540 541 542 543

	__rb_link = &mm->mm_rb.rb_node;
	rb_prev = __rb_parent = NULL;

	while (*__rb_link) {
		struct vm_area_struct *vma_tmp;

		__rb_parent = *__rb_link;
		vma_tmp = rb_entry(__rb_parent, struct vm_area_struct, vm_rb);

		if (vma_tmp->vm_end > addr) {
544 545 546
			/* Fail if an existing vma overlaps the area */
			if (vma_tmp->vm_start < end)
				return -ENOMEM;
L
Linus Torvalds 已提交
547 548 549 550 551 552 553 554 555 556 557 558
			__rb_link = &__rb_parent->rb_left;
		} else {
			rb_prev = __rb_parent;
			__rb_link = &__rb_parent->rb_right;
		}
	}

	*pprev = NULL;
	if (rb_prev)
		*pprev = rb_entry(rb_prev, struct vm_area_struct, vm_rb);
	*rb_link = __rb_link;
	*rb_parent = __rb_parent;
559
	return 0;
L
Linus Torvalds 已提交
560 561
}

562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589
static unsigned long count_vma_pages_range(struct mm_struct *mm,
		unsigned long addr, unsigned long end)
{
	unsigned long nr_pages = 0;
	struct vm_area_struct *vma;

	/* Find first overlaping mapping */
	vma = find_vma_intersection(mm, addr, end);
	if (!vma)
		return 0;

	nr_pages = (min(end, vma->vm_end) -
		max(addr, vma->vm_start)) >> PAGE_SHIFT;

	/* Iterate over the rest of the overlaps */
	for (vma = vma->vm_next; vma; vma = vma->vm_next) {
		unsigned long overlap_len;

		if (vma->vm_start > end)
			break;

		overlap_len = min(end, vma->vm_end) - vma->vm_start;
		nr_pages += overlap_len >> PAGE_SHIFT;
	}

	return nr_pages;
}

L
Linus Torvalds 已提交
590 591 592
void __vma_link_rb(struct mm_struct *mm, struct vm_area_struct *vma,
		struct rb_node **rb_link, struct rb_node *rb_parent)
{
593 594 595 596 597 598 599 600 601 602 603 604 605 606 607
	/* Update tracking information for the gap following the new vma. */
	if (vma->vm_next)
		vma_gap_update(vma->vm_next);
	else
		mm->highest_vm_end = vma->vm_end;

	/*
	 * vma->vm_prev wasn't known when we followed the rbtree to find the
	 * correct insertion point for that vma. As a result, we could not
	 * update the vma vm_rb parents rb_subtree_gap values on the way down.
	 * So, we first insert the vma with a zero rb_subtree_gap value
	 * (to be consistent with what we did on the way down), and then
	 * immediately update the gap to the correct value. Finally we
	 * rebalance the rbtree after all augmented values have been set.
	 */
L
Linus Torvalds 已提交
608
	rb_link_node(&vma->vm_rb, rb_parent, rb_link);
609 610 611
	vma->rb_subtree_gap = 0;
	vma_gap_update(vma);
	vma_rb_insert(vma, &mm->mm_rb);
L
Linus Torvalds 已提交
612 613
}

D
Denys Vlasenko 已提交
614
static void __vma_link_file(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
615
{
Z
ZhenwenXu 已提交
616
	struct file *file;
L
Linus Torvalds 已提交
617 618 619 620 621 622

	file = vma->vm_file;
	if (file) {
		struct address_space *mapping = file->f_mapping;

		if (vma->vm_flags & VM_DENYWRITE)
A
Al Viro 已提交
623
			atomic_dec(&file_inode(file)->i_writecount);
L
Linus Torvalds 已提交
624
		if (vma->vm_flags & VM_SHARED)
625
			atomic_inc(&mapping->i_mmap_writable);
L
Linus Torvalds 已提交
626 627 628 629 630

		flush_dcache_mmap_lock(mapping);
		if (unlikely(vma->vm_flags & VM_NONLINEAR))
			vma_nonlinear_insert(vma, &mapping->i_mmap_nonlinear);
		else
631
			vma_interval_tree_insert(vma, &mapping->i_mmap);
L
Linus Torvalds 已提交
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
		flush_dcache_mmap_unlock(mapping);
	}
}

static void
__vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
	struct vm_area_struct *prev, struct rb_node **rb_link,
	struct rb_node *rb_parent)
{
	__vma_link_list(mm, vma, prev, rb_parent);
	__vma_link_rb(mm, vma, rb_link, rb_parent);
}

static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
			struct vm_area_struct *prev, struct rb_node **rb_link,
			struct rb_node *rb_parent)
{
	struct address_space *mapping = NULL;

651
	if (vma->vm_file) {
L
Linus Torvalds 已提交
652
		mapping = vma->vm_file->f_mapping;
653
		mutex_lock(&mapping->i_mmap_mutex);
654
	}
L
Linus Torvalds 已提交
655 656 657 658 659

	__vma_link(mm, vma, prev, rb_link, rb_parent);
	__vma_link_file(vma);

	if (mapping)
660
		mutex_unlock(&mapping->i_mmap_mutex);
L
Linus Torvalds 已提交
661 662 663 664 665 666

	mm->map_count++;
	validate_mm(mm);
}

/*
667
 * Helper for vma_adjust() in the split_vma insert case: insert a vma into the
668
 * mm's list and rbtree.  It has already been inserted into the interval tree.
L
Linus Torvalds 已提交
669
 */
Z
ZhenwenXu 已提交
670
static void __insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
671
{
672
	struct vm_area_struct *prev;
Z
ZhenwenXu 已提交
673
	struct rb_node **rb_link, *rb_parent;
L
Linus Torvalds 已提交
674

675 676 677
	if (find_vma_links(mm, vma->vm_start, vma->vm_end,
			   &prev, &rb_link, &rb_parent))
		BUG();
L
Linus Torvalds 已提交
678 679 680 681 682 683 684 685
	__vma_link(mm, vma, prev, rb_link, rb_parent);
	mm->map_count++;
}

static inline void
__vma_unlink(struct mm_struct *mm, struct vm_area_struct *vma,
		struct vm_area_struct *prev)
{
686
	struct vm_area_struct *next;
687

688 689
	vma_rb_erase(vma, &mm->mm_rb);
	prev->vm_next = next = vma->vm_next;
690 691
	if (next)
		next->vm_prev = prev;
D
Davidlohr Bueso 已提交
692 693 694

	/* Kill the cache */
	vmacache_invalidate(mm);
L
Linus Torvalds 已提交
695 696 697 698 699 700 701 702 703
}

/*
 * We cannot adjust vm_start, vm_end, vm_pgoff fields of a vma that
 * is already present in an i_mmap tree without adjusting the tree.
 * The following helper function should be used when such adjustments
 * are necessary.  The "insert" vma (if any) is to be inserted
 * before we drop the necessary locks.
 */
704
int vma_adjust(struct vm_area_struct *vma, unsigned long start,
L
Linus Torvalds 已提交
705 706 707 708 709 710
	unsigned long end, pgoff_t pgoff, struct vm_area_struct *insert)
{
	struct mm_struct *mm = vma->vm_mm;
	struct vm_area_struct *next = vma->vm_next;
	struct vm_area_struct *importer = NULL;
	struct address_space *mapping = NULL;
711
	struct rb_root *root = NULL;
712
	struct anon_vma *anon_vma = NULL;
L
Linus Torvalds 已提交
713
	struct file *file = vma->vm_file;
714
	bool start_changed = false, end_changed = false;
L
Linus Torvalds 已提交
715 716 717 718
	long adjust_next = 0;
	int remove_next = 0;

	if (next && !insert) {
719 720
		struct vm_area_struct *exporter = NULL;

L
Linus Torvalds 已提交
721 722 723 724 725 726 727
		if (end >= next->vm_end) {
			/*
			 * vma expands, overlapping all the next, and
			 * perhaps the one after too (mprotect case 6).
			 */
again:			remove_next = 1 + (end > next->vm_end);
			end = next->vm_end;
728
			exporter = next;
L
Linus Torvalds 已提交
729 730 731 732 733 734 735
			importer = vma;
		} else if (end > next->vm_start) {
			/*
			 * vma expands, overlapping part of the next:
			 * mprotect case 5 shifting the boundary up.
			 */
			adjust_next = (end - next->vm_start) >> PAGE_SHIFT;
736
			exporter = next;
L
Linus Torvalds 已提交
737 738 739 740 741 742 743
			importer = vma;
		} else if (end < vma->vm_end) {
			/*
			 * vma shrinks, and !insert tells it's not
			 * split_vma inserting another: so it must be
			 * mprotect case 4 shifting the boundary down.
			 */
V
vishnu.ps 已提交
744
			adjust_next = -((vma->vm_end - end) >> PAGE_SHIFT);
745
			exporter = vma;
L
Linus Torvalds 已提交
746 747 748
			importer = next;
		}

749 750 751 752 753
		/*
		 * Easily overlooked: when mprotect shifts the boundary,
		 * make sure the expanding vma has anon_vma set if the
		 * shrinking vma had, to cover any anon pages imported.
		 */
754 755
		if (exporter && exporter->anon_vma && !importer->anon_vma) {
			if (anon_vma_clone(importer, exporter))
756
				return -ENOMEM;
757
			importer->anon_vma = exporter->anon_vma;
758 759 760
		}
	}

L
Linus Torvalds 已提交
761 762
	if (file) {
		mapping = file->f_mapping;
763
		if (!(vma->vm_flags & VM_NONLINEAR)) {
L
Linus Torvalds 已提交
764
			root = &mapping->i_mmap;
765
			uprobe_munmap(vma, vma->vm_start, vma->vm_end);
766 767

			if (adjust_next)
768 769
				uprobe_munmap(next, next->vm_start,
							next->vm_end);
770 771
		}

772
		mutex_lock(&mapping->i_mmap_mutex);
L
Linus Torvalds 已提交
773 774
		if (insert) {
			/*
775
			 * Put into interval tree now, so instantiated pages
L
Linus Torvalds 已提交
776 777 778 779 780 781 782 783
			 * are visible to arm/parisc __flush_dcache_page
			 * throughout; but we cannot insert into address
			 * space until vma start or end is updated.
			 */
			__vma_link_file(insert);
		}
	}

784 785
	vma_adjust_trans_huge(vma, start, end, adjust_next);

786 787 788 789
	anon_vma = vma->anon_vma;
	if (!anon_vma && adjust_next)
		anon_vma = next->anon_vma;
	if (anon_vma) {
790 791
		VM_BUG_ON(adjust_next && next->anon_vma &&
			  anon_vma != next->anon_vma);
792
		anon_vma_lock_write(anon_vma);
793 794 795 796
		anon_vma_interval_tree_pre_update_vma(vma);
		if (adjust_next)
			anon_vma_interval_tree_pre_update_vma(next);
	}
797

L
Linus Torvalds 已提交
798 799
	if (root) {
		flush_dcache_mmap_lock(mapping);
800
		vma_interval_tree_remove(vma, root);
L
Linus Torvalds 已提交
801
		if (adjust_next)
802
			vma_interval_tree_remove(next, root);
L
Linus Torvalds 已提交
803 804
	}

805 806 807 808 809 810 811 812
	if (start != vma->vm_start) {
		vma->vm_start = start;
		start_changed = true;
	}
	if (end != vma->vm_end) {
		vma->vm_end = end;
		end_changed = true;
	}
L
Linus Torvalds 已提交
813 814 815 816 817 818 819 820
	vma->vm_pgoff = pgoff;
	if (adjust_next) {
		next->vm_start += adjust_next << PAGE_SHIFT;
		next->vm_pgoff += adjust_next;
	}

	if (root) {
		if (adjust_next)
821 822
			vma_interval_tree_insert(next, root);
		vma_interval_tree_insert(vma, root);
L
Linus Torvalds 已提交
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
		flush_dcache_mmap_unlock(mapping);
	}

	if (remove_next) {
		/*
		 * vma_merge has merged next into vma, and needs
		 * us to remove next before dropping the locks.
		 */
		__vma_unlink(mm, next, vma);
		if (file)
			__remove_shared_vm_struct(next, file, mapping);
	} else if (insert) {
		/*
		 * split_vma has split insert from vma, and needs
		 * us to insert it before dropping the locks
		 * (it may either follow vma or precede it).
		 */
		__insert_vm_struct(mm, insert);
841 842 843 844 845 846 847 848 849
	} else {
		if (start_changed)
			vma_gap_update(vma);
		if (end_changed) {
			if (!next)
				mm->highest_vm_end = end;
			else if (!adjust_next)
				vma_gap_update(next);
		}
L
Linus Torvalds 已提交
850 851
	}

852 853 854 855
	if (anon_vma) {
		anon_vma_interval_tree_post_update_vma(vma);
		if (adjust_next)
			anon_vma_interval_tree_post_update_vma(next);
856
		anon_vma_unlock_write(anon_vma);
857
	}
L
Linus Torvalds 已提交
858
	if (mapping)
859
		mutex_unlock(&mapping->i_mmap_mutex);
L
Linus Torvalds 已提交
860

861
	if (root) {
862
		uprobe_mmap(vma);
863 864

		if (adjust_next)
865
			uprobe_mmap(next);
866 867
	}

L
Linus Torvalds 已提交
868
	if (remove_next) {
M
Matt Helsley 已提交
869
		if (file) {
870
			uprobe_munmap(next, next->vm_start, next->vm_end);
L
Linus Torvalds 已提交
871
			fput(file);
M
Matt Helsley 已提交
872
		}
873 874
		if (next->anon_vma)
			anon_vma_merge(vma, next);
L
Linus Torvalds 已提交
875
		mm->map_count--;
876
		mpol_put(vma_policy(next));
L
Linus Torvalds 已提交
877 878 879 880 881 882
		kmem_cache_free(vm_area_cachep, next);
		/*
		 * In mprotect's case 6 (see comments on vma_merge),
		 * we must remove another next too. It would clutter
		 * up the code too much to do both in one go.
		 */
883 884
		next = vma->vm_next;
		if (remove_next == 2)
L
Linus Torvalds 已提交
885
			goto again;
886 887 888 889
		else if (next)
			vma_gap_update(next);
		else
			mm->highest_vm_end = end;
L
Linus Torvalds 已提交
890
	}
891
	if (insert && file)
892
		uprobe_mmap(insert);
L
Linus Torvalds 已提交
893 894

	validate_mm(mm);
895 896

	return 0;
L
Linus Torvalds 已提交
897 898 899 900 901 902 903 904 905
}

/*
 * If the vma has a ->close operation then the driver probably needs to release
 * per-vma resources, so we don't attempt to merge those.
 */
static inline int is_mergeable_vma(struct vm_area_struct *vma,
			struct file *file, unsigned long vm_flags)
{
906 907 908 909 910 911 912 913 914
	/*
	 * VM_SOFTDIRTY should not prevent from VMA merging, if we
	 * match the flags but dirty bit -- the caller should mark
	 * merged VMA as dirty. If dirty bit won't be excluded from
	 * comparison, we increase pressue on the memory system forcing
	 * the kernel to generate new VMAs when old one could be
	 * extended instead.
	 */
	if ((vma->vm_flags ^ vm_flags) & ~VM_SOFTDIRTY)
L
Linus Torvalds 已提交
915 916 917 918 919 920 921 922 923
		return 0;
	if (vma->vm_file != file)
		return 0;
	if (vma->vm_ops && vma->vm_ops->close)
		return 0;
	return 1;
}

static inline int is_mergeable_anon_vma(struct anon_vma *anon_vma1,
S
Shaohua Li 已提交
924 925
					struct anon_vma *anon_vma2,
					struct vm_area_struct *vma)
L
Linus Torvalds 已提交
926
{
S
Shaohua Li 已提交
927 928 929 930 931 932 933 934
	/*
	 * The list_is_singular() test is to avoid merging VMA cloned from
	 * parents. This can improve scalability caused by anon_vma lock.
	 */
	if ((!anon_vma1 || !anon_vma2) && (!vma ||
		list_is_singular(&vma->anon_vma_chain)))
		return 1;
	return anon_vma1 == anon_vma2;
L
Linus Torvalds 已提交
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952
}

/*
 * Return true if we can merge this (vm_flags,anon_vma,file,vm_pgoff)
 * in front of (at a lower virtual address and file offset than) the vma.
 *
 * We cannot merge two vmas if they have differently assigned (non-NULL)
 * anon_vmas, nor if same anon_vma is assigned but offsets incompatible.
 *
 * We don't check here for the merged mmap wrapping around the end of pagecache
 * indices (16TB on ia32) because do_mmap_pgoff() does not permit mmap's which
 * wrap, nor mmaps which cover the final page at index -1UL.
 */
static int
can_vma_merge_before(struct vm_area_struct *vma, unsigned long vm_flags,
	struct anon_vma *anon_vma, struct file *file, pgoff_t vm_pgoff)
{
	if (is_mergeable_vma(vma, file, vm_flags) &&
S
Shaohua Li 已提交
953
	    is_mergeable_anon_vma(anon_vma, vma->anon_vma, vma)) {
L
Linus Torvalds 已提交
954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
		if (vma->vm_pgoff == vm_pgoff)
			return 1;
	}
	return 0;
}

/*
 * Return true if we can merge this (vm_flags,anon_vma,file,vm_pgoff)
 * beyond (at a higher virtual address and file offset than) the vma.
 *
 * We cannot merge two vmas if they have differently assigned (non-NULL)
 * anon_vmas, nor if same anon_vma is assigned but offsets incompatible.
 */
static int
can_vma_merge_after(struct vm_area_struct *vma, unsigned long vm_flags,
	struct anon_vma *anon_vma, struct file *file, pgoff_t vm_pgoff)
{
	if (is_mergeable_vma(vma, file, vm_flags) &&
S
Shaohua Li 已提交
972
	    is_mergeable_anon_vma(anon_vma, vma->anon_vma, vma)) {
L
Linus Torvalds 已提交
973
		pgoff_t vm_pglen;
974
		vm_pglen = vma_pages(vma);
L
Linus Torvalds 已提交
975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012
		if (vma->vm_pgoff + vm_pglen == vm_pgoff)
			return 1;
	}
	return 0;
}

/*
 * Given a mapping request (addr,end,vm_flags,file,pgoff), figure out
 * whether that can be merged with its predecessor or its successor.
 * Or both (it neatly fills a hole).
 *
 * In most cases - when called for mmap, brk or mremap - [addr,end) is
 * certain not to be mapped by the time vma_merge is called; but when
 * called for mprotect, it is certain to be already mapped (either at
 * an offset within prev, or at the start of next), and the flags of
 * this area are about to be changed to vm_flags - and the no-change
 * case has already been eliminated.
 *
 * The following mprotect cases have to be considered, where AAAA is
 * the area passed down from mprotect_fixup, never extending beyond one
 * vma, PPPPPP is the prev vma specified, and NNNNNN the next vma after:
 *
 *     AAAA             AAAA                AAAA          AAAA
 *    PPPPPPNNNNNN    PPPPPPNNNNNN    PPPPPPNNNNNN    PPPPNNNNXXXX
 *    cannot merge    might become    might become    might become
 *                    PPNNNNNNNNNN    PPPPPPPPPPNN    PPPPPPPPPPPP 6 or
 *    mmap, brk or    case 4 below    case 5 below    PPPPPPPPXXXX 7 or
 *    mremap move:                                    PPPPNNNNNNNN 8
 *        AAAA
 *    PPPP    NNNN    PPPPPPPPPPPP    PPPPPPPPNNNN    PPPPNNNNNNNN
 *    might become    case 1 below    case 2 below    case 3 below
 *
 * Odd one out? Case 8, because it extends NNNN but needs flags of XXXX:
 * mprotect_fixup updates vm_flags & vm_page_prot on successful return.
 */
struct vm_area_struct *vma_merge(struct mm_struct *mm,
			struct vm_area_struct *prev, unsigned long addr,
			unsigned long end, unsigned long vm_flags,
V
vishnu.ps 已提交
1013
			struct anon_vma *anon_vma, struct file *file,
L
Linus Torvalds 已提交
1014 1015 1016 1017
			pgoff_t pgoff, struct mempolicy *policy)
{
	pgoff_t pglen = (end - addr) >> PAGE_SHIFT;
	struct vm_area_struct *area, *next;
1018
	int err;
L
Linus Torvalds 已提交
1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038

	/*
	 * We later require that vma->vm_flags == vm_flags,
	 * so this tests vma->vm_flags & VM_SPECIAL, too.
	 */
	if (vm_flags & VM_SPECIAL)
		return NULL;

	if (prev)
		next = prev->vm_next;
	else
		next = mm->mmap;
	area = next;
	if (next && next->vm_end == end)		/* cases 6, 7, 8 */
		next = next->vm_next;

	/*
	 * Can it merge with the predecessor?
	 */
	if (prev && prev->vm_end == addr &&
V
vishnu.ps 已提交
1039
			mpol_equal(vma_policy(prev), policy) &&
L
Linus Torvalds 已提交
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
			can_vma_merge_after(prev, vm_flags,
						anon_vma, file, pgoff)) {
		/*
		 * OK, it can.  Can we now merge in the successor as well?
		 */
		if (next && end == next->vm_start &&
				mpol_equal(policy, vma_policy(next)) &&
				can_vma_merge_before(next, vm_flags,
					anon_vma, file, pgoff+pglen) &&
				is_mergeable_anon_vma(prev->anon_vma,
S
Shaohua Li 已提交
1050
						      next->anon_vma, NULL)) {
L
Linus Torvalds 已提交
1051
							/* cases 1, 6 */
1052
			err = vma_adjust(prev, prev->vm_start,
L
Linus Torvalds 已提交
1053 1054
				next->vm_end, prev->vm_pgoff, NULL);
		} else					/* cases 2, 5, 7 */
1055
			err = vma_adjust(prev, prev->vm_start,
L
Linus Torvalds 已提交
1056
				end, prev->vm_pgoff, NULL);
1057 1058
		if (err)
			return NULL;
A
Andrea Arcangeli 已提交
1059
		khugepaged_enter_vma_merge(prev);
L
Linus Torvalds 已提交
1060 1061 1062 1063 1064 1065 1066
		return prev;
	}

	/*
	 * Can this new request be merged in front of next?
	 */
	if (next && end == next->vm_start &&
V
vishnu.ps 已提交
1067
			mpol_equal(policy, vma_policy(next)) &&
L
Linus Torvalds 已提交
1068 1069 1070
			can_vma_merge_before(next, vm_flags,
					anon_vma, file, pgoff+pglen)) {
		if (prev && addr < prev->vm_end)	/* case 4 */
1071
			err = vma_adjust(prev, prev->vm_start,
L
Linus Torvalds 已提交
1072 1073
				addr, prev->vm_pgoff, NULL);
		else					/* cases 3, 8 */
1074
			err = vma_adjust(area, addr, next->vm_end,
L
Linus Torvalds 已提交
1075
				next->vm_pgoff - pglen, NULL);
1076 1077
		if (err)
			return NULL;
A
Andrea Arcangeli 已提交
1078
		khugepaged_enter_vma_merge(area);
L
Linus Torvalds 已提交
1079 1080 1081 1082 1083 1084
		return area;
	}

	return NULL;
}

1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
/*
 * Rough compatbility check to quickly see if it's even worth looking
 * at sharing an anon_vma.
 *
 * They need to have the same vm_file, and the flags can only differ
 * in things that mprotect may change.
 *
 * NOTE! The fact that we share an anon_vma doesn't _have_ to mean that
 * we can merge the two vma's. For example, we refuse to merge a vma if
 * there is a vm_ops->close() function, because that indicates that the
 * driver is doing some kind of reference counting. But that doesn't
 * really matter for the anon_vma sharing case.
 */
static int anon_vma_compatible(struct vm_area_struct *a, struct vm_area_struct *b)
{
	return a->vm_end == b->vm_start &&
		mpol_equal(vma_policy(a), vma_policy(b)) &&
		a->vm_file == b->vm_file &&
1103
		!((a->vm_flags ^ b->vm_flags) & ~(VM_READ|VM_WRITE|VM_EXEC|VM_SOFTDIRTY)) &&
1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
		b->vm_pgoff == a->vm_pgoff + ((b->vm_start - a->vm_start) >> PAGE_SHIFT);
}

/*
 * Do some basic sanity checking to see if we can re-use the anon_vma
 * from 'old'. The 'a'/'b' vma's are in VM order - one of them will be
 * the same as 'old', the other will be the new one that is trying
 * to share the anon_vma.
 *
 * NOTE! This runs with mm_sem held for reading, so it is possible that
 * the anon_vma of 'old' is concurrently in the process of being set up
 * by another page fault trying to merge _that_. But that's ok: if it
 * is being set up, that automatically means that it will be a singleton
 * acceptable for merging, so we can do all of this optimistically. But
 * we do that ACCESS_ONCE() to make sure that we never re-load the pointer.
 *
 * IOW: that the "list_is_singular()" test on the anon_vma_chain only
 * matters for the 'stable anon_vma' case (ie the thing we want to avoid
 * is to return an anon_vma that is "complex" due to having gone through
 * a fork).
 *
 * We also make sure that the two vma's are compatible (adjacent,
 * and with the same memory policies). That's all stable, even with just
 * a read lock on the mm_sem.
 */
static struct anon_vma *reusable_anon_vma(struct vm_area_struct *old, struct vm_area_struct *a, struct vm_area_struct *b)
{
	if (anon_vma_compatible(a, b)) {
		struct anon_vma *anon_vma = ACCESS_ONCE(old->anon_vma);

		if (anon_vma && list_is_singular(&old->anon_vma_chain))
			return anon_vma;
	}
	return NULL;
}

L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149
/*
 * find_mergeable_anon_vma is used by anon_vma_prepare, to check
 * neighbouring vmas for a suitable anon_vma, before it goes off
 * to allocate a new anon_vma.  It checks because a repetitive
 * sequence of mprotects and faults may otherwise lead to distinct
 * anon_vmas being allocated, preventing vma merge in subsequent
 * mprotect.
 */
struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *vma)
{
1150
	struct anon_vma *anon_vma;
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156
	struct vm_area_struct *near;

	near = vma->vm_next;
	if (!near)
		goto try_prev;

1157 1158 1159
	anon_vma = reusable_anon_vma(near, vma, near);
	if (anon_vma)
		return anon_vma;
L
Linus Torvalds 已提交
1160
try_prev:
1161
	near = vma->vm_prev;
L
Linus Torvalds 已提交
1162 1163 1164
	if (!near)
		goto none;

1165 1166 1167
	anon_vma = reusable_anon_vma(near, near, vma);
	if (anon_vma)
		return anon_vma;
L
Linus Torvalds 已提交
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
none:
	/*
	 * There's no absolute need to look only at touching neighbours:
	 * we could search further afield for "compatible" anon_vmas.
	 * But it would probably just be a waste of time searching,
	 * or lead to too many vmas hanging off the same anon_vma.
	 * We're trying to allow mprotect remerging later on,
	 * not trying to minimize memory used for anon_vmas.
	 */
	return NULL;
}

#ifdef CONFIG_PROC_FS
1181
void vm_stat_account(struct mm_struct *mm, unsigned long flags,
L
Linus Torvalds 已提交
1182 1183 1184 1185 1186
						struct file *file, long pages)
{
	const unsigned long stack_flags
		= VM_STACK_FLAGS & (VM_GROWSUP|VM_GROWSDOWN);

1187 1188
	mm->total_vm += pages;

L
Linus Torvalds 已提交
1189 1190 1191 1192 1193 1194 1195 1196 1197
	if (file) {
		mm->shared_vm += pages;
		if ((flags & (VM_EXEC|VM_WRITE)) == VM_EXEC)
			mm->exec_vm += pages;
	} else if (flags & stack_flags)
		mm->stack_vm += pages;
}
#endif /* CONFIG_PROC_FS */

A
Al Viro 已提交
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
/*
 * If a hint addr is less than mmap_min_addr change hint to be as
 * low as possible but still greater than mmap_min_addr
 */
static inline unsigned long round_hint_to_min(unsigned long hint)
{
	hint &= PAGE_MASK;
	if (((void *)hint != NULL) &&
	    (hint < mmap_min_addr))
		return PAGE_ALIGN(mmap_min_addr);
	return hint;
}

1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
static inline int mlock_future_check(struct mm_struct *mm,
				     unsigned long flags,
				     unsigned long len)
{
	unsigned long locked, lock_limit;

	/*  mlock MCL_FUTURE? */
	if (flags & VM_LOCKED) {
		locked = len >> PAGE_SHIFT;
		locked += mm->locked_vm;
		lock_limit = rlimit(RLIMIT_MEMLOCK);
		lock_limit >>= PAGE_SHIFT;
		if (locked > lock_limit && !capable(CAP_IPC_LOCK))
			return -EAGAIN;
	}
	return 0;
}

L
Linus Torvalds 已提交
1229
/*
1230
 * The caller must hold down_write(&current->mm->mmap_sem).
L
Linus Torvalds 已提交
1231 1232
 */

1233
unsigned long do_mmap_pgoff(struct file *file, unsigned long addr,
L
Linus Torvalds 已提交
1234
			unsigned long len, unsigned long prot,
1235
			unsigned long flags, unsigned long pgoff,
1236
			unsigned long *populate)
L
Linus Torvalds 已提交
1237
{
V
vishnu.ps 已提交
1238
	struct mm_struct *mm = current->mm;
1239
	vm_flags_t vm_flags;
L
Linus Torvalds 已提交
1240

1241
	*populate = 0;
1242

L
Linus Torvalds 已提交
1243 1244 1245 1246 1247 1248 1249
	/*
	 * Does the application expect PROT_READ to imply PROT_EXEC?
	 *
	 * (the exception is when the underlying filesystem is noexec
	 *  mounted, in which case we dont add PROT_EXEC.)
	 */
	if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC))
1250
		if (!(file && (file->f_path.mnt->mnt_flags & MNT_NOEXEC)))
L
Linus Torvalds 已提交
1251 1252 1253 1254 1255
			prot |= PROT_EXEC;

	if (!len)
		return -EINVAL;

1256 1257 1258
	if (!(flags & MAP_FIXED))
		addr = round_hint_to_min(addr);

L
Linus Torvalds 已提交
1259 1260
	/* Careful about overflows.. */
	len = PAGE_ALIGN(len);
1261
	if (!len)
L
Linus Torvalds 已提交
1262 1263 1264 1265
		return -ENOMEM;

	/* offset overflow? */
	if ((pgoff + (len >> PAGE_SHIFT)) < pgoff)
V
vishnu.ps 已提交
1266
		return -EOVERFLOW;
L
Linus Torvalds 已提交
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285

	/* Too many mappings? */
	if (mm->map_count > sysctl_max_map_count)
		return -ENOMEM;

	/* Obtain the address to map to. we verify (or select) it and ensure
	 * that it represents a valid section of the address space.
	 */
	addr = get_unmapped_area(file, addr, len, pgoff, flags);
	if (addr & ~PAGE_MASK)
		return addr;

	/* Do simple checking here so the lower-level routines won't have
	 * to. we assume access permissions have been handled by the open
	 * of the memory object, so we don't do any here.
	 */
	vm_flags = calc_vm_prot_bits(prot) | calc_vm_flag_bits(flags) |
			mm->def_flags | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC;

H
Huang Shijie 已提交
1286
	if (flags & MAP_LOCKED)
L
Linus Torvalds 已提交
1287 1288
		if (!can_do_mlock())
			return -EPERM;
1289

1290 1291
	if (mlock_future_check(mm, vm_flags, len))
		return -EAGAIN;
L
Linus Torvalds 已提交
1292 1293

	if (file) {
1294 1295
		struct inode *inode = file_inode(file);

L
Linus Torvalds 已提交
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
		switch (flags & MAP_TYPE) {
		case MAP_SHARED:
			if ((prot&PROT_WRITE) && !(file->f_mode&FMODE_WRITE))
				return -EACCES;

			/*
			 * Make sure we don't allow writing to an append-only
			 * file..
			 */
			if (IS_APPEND(inode) && (file->f_mode & FMODE_WRITE))
				return -EACCES;

			/*
			 * Make sure there are no mandatory locks on the file.
			 */
1311
			if (locks_verify_locked(file))
L
Linus Torvalds 已提交
1312 1313 1314 1315 1316 1317 1318 1319 1320 1321
				return -EAGAIN;

			vm_flags |= VM_SHARED | VM_MAYSHARE;
			if (!(file->f_mode & FMODE_WRITE))
				vm_flags &= ~(VM_MAYWRITE | VM_SHARED);

			/* fall through */
		case MAP_PRIVATE:
			if (!(file->f_mode & FMODE_READ))
				return -EACCES;
1322
			if (file->f_path.mnt->mnt_flags & MNT_NOEXEC) {
L
Linus Torvalds 已提交
1323 1324 1325 1326 1327
				if (vm_flags & VM_EXEC)
					return -EPERM;
				vm_flags &= ~VM_MAYEXEC;
			}

A
Al Viro 已提交
1328
			if (!file->f_op->mmap)
L
Linus Torvalds 已提交
1329
				return -ENODEV;
1330 1331
			if (vm_flags & (VM_GROWSDOWN|VM_GROWSUP))
				return -EINVAL;
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337 1338 1339
			break;

		default:
			return -EINVAL;
		}
	} else {
		switch (flags & MAP_TYPE) {
		case MAP_SHARED:
1340 1341
			if (vm_flags & (VM_GROWSDOWN|VM_GROWSUP))
				return -EINVAL;
1342 1343 1344 1345
			/*
			 * Ignore pgoff.
			 */
			pgoff = 0;
L
Linus Torvalds 已提交
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
			vm_flags |= VM_SHARED | VM_MAYSHARE;
			break;
		case MAP_PRIVATE:
			/*
			 * Set pgoff according to addr for anon_vma.
			 */
			pgoff = addr >> PAGE_SHIFT;
			break;
		default:
			return -EINVAL;
		}
	}

1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
	/*
	 * Set 'VM_NORESERVE' if we should not account for the
	 * memory use of this mapping.
	 */
	if (flags & MAP_NORESERVE) {
		/* We honor MAP_NORESERVE if allowed to overcommit */
		if (sysctl_overcommit_memory != OVERCOMMIT_NEVER)
			vm_flags |= VM_NORESERVE;

		/* hugetlb applies strict overcommit unless MAP_NORESERVE */
		if (file && is_file_hugepages(file))
			vm_flags |= VM_NORESERVE;
	}

	addr = mmap_region(file, addr, len, vm_flags, pgoff);
1374 1375 1376
	if (!IS_ERR_VALUE(addr) &&
	    ((vm_flags & VM_LOCKED) ||
	     (flags & (MAP_POPULATE | MAP_NONBLOCK)) == MAP_POPULATE))
1377
		*populate = len;
1378
	return addr;
M
Miklos Szeredi 已提交
1379
}
1380

H
Hugh Dickins 已提交
1381 1382 1383 1384 1385 1386 1387 1388
SYSCALL_DEFINE6(mmap_pgoff, unsigned long, addr, unsigned long, len,
		unsigned long, prot, unsigned long, flags,
		unsigned long, fd, unsigned long, pgoff)
{
	struct file *file = NULL;
	unsigned long retval = -EBADF;

	if (!(flags & MAP_ANONYMOUS)) {
A
Al Viro 已提交
1389
		audit_mmap_fd(fd, flags);
H
Hugh Dickins 已提交
1390 1391 1392
		file = fget(fd);
		if (!file)
			goto out;
1393 1394
		if (is_file_hugepages(file))
			len = ALIGN(len, huge_page_size(hstate_file(file)));
1395 1396 1397
		retval = -EINVAL;
		if (unlikely(flags & MAP_HUGETLB && !is_file_hugepages(file)))
			goto out_fput;
H
Hugh Dickins 已提交
1398 1399
	} else if (flags & MAP_HUGETLB) {
		struct user_struct *user = NULL;
1400
		struct hstate *hs;
1401

1402
		hs = hstate_sizelog((flags >> MAP_HUGE_SHIFT) & SHM_HUGE_MASK);
1403 1404 1405 1406
		if (!hs)
			return -EINVAL;

		len = ALIGN(len, huge_page_size(hs));
H
Hugh Dickins 已提交
1407 1408 1409 1410 1411 1412
		/*
		 * VM_NORESERVE is used because the reservations will be
		 * taken when vm_ops->mmap() is called
		 * A dummy user value is used because we are not locking
		 * memory so no accounting is necessary
		 */
1413
		file = hugetlb_file_setup(HUGETLB_ANON_FILE, len,
1414 1415 1416
				VM_NORESERVE,
				&user, HUGETLB_ANONHUGE_INODE,
				(flags >> MAP_HUGE_SHIFT) & MAP_HUGE_MASK);
H
Hugh Dickins 已提交
1417 1418 1419 1420 1421 1422
		if (IS_ERR(file))
			return PTR_ERR(file);
	}

	flags &= ~(MAP_EXECUTABLE | MAP_DENYWRITE);

A
Al Viro 已提交
1423
	retval = vm_mmap_pgoff(file, addr, len, prot, flags, pgoff);
1424
out_fput:
H
Hugh Dickins 已提交
1425 1426 1427 1428 1429 1430
	if (file)
		fput(file);
out:
	return retval;
}

C
Christoph Hellwig 已提交
1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
#ifdef __ARCH_WANT_SYS_OLD_MMAP
struct mmap_arg_struct {
	unsigned long addr;
	unsigned long len;
	unsigned long prot;
	unsigned long flags;
	unsigned long fd;
	unsigned long offset;
};

SYSCALL_DEFINE1(old_mmap, struct mmap_arg_struct __user *, arg)
{
	struct mmap_arg_struct a;

	if (copy_from_user(&a, arg, sizeof(a)))
		return -EFAULT;
	if (a.offset & ~PAGE_MASK)
		return -EINVAL;

	return sys_mmap_pgoff(a.addr, a.len, a.prot, a.flags, a.fd,
			      a.offset >> PAGE_SHIFT);
}
#endif /* __ARCH_WANT_SYS_OLD_MMAP */

A
Alexey Dobriyan 已提交
1455 1456 1457 1458 1459 1460 1461 1462
/*
 * Some shared mappigns will want the pages marked read-only
 * to track write events. If so, we'll downgrade vm_page_prot
 * to the private version (using protection_map[] without the
 * VM_SHARED bit).
 */
int vma_wants_writenotify(struct vm_area_struct *vma)
{
1463
	vm_flags_t vm_flags = vma->vm_flags;
A
Alexey Dobriyan 已提交
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474

	/* If it was private or non-writable, the write bit is already clear */
	if ((vm_flags & (VM_WRITE|VM_SHARED)) != ((VM_WRITE|VM_SHARED)))
		return 0;

	/* The backer wishes to know when pages are first written to? */
	if (vma->vm_ops && vma->vm_ops->page_mkwrite)
		return 1;

	/* The open routine did something to the protections already? */
	if (pgprot_val(vma->vm_page_prot) !=
1475
	    pgprot_val(vm_get_page_prot(vm_flags)))
A
Alexey Dobriyan 已提交
1476 1477 1478
		return 0;

	/* Specialty mapping? */
1479
	if (vm_flags & VM_PFNMAP)
A
Alexey Dobriyan 已提交
1480 1481 1482 1483 1484 1485 1486
		return 0;

	/* Can the mapping track the dirty pages? */
	return vma->vm_file && vma->vm_file->f_mapping &&
		mapping_cap_account_dirty(vma->vm_file->f_mapping);
}

1487 1488
/*
 * We account for memory if it's a private writeable mapping,
1489
 * not hugepages and VM_NORESERVE wasn't set.
1490
 */
1491
static inline int accountable_mapping(struct file *file, vm_flags_t vm_flags)
1492
{
1493 1494 1495 1496 1497 1498 1499
	/*
	 * hugetlb has its own accounting separate from the core VM
	 * VM_HUGETLB may not be set yet so we cannot check for that flag.
	 */
	if (file && is_file_hugepages(file))
		return 0;

1500 1501 1502
	return (vm_flags & (VM_NORESERVE | VM_SHARED | VM_WRITE)) == VM_WRITE;
}

M
Miklos Szeredi 已提交
1503
unsigned long mmap_region(struct file *file, unsigned long addr,
1504
		unsigned long len, vm_flags_t vm_flags, unsigned long pgoff)
M
Miklos Szeredi 已提交
1505 1506 1507 1508 1509 1510 1511
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma, *prev;
	int error;
	struct rb_node **rb_link, *rb_parent;
	unsigned long charged = 0;

1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528
	/* Check against address space limit. */
	if (!may_expand_vm(mm, len >> PAGE_SHIFT)) {
		unsigned long nr_pages;

		/*
		 * MAP_FIXED may remove pages of mappings that intersects with
		 * requested mapping. Account for the pages it would unmap.
		 */
		if (!(vm_flags & MAP_FIXED))
			return -ENOMEM;

		nr_pages = count_vma_pages_range(mm, addr, addr + len);

		if (!may_expand_vm(mm, (len >> PAGE_SHIFT) - nr_pages))
			return -ENOMEM;
	}

L
Linus Torvalds 已提交
1529 1530 1531
	/* Clear old maps */
	error = -ENOMEM;
munmap_back:
1532
	if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent)) {
L
Linus Torvalds 已提交
1533 1534 1535 1536 1537
		if (do_munmap(mm, addr, len))
			return -ENOMEM;
		goto munmap_back;
	}

1538 1539 1540
	/*
	 * Private writable mapping: check memory availability
	 */
1541
	if (accountable_mapping(file, vm_flags)) {
1542
		charged = len >> PAGE_SHIFT;
1543
		if (security_vm_enough_memory_mm(mm, charged))
1544 1545
			return -ENOMEM;
		vm_flags |= VM_ACCOUNT;
L
Linus Torvalds 已提交
1546 1547 1548
	}

	/*
1549
	 * Can we just expand an old mapping?
L
Linus Torvalds 已提交
1550
	 */
1551 1552 1553
	vma = vma_merge(mm, prev, addr, addr + len, vm_flags, NULL, file, pgoff, NULL);
	if (vma)
		goto out;
L
Linus Torvalds 已提交
1554 1555 1556 1557 1558 1559

	/*
	 * Determine the object being mapped and call the appropriate
	 * specific mapper. the address has already been validated, but
	 * not unmapped, but the maps are removed from the list.
	 */
P
Pekka Enberg 已提交
1560
	vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
1561 1562 1563 1564 1565 1566 1567 1568 1569
	if (!vma) {
		error = -ENOMEM;
		goto unacct_error;
	}

	vma->vm_mm = mm;
	vma->vm_start = addr;
	vma->vm_end = addr + len;
	vma->vm_flags = vm_flags;
1570
	vma->vm_page_prot = vm_get_page_prot(vm_flags);
L
Linus Torvalds 已提交
1571
	vma->vm_pgoff = pgoff;
1572
	INIT_LIST_HEAD(&vma->anon_vma_chain);
L
Linus Torvalds 已提交
1573 1574 1575 1576 1577 1578 1579

	if (file) {
		if (vm_flags & VM_DENYWRITE) {
			error = deny_write_access(file);
			if (error)
				goto free_vma;
		}
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
		if (vm_flags & VM_SHARED) {
			error = mapping_map_writable(file->f_mapping);
			if (error)
				goto allow_write_and_free_vma;
		}

		/* ->mmap() can change vma->vm_file, but must guarantee that
		 * vma_link() below can deny write-access if VM_DENYWRITE is set
		 * and map writably if VM_SHARED is set. This usually means the
		 * new file must not have been exposed to user-space, yet.
		 */
A
Al Viro 已提交
1591
		vma->vm_file = get_file(file);
L
Linus Torvalds 已提交
1592 1593 1594
		error = file->f_op->mmap(file, vma);
		if (error)
			goto unmap_and_free_vma;
1595 1596 1597 1598 1599

		/* Can addr have changed??
		 *
		 * Answer: Yes, several device drivers can do it in their
		 *         f_op->mmap method. -DaveM
1600 1601
		 * Bug: If addr is changed, prev, rb_link, rb_parent should
		 *      be updated for vma_link()
1602
		 */
1603 1604
		WARN_ON_ONCE(addr != vma->vm_start);

1605 1606
		addr = vma->vm_start;
		vm_flags = vma->vm_flags;
L
Linus Torvalds 已提交
1607 1608 1609 1610 1611 1612
	} else if (vm_flags & VM_SHARED) {
		error = shmem_zero_setup(vma);
		if (error)
			goto free_vma;
	}

1613 1614 1615 1616 1617 1618 1619 1620 1621 1622
	if (vma_wants_writenotify(vma)) {
		pgprot_t pprot = vma->vm_page_prot;

		/* Can vma->vm_page_prot have changed??
		 *
		 * Answer: Yes, drivers may have changed it in their
		 *         f_op->mmap method.
		 *
		 * Ensures that vmas marked as uncached stay that way.
		 */
1623
		vma->vm_page_prot = vm_get_page_prot(vm_flags & ~VM_SHARED);
1624 1625 1626
		if (pgprot_val(pprot) == pgprot_val(pgprot_noncached(pprot)))
			vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
	}
1627

1628
	vma_link(mm, vma, prev, rb_link, rb_parent);
1629
	/* Once vma denies write, undo our temporary denial count */
1630 1631 1632 1633 1634 1635
	if (file) {
		if (vm_flags & VM_SHARED)
			mapping_unmap_writable(file->f_mapping);
		if (vm_flags & VM_DENYWRITE)
			allow_write_access(file);
	}
1636
	file = vma->vm_file;
1637
out:
1638
	perf_event_mmap(vma);
1639

1640
	vm_stat_account(mm, vm_flags, file, len >> PAGE_SHIFT);
L
Linus Torvalds 已提交
1641
	if (vm_flags & VM_LOCKED) {
1642 1643
		if (!((vm_flags & VM_SPECIAL) || is_vm_hugetlb_page(vma) ||
					vma == get_gate_vma(current->mm)))
1644
			mm->locked_vm += (len >> PAGE_SHIFT);
1645 1646 1647
		else
			vma->vm_flags &= ~VM_LOCKED;
	}
1648

1649 1650
	if (file)
		uprobe_mmap(vma);
1651

1652 1653 1654 1655 1656 1657 1658 1659 1660
	/*
	 * New (or expanded) vma always get soft dirty status.
	 * Otherwise user-space soft-dirty page tracker won't
	 * be able to distinguish situation when vma area unmapped,
	 * then new mapped in-place (which must be aimed as
	 * a completely new data area).
	 */
	vma->vm_flags |= VM_SOFTDIRTY;

L
Linus Torvalds 已提交
1661 1662 1663 1664 1665 1666 1667
	return addr;

unmap_and_free_vma:
	vma->vm_file = NULL;
	fput(file);

	/* Undo any partial mapping done by a device driver. */
1668 1669
	unmap_region(mm, vma, prev, vma->vm_start, vma->vm_end);
	charged = 0;
1670 1671 1672 1673 1674
	if (vm_flags & VM_SHARED)
		mapping_unmap_writable(file->f_mapping);
allow_write_and_free_vma:
	if (vm_flags & VM_DENYWRITE)
		allow_write_access(file);
L
Linus Torvalds 已提交
1675 1676 1677 1678 1679 1680 1681 1682
free_vma:
	kmem_cache_free(vm_area_cachep, vma);
unacct_error:
	if (charged)
		vm_unacct_memory(charged);
	return error;
}

1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
unsigned long unmapped_area(struct vm_unmapped_area_info *info)
{
	/*
	 * We implement the search by looking for an rbtree node that
	 * immediately follows a suitable gap. That is,
	 * - gap_start = vma->vm_prev->vm_end <= info->high_limit - length;
	 * - gap_end   = vma->vm_start        >= info->low_limit  + length;
	 * - gap_end - gap_start >= length
	 */

	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	unsigned long length, low_limit, high_limit, gap_start, gap_end;

	/* Adjust search length to account for worst case alignment overhead */
	length = info->length + info->align_mask;
	if (length < info->length)
		return -ENOMEM;

	/* Adjust search limits by the desired length */
	if (info->high_limit < length)
		return -ENOMEM;
	high_limit = info->high_limit - length;

	if (info->low_limit > high_limit)
		return -ENOMEM;
	low_limit = info->low_limit + length;

	/* Check if rbtree root looks promising */
	if (RB_EMPTY_ROOT(&mm->mm_rb))
		goto check_highest;
	vma = rb_entry(mm->mm_rb.rb_node, struct vm_area_struct, vm_rb);
	if (vma->rb_subtree_gap < length)
		goto check_highest;

	while (true) {
		/* Visit left subtree if it looks promising */
		gap_end = vma->vm_start;
		if (gap_end >= low_limit && vma->vm_rb.rb_left) {
			struct vm_area_struct *left =
				rb_entry(vma->vm_rb.rb_left,
					 struct vm_area_struct, vm_rb);
			if (left->rb_subtree_gap >= length) {
				vma = left;
				continue;
			}
		}

		gap_start = vma->vm_prev ? vma->vm_prev->vm_end : 0;
check_current:
		/* Check if current node has a suitable gap */
		if (gap_start > high_limit)
			return -ENOMEM;
		if (gap_end >= low_limit && gap_end - gap_start >= length)
			goto found;

		/* Visit right subtree if it looks promising */
		if (vma->vm_rb.rb_right) {
			struct vm_area_struct *right =
				rb_entry(vma->vm_rb.rb_right,
					 struct vm_area_struct, vm_rb);
			if (right->rb_subtree_gap >= length) {
				vma = right;
				continue;
			}
		}

		/* Go back up the rbtree to find next candidate node */
		while (true) {
			struct rb_node *prev = &vma->vm_rb;
			if (!rb_parent(prev))
				goto check_highest;
			vma = rb_entry(rb_parent(prev),
				       struct vm_area_struct, vm_rb);
			if (prev == vma->vm_rb.rb_left) {
				gap_start = vma->vm_prev->vm_end;
				gap_end = vma->vm_start;
				goto check_current;
			}
		}
	}

check_highest:
	/* Check highest gap, which does not precede any rbtree node */
	gap_start = mm->highest_vm_end;
	gap_end = ULONG_MAX;  /* Only for VM_BUG_ON below */
	if (gap_start > high_limit)
		return -ENOMEM;

found:
	/* We found a suitable gap. Clip it with the original low_limit. */
	if (gap_start < info->low_limit)
		gap_start = info->low_limit;

	/* Adjust gap address to the desired alignment */
	gap_start += (info->align_offset - gap_start) & info->align_mask;

	VM_BUG_ON(gap_start + info->length > info->high_limit);
	VM_BUG_ON(gap_start + info->length > gap_end);
	return gap_start;
}

unsigned long unmapped_area_topdown(struct vm_unmapped_area_info *info)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	unsigned long length, low_limit, high_limit, gap_start, gap_end;

	/* Adjust search length to account for worst case alignment overhead */
	length = info->length + info->align_mask;
	if (length < info->length)
		return -ENOMEM;

	/*
	 * Adjust search limits by the desired length.
	 * See implementation comment at top of unmapped_area().
	 */
	gap_end = info->high_limit;
	if (gap_end < length)
		return -ENOMEM;
	high_limit = gap_end - length;

	if (info->low_limit > high_limit)
		return -ENOMEM;
	low_limit = info->low_limit + length;

	/* Check highest gap, which does not precede any rbtree node */
	gap_start = mm->highest_vm_end;
	if (gap_start <= high_limit)
		goto found_highest;

	/* Check if rbtree root looks promising */
	if (RB_EMPTY_ROOT(&mm->mm_rb))
		return -ENOMEM;
	vma = rb_entry(mm->mm_rb.rb_node, struct vm_area_struct, vm_rb);
	if (vma->rb_subtree_gap < length)
		return -ENOMEM;

	while (true) {
		/* Visit right subtree if it looks promising */
		gap_start = vma->vm_prev ? vma->vm_prev->vm_end : 0;
		if (gap_start <= high_limit && vma->vm_rb.rb_right) {
			struct vm_area_struct *right =
				rb_entry(vma->vm_rb.rb_right,
					 struct vm_area_struct, vm_rb);
			if (right->rb_subtree_gap >= length) {
				vma = right;
				continue;
			}
		}

check_current:
		/* Check if current node has a suitable gap */
		gap_end = vma->vm_start;
		if (gap_end < low_limit)
			return -ENOMEM;
		if (gap_start <= high_limit && gap_end - gap_start >= length)
			goto found;

		/* Visit left subtree if it looks promising */
		if (vma->vm_rb.rb_left) {
			struct vm_area_struct *left =
				rb_entry(vma->vm_rb.rb_left,
					 struct vm_area_struct, vm_rb);
			if (left->rb_subtree_gap >= length) {
				vma = left;
				continue;
			}
		}

		/* Go back up the rbtree to find next candidate node */
		while (true) {
			struct rb_node *prev = &vma->vm_rb;
			if (!rb_parent(prev))
				return -ENOMEM;
			vma = rb_entry(rb_parent(prev),
				       struct vm_area_struct, vm_rb);
			if (prev == vma->vm_rb.rb_right) {
				gap_start = vma->vm_prev ?
					vma->vm_prev->vm_end : 0;
				goto check_current;
			}
		}
	}

found:
	/* We found a suitable gap. Clip it with the original high_limit. */
	if (gap_end > info->high_limit)
		gap_end = info->high_limit;

found_highest:
	/* Compute highest gap address at the desired alignment */
	gap_end -= info->length;
	gap_end -= (gap_end - info->align_offset) & info->align_mask;

	VM_BUG_ON(gap_end < info->low_limit);
	VM_BUG_ON(gap_end < gap_start);
	return gap_end;
}

L
Linus Torvalds 已提交
1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900
/* Get an address range which is currently unmapped.
 * For shmat() with addr=0.
 *
 * Ugly calling convention alert:
 * Return value with the low bits set means error value,
 * ie
 *	if (ret & ~PAGE_MASK)
 *		error = ret;
 *
 * This function "knows" that -ENOMEM has the bits set.
 */
#ifndef HAVE_ARCH_UNMAPPED_AREA
unsigned long
arch_get_unmapped_area(struct file *filp, unsigned long addr,
		unsigned long len, unsigned long pgoff, unsigned long flags)
{
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
1901
	struct vm_unmapped_area_info info;
L
Linus Torvalds 已提交
1902

1903
	if (len > TASK_SIZE - mmap_min_addr)
L
Linus Torvalds 已提交
1904 1905
		return -ENOMEM;

1906 1907 1908
	if (flags & MAP_FIXED)
		return addr;

L
Linus Torvalds 已提交
1909 1910 1911
	if (addr) {
		addr = PAGE_ALIGN(addr);
		vma = find_vma(mm, addr);
1912
		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
L
Linus Torvalds 已提交
1913 1914 1915 1916
		    (!vma || addr + len <= vma->vm_start))
			return addr;
	}

1917 1918
	info.flags = 0;
	info.length = len;
1919
	info.low_limit = mm->mmap_base;
1920 1921 1922
	info.high_limit = TASK_SIZE;
	info.align_mask = 0;
	return vm_unmapped_area(&info);
L
Linus Torvalds 已提交
1923
}
V
vishnu.ps 已提交
1924
#endif
L
Linus Torvalds 已提交
1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937

/*
 * This mmap-allocator allocates new areas top-down from below the
 * stack's low limit (the base):
 */
#ifndef HAVE_ARCH_UNMAPPED_AREA_TOPDOWN
unsigned long
arch_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
			  const unsigned long len, const unsigned long pgoff,
			  const unsigned long flags)
{
	struct vm_area_struct *vma;
	struct mm_struct *mm = current->mm;
1938 1939
	unsigned long addr = addr0;
	struct vm_unmapped_area_info info;
L
Linus Torvalds 已提交
1940 1941

	/* requested length too big for entire address space */
1942
	if (len > TASK_SIZE - mmap_min_addr)
L
Linus Torvalds 已提交
1943 1944
		return -ENOMEM;

1945 1946 1947
	if (flags & MAP_FIXED)
		return addr;

L
Linus Torvalds 已提交
1948 1949 1950 1951
	/* requesting a specific address */
	if (addr) {
		addr = PAGE_ALIGN(addr);
		vma = find_vma(mm, addr);
1952
		if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
L
Linus Torvalds 已提交
1953 1954 1955 1956
				(!vma || addr + len <= vma->vm_start))
			return addr;
	}

1957 1958
	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	info.length = len;
1959
	info.low_limit = max(PAGE_SIZE, mmap_min_addr);
1960 1961 1962
	info.high_limit = mm->mmap_base;
	info.align_mask = 0;
	addr = vm_unmapped_area(&info);
1963

L
Linus Torvalds 已提交
1964 1965 1966 1967 1968 1969
	/*
	 * A failed mmap() very likely causes application failure,
	 * so fall back to the bottom-up function here. This scenario
	 * can happen with large stack limits and large mmap()
	 * allocations.
	 */
1970 1971 1972 1973 1974 1975 1976
	if (addr & ~PAGE_MASK) {
		VM_BUG_ON(addr != -ENOMEM);
		info.flags = 0;
		info.low_limit = TASK_UNMAPPED_BASE;
		info.high_limit = TASK_SIZE;
		addr = vm_unmapped_area(&info);
	}
L
Linus Torvalds 已提交
1977 1978 1979 1980 1981 1982 1983 1984 1985

	return addr;
}
#endif

unsigned long
get_unmapped_area(struct file *file, unsigned long addr, unsigned long len,
		unsigned long pgoff, unsigned long flags)
{
1986 1987 1988
	unsigned long (*get_area)(struct file *, unsigned long,
				  unsigned long, unsigned long, unsigned long);

1989 1990 1991 1992 1993 1994 1995 1996
	unsigned long error = arch_mmap_check(addr, len, flags);
	if (error)
		return error;

	/* Careful about overflows.. */
	if (len > TASK_SIZE)
		return -ENOMEM;

1997
	get_area = current->mm->get_unmapped_area;
A
Al Viro 已提交
1998
	if (file && file->f_op->get_unmapped_area)
1999 2000 2001 2002
		get_area = file->f_op->get_unmapped_area;
	addr = get_area(file, addr, len, pgoff, flags);
	if (IS_ERR_VALUE(addr))
		return addr;
L
Linus Torvalds 已提交
2003

2004 2005 2006 2007
	if (addr > TASK_SIZE - len)
		return -ENOMEM;
	if (addr & ~PAGE_MASK)
		return -EINVAL;
2008

2009 2010 2011
	addr = arch_rebalance_pgtables(addr, len);
	error = security_mmap_addr(addr);
	return error ? error : addr;
L
Linus Torvalds 已提交
2012 2013 2014 2015 2016
}

EXPORT_SYMBOL(get_unmapped_area);

/* Look up the first VMA which satisfies  addr < vm_end,  NULL if none. */
Z
ZhenwenXu 已提交
2017
struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr)
L
Linus Torvalds 已提交
2018
{
D
Davidlohr Bueso 已提交
2019 2020
	struct rb_node *rb_node;
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
2021

2022
	/* Check the cache first. */
D
Davidlohr Bueso 已提交
2023 2024 2025
	vma = vmacache_find(mm, addr);
	if (likely(vma))
		return vma;
2026

D
Davidlohr Bueso 已提交
2027 2028
	rb_node = mm->mm_rb.rb_node;
	vma = NULL;
2029

D
Davidlohr Bueso 已提交
2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041
	while (rb_node) {
		struct vm_area_struct *tmp;

		tmp = rb_entry(rb_node, struct vm_area_struct, vm_rb);

		if (tmp->vm_end > addr) {
			vma = tmp;
			if (tmp->vm_start <= addr)
				break;
			rb_node = rb_node->rb_left;
		} else
			rb_node = rb_node->rb_right;
L
Linus Torvalds 已提交
2042
	}
D
Davidlohr Bueso 已提交
2043 2044 2045

	if (vma)
		vmacache_update(addr, vma);
L
Linus Torvalds 已提交
2046 2047 2048 2049 2050
	return vma;
}

EXPORT_SYMBOL(find_vma);

K
KOSAKI Motohiro 已提交
2051 2052 2053
/*
 * Same as find_vma, but also return a pointer to the previous VMA in *pprev.
 */
L
Linus Torvalds 已提交
2054 2055 2056 2057
struct vm_area_struct *
find_vma_prev(struct mm_struct *mm, unsigned long addr,
			struct vm_area_struct **pprev)
{
K
KOSAKI Motohiro 已提交
2058
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
2059

K
KOSAKI Motohiro 已提交
2060
	vma = find_vma(mm, addr);
M
Mikulas Patocka 已提交
2061 2062 2063 2064 2065 2066 2067 2068 2069 2070
	if (vma) {
		*pprev = vma->vm_prev;
	} else {
		struct rb_node *rb_node = mm->mm_rb.rb_node;
		*pprev = NULL;
		while (rb_node) {
			*pprev = rb_entry(rb_node, struct vm_area_struct, vm_rb);
			rb_node = rb_node->rb_right;
		}
	}
K
KOSAKI Motohiro 已提交
2071
	return vma;
L
Linus Torvalds 已提交
2072 2073 2074 2075 2076 2077 2078
}

/*
 * Verify that the stack growth is acceptable and
 * update accounting. This is shared with both the
 * grow-up and grow-down cases.
 */
Z
ZhenwenXu 已提交
2079
static int acct_stack_growth(struct vm_area_struct *vma, unsigned long size, unsigned long grow)
L
Linus Torvalds 已提交
2080 2081 2082
{
	struct mm_struct *mm = vma->vm_mm;
	struct rlimit *rlim = current->signal->rlim;
2083
	unsigned long new_start;
L
Linus Torvalds 已提交
2084 2085

	/* address space limit tests */
A
akpm@osdl.org 已提交
2086
	if (!may_expand_vm(mm, grow))
L
Linus Torvalds 已提交
2087 2088 2089
		return -ENOMEM;

	/* Stack limit test */
J
Jiri Slaby 已提交
2090
	if (size > ACCESS_ONCE(rlim[RLIMIT_STACK].rlim_cur))
L
Linus Torvalds 已提交
2091 2092 2093 2094 2095 2096 2097
		return -ENOMEM;

	/* mlock limit tests */
	if (vma->vm_flags & VM_LOCKED) {
		unsigned long locked;
		unsigned long limit;
		locked = mm->locked_vm + grow;
J
Jiri Slaby 已提交
2098 2099
		limit = ACCESS_ONCE(rlim[RLIMIT_MEMLOCK].rlim_cur);
		limit >>= PAGE_SHIFT;
L
Linus Torvalds 已提交
2100 2101 2102 2103
		if (locked > limit && !capable(CAP_IPC_LOCK))
			return -ENOMEM;
	}

2104 2105 2106 2107 2108 2109
	/* Check to ensure the stack will not grow into a hugetlb-only region */
	new_start = (vma->vm_flags & VM_GROWSUP) ? vma->vm_start :
			vma->vm_end - size;
	if (is_hugepage_only_range(vma->vm_mm, new_start, size))
		return -EFAULT;

L
Linus Torvalds 已提交
2110 2111 2112 2113
	/*
	 * Overcommit..  This must be the final test, as it will
	 * update security statistics.
	 */
2114
	if (security_vm_enough_memory_mm(mm, grow))
L
Linus Torvalds 已提交
2115 2116 2117 2118 2119
		return -ENOMEM;

	/* Ok, everything looks good - let it rip */
	if (vma->vm_flags & VM_LOCKED)
		mm->locked_vm += grow;
2120
	vm_stat_account(mm, vma->vm_flags, vma->vm_file, grow);
L
Linus Torvalds 已提交
2121 2122 2123
	return 0;
}

H
Hugh Dickins 已提交
2124
#if defined(CONFIG_STACK_GROWSUP) || defined(CONFIG_IA64)
L
Linus Torvalds 已提交
2125
/*
H
Hugh Dickins 已提交
2126 2127
 * PA-RISC uses this for its stack; IA64 for its Register Backing Store.
 * vma is the last one with address > vma->vm_end.  Have to extend vma.
L
Linus Torvalds 已提交
2128
 */
H
Hugh Dickins 已提交
2129
int expand_upwards(struct vm_area_struct *vma, unsigned long address)
L
Linus Torvalds 已提交
2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141
{
	int error;

	if (!(vma->vm_flags & VM_GROWSUP))
		return -EFAULT;

	/*
	 * We must make sure the anon_vma is allocated
	 * so that the anon_vma locking is not a noop.
	 */
	if (unlikely(anon_vma_prepare(vma)))
		return -ENOMEM;
2142
	vma_lock_anon_vma(vma);
L
Linus Torvalds 已提交
2143 2144 2145 2146 2147

	/*
	 * vma->vm_start/vm_end cannot change under us because the caller
	 * is required to hold the mmap_sem in read mode.  We need the
	 * anon_vma lock to serialize against concurrent expand_stacks.
2148
	 * Also guard against wrapping around to address 0.
L
Linus Torvalds 已提交
2149
	 */
2150 2151 2152
	if (address < PAGE_ALIGN(address+4))
		address = PAGE_ALIGN(address+4);
	else {
2153
		vma_unlock_anon_vma(vma);
2154 2155
		return -ENOMEM;
	}
L
Linus Torvalds 已提交
2156 2157 2158 2159 2160 2161 2162 2163 2164
	error = 0;

	/* Somebody else might have raced and expanded it already */
	if (address > vma->vm_end) {
		unsigned long size, grow;

		size = address - vma->vm_start;
		grow = (address - vma->vm_end) >> PAGE_SHIFT;

2165 2166 2167 2168
		error = -ENOMEM;
		if (vma->vm_pgoff + (size >> PAGE_SHIFT) >= vma->vm_pgoff) {
			error = acct_stack_growth(vma, size, grow);
			if (!error) {
2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
				/*
				 * vma_gap_update() doesn't support concurrent
				 * updates, but we only hold a shared mmap_sem
				 * lock here, so we need to protect against
				 * concurrent vma expansions.
				 * vma_lock_anon_vma() doesn't help here, as
				 * we don't guarantee that all growable vmas
				 * in a mm share the same root anon vma.
				 * So, we reuse mm->page_table_lock to guard
				 * against concurrent vma expansions.
				 */
				spin_lock(&vma->vm_mm->page_table_lock);
2181
				anon_vma_interval_tree_pre_update_vma(vma);
2182
				vma->vm_end = address;
2183
				anon_vma_interval_tree_post_update_vma(vma);
2184 2185 2186 2187
				if (vma->vm_next)
					vma_gap_update(vma->vm_next);
				else
					vma->vm_mm->highest_vm_end = address;
2188 2189
				spin_unlock(&vma->vm_mm->page_table_lock);

2190 2191
				perf_event_mmap(vma);
			}
2192
		}
L
Linus Torvalds 已提交
2193
	}
2194
	vma_unlock_anon_vma(vma);
A
Andrea Arcangeli 已提交
2195
	khugepaged_enter_vma_merge(vma);
2196
	validate_mm(vma->vm_mm);
L
Linus Torvalds 已提交
2197 2198
	return error;
}
H
Hugh Dickins 已提交
2199 2200
#endif /* CONFIG_STACK_GROWSUP || CONFIG_IA64 */

L
Linus Torvalds 已提交
2201 2202 2203
/*
 * vma is the first one with address < vma->vm_start.  Have to extend vma.
 */
2204
int expand_downwards(struct vm_area_struct *vma,
2205
				   unsigned long address)
L
Linus Torvalds 已提交
2206 2207 2208 2209 2210 2211 2212 2213 2214
{
	int error;

	/*
	 * We must make sure the anon_vma is allocated
	 * so that the anon_vma locking is not a noop.
	 */
	if (unlikely(anon_vma_prepare(vma)))
		return -ENOMEM;
2215 2216

	address &= PAGE_MASK;
2217
	error = security_mmap_addr(address);
2218 2219 2220
	if (error)
		return error;

2221
	vma_lock_anon_vma(vma);
L
Linus Torvalds 已提交
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235

	/*
	 * vma->vm_start/vm_end cannot change under us because the caller
	 * is required to hold the mmap_sem in read mode.  We need the
	 * anon_vma lock to serialize against concurrent expand_stacks.
	 */

	/* Somebody else might have raced and expanded it already */
	if (address < vma->vm_start) {
		unsigned long size, grow;

		size = vma->vm_end - address;
		grow = (vma->vm_start - address) >> PAGE_SHIFT;

2236 2237 2238 2239
		error = -ENOMEM;
		if (grow <= vma->vm_pgoff) {
			error = acct_stack_growth(vma, size, grow);
			if (!error) {
2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251
				/*
				 * vma_gap_update() doesn't support concurrent
				 * updates, but we only hold a shared mmap_sem
				 * lock here, so we need to protect against
				 * concurrent vma expansions.
				 * vma_lock_anon_vma() doesn't help here, as
				 * we don't guarantee that all growable vmas
				 * in a mm share the same root anon vma.
				 * So, we reuse mm->page_table_lock to guard
				 * against concurrent vma expansions.
				 */
				spin_lock(&vma->vm_mm->page_table_lock);
2252
				anon_vma_interval_tree_pre_update_vma(vma);
2253 2254
				vma->vm_start = address;
				vma->vm_pgoff -= grow;
2255
				anon_vma_interval_tree_post_update_vma(vma);
2256
				vma_gap_update(vma);
2257 2258
				spin_unlock(&vma->vm_mm->page_table_lock);

2259 2260
				perf_event_mmap(vma);
			}
L
Linus Torvalds 已提交
2261 2262
		}
	}
2263
	vma_unlock_anon_vma(vma);
A
Andrea Arcangeli 已提交
2264
	khugepaged_enter_vma_merge(vma);
2265
	validate_mm(vma->vm_mm);
L
Linus Torvalds 已提交
2266 2267 2268
	return error;
}

2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279
/*
 * Note how expand_stack() refuses to expand the stack all the way to
 * abut the next virtual mapping, *unless* that mapping itself is also
 * a stack mapping. We want to leave room for a guard page, after all
 * (the guard page itself is not added here, that is done by the
 * actual page faulting logic)
 *
 * This matches the behavior of the guard page logic (see mm/memory.c:
 * check_stack_guard_page()), which only allows the guard page to be
 * removed under these circumstances.
 */
2280 2281 2282
#ifdef CONFIG_STACK_GROWSUP
int expand_stack(struct vm_area_struct *vma, unsigned long address)
{
2283 2284 2285 2286 2287 2288 2289 2290
	struct vm_area_struct *next;

	address &= PAGE_MASK;
	next = vma->vm_next;
	if (next && next->vm_start == address + PAGE_SIZE) {
		if (!(next->vm_flags & VM_GROWSUP))
			return -ENOMEM;
	}
2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302
	return expand_upwards(vma, address);
}

struct vm_area_struct *
find_extend_vma(struct mm_struct *mm, unsigned long addr)
{
	struct vm_area_struct *vma, *prev;

	addr &= PAGE_MASK;
	vma = find_vma_prev(mm, addr, &prev);
	if (vma && (vma->vm_start <= addr))
		return vma;
2303
	if (!prev || expand_stack(prev, addr))
2304
		return NULL;
2305 2306
	if (prev->vm_flags & VM_LOCKED)
		__mlock_vma_pages_range(prev, addr, prev->vm_end, NULL);
2307 2308 2309 2310 2311
	return prev;
}
#else
int expand_stack(struct vm_area_struct *vma, unsigned long address)
{
2312 2313 2314 2315 2316 2317 2318 2319
	struct vm_area_struct *prev;

	address &= PAGE_MASK;
	prev = vma->vm_prev;
	if (prev && prev->vm_end == address) {
		if (!(prev->vm_flags & VM_GROWSDOWN))
			return -ENOMEM;
	}
2320 2321 2322
	return expand_downwards(vma, address);
}

L
Linus Torvalds 已提交
2323
struct vm_area_struct *
V
vishnu.ps 已提交
2324
find_extend_vma(struct mm_struct *mm, unsigned long addr)
L
Linus Torvalds 已提交
2325
{
V
vishnu.ps 已提交
2326
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
2327 2328 2329
	unsigned long start;

	addr &= PAGE_MASK;
V
vishnu.ps 已提交
2330
	vma = find_vma(mm, addr);
L
Linus Torvalds 已提交
2331 2332 2333 2334 2335 2336 2337 2338 2339
	if (!vma)
		return NULL;
	if (vma->vm_start <= addr)
		return vma;
	if (!(vma->vm_flags & VM_GROWSDOWN))
		return NULL;
	start = vma->vm_start;
	if (expand_stack(vma, addr))
		return NULL;
2340 2341
	if (vma->vm_flags & VM_LOCKED)
		__mlock_vma_pages_range(vma, addr, start, NULL);
L
Linus Torvalds 已提交
2342 2343 2344 2345 2346
	return vma;
}
#endif

/*
2347
 * Ok - we have the memory areas we should free on the vma list,
L
Linus Torvalds 已提交
2348
 * so release them, and do the vma updates.
2349 2350
 *
 * Called with the mm semaphore held.
L
Linus Torvalds 已提交
2351
 */
2352
static void remove_vma_list(struct mm_struct *mm, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
2353
{
2354 2355
	unsigned long nr_accounted = 0;

2356 2357
	/* Update high watermark before we lower total_vm */
	update_hiwater_vm(mm);
L
Linus Torvalds 已提交
2358
	do {
2359 2360
		long nrpages = vma_pages(vma);

2361 2362
		if (vma->vm_flags & VM_ACCOUNT)
			nr_accounted += nrpages;
2363
		vm_stat_account(mm, vma->vm_flags, vma->vm_file, -nrpages);
2364
		vma = remove_vma(vma);
2365
	} while (vma);
2366
	vm_unacct_memory(nr_accounted);
L
Linus Torvalds 已提交
2367 2368 2369 2370 2371 2372
	validate_mm(mm);
}

/*
 * Get rid of page table information in the indicated region.
 *
2373
 * Called with the mm semaphore held.
L
Linus Torvalds 已提交
2374 2375
 */
static void unmap_region(struct mm_struct *mm,
2376 2377
		struct vm_area_struct *vma, struct vm_area_struct *prev,
		unsigned long start, unsigned long end)
L
Linus Torvalds 已提交
2378
{
V
vishnu.ps 已提交
2379
	struct vm_area_struct *next = prev ? prev->vm_next : mm->mmap;
P
Peter Zijlstra 已提交
2380
	struct mmu_gather tlb;
L
Linus Torvalds 已提交
2381 2382

	lru_add_drain();
2383
	tlb_gather_mmu(&tlb, mm, start, end);
2384
	update_hiwater_rss(mm);
2385
	unmap_vmas(&tlb, vma, start, end);
P
Peter Zijlstra 已提交
2386
	free_pgtables(&tlb, vma, prev ? prev->vm_end : FIRST_USER_ADDRESS,
2387
				 next ? next->vm_start : USER_PGTABLES_CEILING);
P
Peter Zijlstra 已提交
2388
	tlb_finish_mmu(&tlb, start, end);
L
Linus Torvalds 已提交
2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402
}

/*
 * Create a list of vma's touched by the unmap, removing them from the mm's
 * vma list as we go..
 */
static void
detach_vmas_to_be_unmapped(struct mm_struct *mm, struct vm_area_struct *vma,
	struct vm_area_struct *prev, unsigned long end)
{
	struct vm_area_struct **insertion_point;
	struct vm_area_struct *tail_vma = NULL;

	insertion_point = (prev ? &prev->vm_next : &mm->mmap);
2403
	vma->vm_prev = NULL;
L
Linus Torvalds 已提交
2404
	do {
2405
		vma_rb_erase(vma, &mm->mm_rb);
L
Linus Torvalds 已提交
2406 2407 2408 2409 2410
		mm->map_count--;
		tail_vma = vma;
		vma = vma->vm_next;
	} while (vma && vma->vm_start < end);
	*insertion_point = vma;
2411
	if (vma) {
2412
		vma->vm_prev = prev;
2413 2414 2415
		vma_gap_update(vma);
	} else
		mm->highest_vm_end = prev ? prev->vm_end : 0;
L
Linus Torvalds 已提交
2416
	tail_vma->vm_next = NULL;
D
Davidlohr Bueso 已提交
2417 2418 2419

	/* Kill the cache */
	vmacache_invalidate(mm);
L
Linus Torvalds 已提交
2420 2421 2422
}

/*
2423 2424
 * __split_vma() bypasses sysctl_max_map_count checking.  We use this on the
 * munmap path where it doesn't make sense to fail.
L
Linus Torvalds 已提交
2425
 */
V
vishnu.ps 已提交
2426
static int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
L
Linus Torvalds 已提交
2427 2428 2429
	      unsigned long addr, int new_below)
{
	struct vm_area_struct *new;
2430
	int err = -ENOMEM;
L
Linus Torvalds 已提交
2431

2432 2433
	if (is_vm_hugetlb_page(vma) && (addr &
					~(huge_page_mask(hstate_vma(vma)))))
L
Linus Torvalds 已提交
2434 2435
		return -EINVAL;

2436
	new = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2437
	if (!new)
2438
		goto out_err;
L
Linus Torvalds 已提交
2439 2440 2441 2442

	/* most fields are the same, copy all, and then fixup */
	*new = *vma;

2443 2444
	INIT_LIST_HEAD(&new->anon_vma_chain);

L
Linus Torvalds 已提交
2445 2446 2447 2448 2449 2450 2451
	if (new_below)
		new->vm_end = addr;
	else {
		new->vm_start = addr;
		new->vm_pgoff += ((addr - vma->vm_start) >> PAGE_SHIFT);
	}

2452 2453
	err = vma_dup_policy(vma, new);
	if (err)
2454
		goto out_free_vma;
L
Linus Torvalds 已提交
2455

2456 2457 2458
	if (anon_vma_clone(new, vma))
		goto out_free_mpol;

2459
	if (new->vm_file)
L
Linus Torvalds 已提交
2460 2461 2462 2463 2464 2465
		get_file(new->vm_file);

	if (new->vm_ops && new->vm_ops->open)
		new->vm_ops->open(new);

	if (new_below)
2466
		err = vma_adjust(vma, addr, vma->vm_end, vma->vm_pgoff +
L
Linus Torvalds 已提交
2467 2468
			((addr - new->vm_start) >> PAGE_SHIFT), new);
	else
2469
		err = vma_adjust(vma, vma->vm_start, addr, vma->vm_pgoff, new);
L
Linus Torvalds 已提交
2470

2471 2472 2473 2474 2475
	/* Success. */
	if (!err)
		return 0;

	/* Clean everything up if vma_adjust failed. */
2476 2477
	if (new->vm_ops && new->vm_ops->close)
		new->vm_ops->close(new);
2478
	if (new->vm_file)
2479
		fput(new->vm_file);
2480
	unlink_anon_vmas(new);
2481
 out_free_mpol:
2482
	mpol_put(vma_policy(new));
2483 2484 2485 2486
 out_free_vma:
	kmem_cache_free(vm_area_cachep, new);
 out_err:
	return err;
L
Linus Torvalds 已提交
2487 2488
}

2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501
/*
 * Split a vma into two pieces at address 'addr', a new vma is allocated
 * either for the first part or the tail.
 */
int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
	      unsigned long addr, int new_below)
{
	if (mm->map_count >= sysctl_max_map_count)
		return -ENOMEM;

	return __split_vma(mm, vma, addr, new_below);
}

L
Linus Torvalds 已提交
2502 2503 2504 2505 2506 2507 2508 2509
/* Munmap is split into 2 main parts -- this part which finds
 * what needs doing, and the areas themselves, which do the
 * work.  This now handles partial unmappings.
 * Jeremy Fitzhardinge <jeremy@goop.org>
 */
int do_munmap(struct mm_struct *mm, unsigned long start, size_t len)
{
	unsigned long end;
2510
	struct vm_area_struct *vma, *prev, *last;
L
Linus Torvalds 已提交
2511 2512 2513 2514

	if ((start & ~PAGE_MASK) || start > TASK_SIZE || len > TASK_SIZE-start)
		return -EINVAL;

V
vishnu.ps 已提交
2515 2516
	len = PAGE_ALIGN(len);
	if (len == 0)
L
Linus Torvalds 已提交
2517 2518 2519
		return -EINVAL;

	/* Find the first overlapping VMA */
2520
	vma = find_vma(mm, start);
2521
	if (!vma)
L
Linus Torvalds 已提交
2522
		return 0;
2523
	prev = vma->vm_prev;
2524
	/* we have  start < vma->vm_end  */
L
Linus Torvalds 已提交
2525 2526 2527

	/* if it doesn't overlap, we have nothing.. */
	end = start + len;
2528
	if (vma->vm_start >= end)
L
Linus Torvalds 已提交
2529 2530 2531 2532 2533 2534 2535 2536 2537
		return 0;

	/*
	 * If we need to split any vma, do it now to save pain later.
	 *
	 * Note: mremap's move_vma VM_ACCOUNT handling assumes a partially
	 * unmapped vm_area_struct will remain in use: so lower split_vma
	 * places tmp vma above, and higher split_vma places tmp vma below.
	 */
2538
	if (start > vma->vm_start) {
2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549
		int error;

		/*
		 * Make sure that map_count on return from munmap() will
		 * not exceed its limit; but let map_count go just above
		 * its limit temporarily, to help free resources as expected.
		 */
		if (end < vma->vm_end && mm->map_count >= sysctl_max_map_count)
			return -ENOMEM;

		error = __split_vma(mm, vma, start, 0);
L
Linus Torvalds 已提交
2550 2551
		if (error)
			return error;
2552
		prev = vma;
L
Linus Torvalds 已提交
2553 2554 2555 2556 2557
	}

	/* Does it split the last one? */
	last = find_vma(mm, end);
	if (last && end > last->vm_start) {
2558
		int error = __split_vma(mm, last, end, 1);
L
Linus Torvalds 已提交
2559 2560 2561
		if (error)
			return error;
	}
V
vishnu.ps 已提交
2562
	vma = prev ? prev->vm_next : mm->mmap;
L
Linus Torvalds 已提交
2563

2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577
	/*
	 * unlock any mlock()ed ranges before detaching vmas
	 */
	if (mm->locked_vm) {
		struct vm_area_struct *tmp = vma;
		while (tmp && tmp->vm_start < end) {
			if (tmp->vm_flags & VM_LOCKED) {
				mm->locked_vm -= vma_pages(tmp);
				munlock_vma_pages_all(tmp);
			}
			tmp = tmp->vm_next;
		}
	}

L
Linus Torvalds 已提交
2578 2579 2580
	/*
	 * Remove the vma's, and unmap the actual pages
	 */
2581 2582
	detach_vmas_to_be_unmapped(mm, vma, prev, end);
	unmap_region(mm, vma, prev, start, end);
L
Linus Torvalds 已提交
2583 2584

	/* Fix up all other VM information */
2585
	remove_vma_list(mm, vma);
L
Linus Torvalds 已提交
2586 2587 2588 2589

	return 0;
}

A
Al Viro 已提交
2590
int vm_munmap(unsigned long start, size_t len)
L
Linus Torvalds 已提交
2591 2592
{
	int ret;
A
Al Viro 已提交
2593
	struct mm_struct *mm = current->mm;
L
Linus Torvalds 已提交
2594 2595

	down_write(&mm->mmap_sem);
2596
	ret = do_munmap(mm, start, len);
L
Linus Torvalds 已提交
2597 2598 2599
	up_write(&mm->mmap_sem);
	return ret;
}
2600 2601 2602 2603 2604
EXPORT_SYMBOL(vm_munmap);

SYSCALL_DEFINE2(munmap, unsigned long, addr, size_t, len)
{
	profile_munmap(addr);
A
Al Viro 已提交
2605
	return vm_munmap(addr, len);
2606
}
L
Linus Torvalds 已提交
2607 2608 2609

static inline void verify_mm_writelocked(struct mm_struct *mm)
{
2610
#ifdef CONFIG_DEBUG_VM
L
Linus Torvalds 已提交
2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622
	if (unlikely(down_read_trylock(&mm->mmap_sem))) {
		WARN_ON(1);
		up_read(&mm->mmap_sem);
	}
#endif
}

/*
 *  this is really a simplified "do_mmap".  it only handles
 *  anonymous maps.  eventually we may be able to do some
 *  brk-specific accounting here.
 */
2623
static unsigned long do_brk(unsigned long addr, unsigned long len)
L
Linus Torvalds 已提交
2624
{
V
vishnu.ps 已提交
2625 2626
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma, *prev;
L
Linus Torvalds 已提交
2627
	unsigned long flags;
V
vishnu.ps 已提交
2628
	struct rb_node **rb_link, *rb_parent;
L
Linus Torvalds 已提交
2629
	pgoff_t pgoff = addr >> PAGE_SHIFT;
2630
	int error;
L
Linus Torvalds 已提交
2631 2632 2633 2634 2635

	len = PAGE_ALIGN(len);
	if (!len)
		return addr;

2636 2637
	flags = VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags;

A
Al Viro 已提交
2638 2639
	error = get_unmapped_area(NULL, addr, len, 0, MAP_FIXED);
	if (error & ~PAGE_MASK)
2640 2641
		return error;

2642 2643 2644
	error = mlock_future_check(mm, mm->def_flags, len);
	if (error)
		return error;
L
Linus Torvalds 已提交
2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655

	/*
	 * mm->mmap_sem is required to protect against another thread
	 * changing the mappings in case we sleep.
	 */
	verify_mm_writelocked(mm);

	/*
	 * Clear old maps.  this also does some error checking for us
	 */
 munmap_back:
2656
	if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent)) {
L
Linus Torvalds 已提交
2657 2658 2659 2660 2661 2662
		if (do_munmap(mm, addr, len))
			return -ENOMEM;
		goto munmap_back;
	}

	/* Check against address space limits *after* clearing old maps... */
A
akpm@osdl.org 已提交
2663
	if (!may_expand_vm(mm, len >> PAGE_SHIFT))
L
Linus Torvalds 已提交
2664 2665 2666 2667 2668
		return -ENOMEM;

	if (mm->map_count > sysctl_max_map_count)
		return -ENOMEM;

2669
	if (security_vm_enough_memory_mm(mm, len >> PAGE_SHIFT))
L
Linus Torvalds 已提交
2670 2671 2672
		return -ENOMEM;

	/* Can we just expand an old private anonymous mapping? */
2673 2674 2675
	vma = vma_merge(mm, prev, addr, addr + len, flags,
					NULL, NULL, pgoff, NULL);
	if (vma)
L
Linus Torvalds 已提交
2676 2677 2678 2679 2680
		goto out;

	/*
	 * create a vma struct for an anonymous mapping
	 */
P
Pekka Enberg 已提交
2681
	vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2682 2683 2684 2685 2686
	if (!vma) {
		vm_unacct_memory(len >> PAGE_SHIFT);
		return -ENOMEM;
	}

2687
	INIT_LIST_HEAD(&vma->anon_vma_chain);
L
Linus Torvalds 已提交
2688 2689 2690 2691 2692
	vma->vm_mm = mm;
	vma->vm_start = addr;
	vma->vm_end = addr + len;
	vma->vm_pgoff = pgoff;
	vma->vm_flags = flags;
2693
	vma->vm_page_prot = vm_get_page_prot(flags);
L
Linus Torvalds 已提交
2694 2695
	vma_link(mm, vma, prev, rb_link, rb_parent);
out:
2696
	perf_event_mmap(vma);
L
Linus Torvalds 已提交
2697
	mm->total_vm += len >> PAGE_SHIFT;
2698 2699
	if (flags & VM_LOCKED)
		mm->locked_vm += (len >> PAGE_SHIFT);
2700
	vma->vm_flags |= VM_SOFTDIRTY;
L
Linus Torvalds 已提交
2701 2702 2703
	return addr;
}

2704 2705 2706 2707
unsigned long vm_brk(unsigned long addr, unsigned long len)
{
	struct mm_struct *mm = current->mm;
	unsigned long ret;
2708
	bool populate;
2709 2710 2711

	down_write(&mm->mmap_sem);
	ret = do_brk(addr, len);
2712
	populate = ((mm->def_flags & VM_LOCKED) != 0);
2713
	up_write(&mm->mmap_sem);
2714 2715
	if (populate)
		mm_populate(addr, len);
2716 2717 2718
	return ret;
}
EXPORT_SYMBOL(vm_brk);
L
Linus Torvalds 已提交
2719 2720 2721 2722

/* Release all mmaps. */
void exit_mmap(struct mm_struct *mm)
{
P
Peter Zijlstra 已提交
2723
	struct mmu_gather tlb;
2724
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
2725 2726
	unsigned long nr_accounted = 0;

2727
	/* mm's last user has gone, and its about to be pulled down */
A
Andrea Arcangeli 已提交
2728
	mmu_notifier_release(mm);
2729

2730 2731 2732 2733 2734 2735 2736 2737
	if (mm->locked_vm) {
		vma = mm->mmap;
		while (vma) {
			if (vma->vm_flags & VM_LOCKED)
				munlock_vma_pages_all(vma);
			vma = vma->vm_next;
		}
	}
2738 2739 2740

	arch_exit_mmap(mm);

2741
	vma = mm->mmap;
2742 2743 2744
	if (!vma)	/* Can happen if dup_mmap() received an OOM */
		return;

L
Linus Torvalds 已提交
2745 2746
	lru_add_drain();
	flush_cache_mm(mm);
2747
	tlb_gather_mmu(&tlb, mm, 0, -1);
2748
	/* update_hiwater_rss(mm) here? but nobody should be looking */
2749
	/* Use -1 here to ensure all VMAs in the mm are unmapped */
2750
	unmap_vmas(&tlb, vma, 0, -1);
H
Hugh Dickins 已提交
2751

2752
	free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, USER_PGTABLES_CEILING);
2753
	tlb_finish_mmu(&tlb, 0, -1);
L
Linus Torvalds 已提交
2754 2755

	/*
2756 2757
	 * Walk the list again, actually closing and freeing it,
	 * with preemption enabled, without holding any MM locks.
L
Linus Torvalds 已提交
2758
	 */
2759 2760 2761
	while (vma) {
		if (vma->vm_flags & VM_ACCOUNT)
			nr_accounted += vma_pages(vma);
2762
		vma = remove_vma(vma);
2763 2764
	}
	vm_unacct_memory(nr_accounted);
2765

2766 2767
	WARN_ON(atomic_long_read(&mm->nr_ptes) >
			(FIRST_USER_ADDRESS+PMD_SIZE-1)>>PMD_SHIFT);
L
Linus Torvalds 已提交
2768 2769 2770 2771
}

/* Insert vm structure into process list sorted by address
 * and into the inode's i_mmap tree.  If vm_file is non-NULL
2772
 * then i_mmap_mutex is taken here.
L
Linus Torvalds 已提交
2773
 */
2774
int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
2775
{
2776 2777
	struct vm_area_struct *prev;
	struct rb_node **rb_link, *rb_parent;
L
Linus Torvalds 已提交
2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794

	/*
	 * The vm_pgoff of a purely anonymous vma should be irrelevant
	 * until its first write fault, when page's anon_vma and index
	 * are set.  But now set the vm_pgoff it will almost certainly
	 * end up with (unless mremap moves it elsewhere before that
	 * first wfault), so /proc/pid/maps tells a consistent story.
	 *
	 * By setting it to reflect the virtual start address of the
	 * vma, merges and splits can happen in a seamless way, just
	 * using the existing file pgoff checks and manipulations.
	 * Similarly in do_mmap_pgoff and in do_brk.
	 */
	if (!vma->vm_file) {
		BUG_ON(vma->anon_vma);
		vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT;
	}
2795 2796
	if (find_vma_links(mm, vma->vm_start, vma->vm_end,
			   &prev, &rb_link, &rb_parent))
L
Linus Torvalds 已提交
2797
		return -ENOMEM;
2798
	if ((vma->vm_flags & VM_ACCOUNT) &&
2799
	     security_vm_enough_memory_mm(mm, vma_pages(vma)))
2800
		return -ENOMEM;
2801

L
Linus Torvalds 已提交
2802 2803 2804 2805 2806 2807 2808 2809 2810
	vma_link(mm, vma, prev, rb_link, rb_parent);
	return 0;
}

/*
 * Copy the vma structure to a new location in the same mm,
 * prior to moving page table entries, to effect an mremap move.
 */
struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
2811 2812
	unsigned long addr, unsigned long len, pgoff_t pgoff,
	bool *need_rmap_locks)
L
Linus Torvalds 已提交
2813 2814 2815 2816 2817 2818
{
	struct vm_area_struct *vma = *vmap;
	unsigned long vma_start = vma->vm_start;
	struct mm_struct *mm = vma->vm_mm;
	struct vm_area_struct *new_vma, *prev;
	struct rb_node **rb_link, *rb_parent;
2819
	bool faulted_in_anon_vma = true;
L
Linus Torvalds 已提交
2820 2821 2822 2823 2824

	/*
	 * If anonymous vma has not yet been faulted, update new pgoff
	 * to match new location, to increase its chance of merging.
	 */
2825
	if (unlikely(!vma->vm_file && !vma->anon_vma)) {
L
Linus Torvalds 已提交
2826
		pgoff = addr >> PAGE_SHIFT;
2827 2828
		faulted_in_anon_vma = false;
	}
L
Linus Torvalds 已提交
2829

2830 2831
	if (find_vma_links(mm, addr, addr + len, &prev, &rb_link, &rb_parent))
		return NULL;	/* should never get here */
L
Linus Torvalds 已提交
2832 2833 2834 2835 2836 2837
	new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags,
			vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma));
	if (new_vma) {
		/*
		 * Source vma may have been merged into new_vma
		 */
2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852
		if (unlikely(vma_start >= new_vma->vm_start &&
			     vma_start < new_vma->vm_end)) {
			/*
			 * The only way we can get a vma_merge with
			 * self during an mremap is if the vma hasn't
			 * been faulted in yet and we were allowed to
			 * reset the dst vma->vm_pgoff to the
			 * destination address of the mremap to allow
			 * the merge to happen. mremap must change the
			 * vm_pgoff linearity between src and dst vmas
			 * (in turn preventing a vma_merge) to be
			 * safe. It is only safe to keep the vm_pgoff
			 * linear if there are no pages mapped yet.
			 */
			VM_BUG_ON(faulted_in_anon_vma);
2853
			*vmap = vma = new_vma;
2854
		}
2855
		*need_rmap_locks = (new_vma->vm_pgoff <= vma->vm_pgoff);
L
Linus Torvalds 已提交
2856
	} else {
2857
		new_vma = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
L
Linus Torvalds 已提交
2858 2859
		if (new_vma) {
			*new_vma = *vma;
2860 2861 2862
			new_vma->vm_start = addr;
			new_vma->vm_end = addr + len;
			new_vma->vm_pgoff = pgoff;
2863
			if (vma_dup_policy(vma, new_vma))
2864 2865 2866 2867
				goto out_free_vma;
			INIT_LIST_HEAD(&new_vma->anon_vma_chain);
			if (anon_vma_clone(new_vma, vma))
				goto out_free_mempol;
2868
			if (new_vma->vm_file)
L
Linus Torvalds 已提交
2869 2870 2871 2872
				get_file(new_vma->vm_file);
			if (new_vma->vm_ops && new_vma->vm_ops->open)
				new_vma->vm_ops->open(new_vma);
			vma_link(mm, new_vma, prev, rb_link, rb_parent);
2873
			*need_rmap_locks = false;
L
Linus Torvalds 已提交
2874 2875 2876
		}
	}
	return new_vma;
2877 2878

 out_free_mempol:
2879
	mpol_put(vma_policy(new_vma));
2880 2881 2882
 out_free_vma:
	kmem_cache_free(vm_area_cachep, new_vma);
	return NULL;
L
Linus Torvalds 已提交
2883
}
A
akpm@osdl.org 已提交
2884 2885 2886 2887 2888 2889 2890 2891 2892 2893

/*
 * Return true if the calling process may expand its vm space by the passed
 * number of pages
 */
int may_expand_vm(struct mm_struct *mm, unsigned long npages)
{
	unsigned long cur = mm->total_vm;	/* pages */
	unsigned long lim;

J
Jiri Slaby 已提交
2894
	lim = rlimit(RLIMIT_AS) >> PAGE_SHIFT;
A
akpm@osdl.org 已提交
2895 2896 2897 2898 2899

	if (cur + npages > lim)
		return 0;
	return 1;
}
2900

2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925
static int special_mapping_fault(struct vm_area_struct *vma,
				 struct vm_fault *vmf);

/*
 * Having a close hook prevents vma merging regardless of flags.
 */
static void special_mapping_close(struct vm_area_struct *vma)
{
}

static const char *special_mapping_name(struct vm_area_struct *vma)
{
	return ((struct vm_special_mapping *)vma->vm_private_data)->name;
}

static const struct vm_operations_struct special_mapping_vmops = {
	.close = special_mapping_close,
	.fault = special_mapping_fault,
	.name = special_mapping_name,
};

static const struct vm_operations_struct legacy_special_mapping_vmops = {
	.close = special_mapping_close,
	.fault = special_mapping_fault,
};
2926

N
Nick Piggin 已提交
2927 2928
static int special_mapping_fault(struct vm_area_struct *vma,
				struct vm_fault *vmf)
2929
{
N
Nick Piggin 已提交
2930
	pgoff_t pgoff;
2931 2932
	struct page **pages;

N
Nick Piggin 已提交
2933 2934 2935 2936 2937 2938 2939
	/*
	 * special mappings have no vm_file, and in that case, the mm
	 * uses vm_pgoff internally. So we have to subtract it from here.
	 * We are allowed to do this because we are the mm; do not copy
	 * this code into drivers!
	 */
	pgoff = vmf->pgoff - vma->vm_pgoff;
2940

2941 2942 2943 2944 2945 2946 2947
	if (vma->vm_ops == &legacy_special_mapping_vmops)
		pages = vma->vm_private_data;
	else
		pages = ((struct vm_special_mapping *)vma->vm_private_data)->
			pages;

	for (; pgoff && *pages; ++pages)
N
Nick Piggin 已提交
2948
		pgoff--;
2949 2950 2951 2952

	if (*pages) {
		struct page *page = *pages;
		get_page(page);
N
Nick Piggin 已提交
2953 2954
		vmf->page = page;
		return 0;
2955 2956
	}

N
Nick Piggin 已提交
2957
	return VM_FAULT_SIGBUS;
2958 2959
}

2960 2961 2962 2963 2964
static struct vm_area_struct *__install_special_mapping(
	struct mm_struct *mm,
	unsigned long addr, unsigned long len,
	unsigned long vm_flags, const struct vm_operations_struct *ops,
	void *priv)
2965
{
2966
	int ret;
2967 2968 2969 2970
	struct vm_area_struct *vma;

	vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
	if (unlikely(vma == NULL))
2971
		return ERR_PTR(-ENOMEM);
2972

2973
	INIT_LIST_HEAD(&vma->anon_vma_chain);
2974 2975 2976 2977
	vma->vm_mm = mm;
	vma->vm_start = addr;
	vma->vm_end = addr + len;

2978
	vma->vm_flags = vm_flags | mm->def_flags | VM_DONTEXPAND | VM_SOFTDIRTY;
2979
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
2980

2981 2982
	vma->vm_ops = ops;
	vma->vm_private_data = priv;
2983

2984 2985 2986
	ret = insert_vm_struct(mm, vma);
	if (ret)
		goto out;
2987 2988 2989

	mm->total_vm += len >> PAGE_SHIFT;

2990
	perf_event_mmap(vma);
2991

2992
	return vma;
2993 2994 2995

out:
	kmem_cache_free(vm_area_cachep, vma);
2996 2997 2998
	return ERR_PTR(ret);
}

2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016
/*
 * Called with mm->mmap_sem held for writing.
 * Insert a new vma covering the given region, with the given flags.
 * Its pages are supplied by the given array of struct page *.
 * The array can be shorter than len >> PAGE_SHIFT if it's null-terminated.
 * The region past the last page supplied will always produce SIGBUS.
 * The array pointer and the pages it points to are assumed to stay alive
 * for as long as this mapping might exist.
 */
struct vm_area_struct *_install_special_mapping(
	struct mm_struct *mm,
	unsigned long addr, unsigned long len,
	unsigned long vm_flags, const struct vm_special_mapping *spec)
{
	return __install_special_mapping(mm, addr, len, vm_flags,
					 &special_mapping_vmops, (void *)spec);
}

3017 3018 3019 3020
int install_special_mapping(struct mm_struct *mm,
			    unsigned long addr, unsigned long len,
			    unsigned long vm_flags, struct page **pages)
{
3021 3022 3023
	struct vm_area_struct *vma = __install_special_mapping(
		mm, addr, len, vm_flags, &legacy_special_mapping_vmops,
		(void *)pages);
3024

3025
	return PTR_ERR_OR_ZERO(vma);
3026
}
3027 3028 3029

static DEFINE_MUTEX(mm_all_locks_mutex);

3030
static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma)
3031
{
3032
	if (!test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) {
3033 3034 3035 3036
		/*
		 * The LSB of head.next can't change from under us
		 * because we hold the mm_all_locks_mutex.
		 */
3037
		down_write_nest_lock(&anon_vma->root->rwsem, &mm->mmap_sem);
3038 3039
		/*
		 * We can safely modify head.next after taking the
3040
		 * anon_vma->root->rwsem. If some other vma in this mm shares
3041 3042 3043 3044
		 * the same anon_vma we won't take it again.
		 *
		 * No need of atomic instructions here, head.next
		 * can't change from under us thanks to the
3045
		 * anon_vma->root->rwsem.
3046 3047
		 */
		if (__test_and_set_bit(0, (unsigned long *)
3048
				       &anon_vma->root->rb_root.rb_node))
3049 3050 3051 3052
			BUG();
	}
}

3053
static void vm_lock_mapping(struct mm_struct *mm, struct address_space *mapping)
3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066
{
	if (!test_bit(AS_MM_ALL_LOCKS, &mapping->flags)) {
		/*
		 * AS_MM_ALL_LOCKS can't change from under us because
		 * we hold the mm_all_locks_mutex.
		 *
		 * Operations on ->flags have to be atomic because
		 * even if AS_MM_ALL_LOCKS is stable thanks to the
		 * mm_all_locks_mutex, there may be other cpus
		 * changing other bitflags in parallel to us.
		 */
		if (test_and_set_bit(AS_MM_ALL_LOCKS, &mapping->flags))
			BUG();
3067
		mutex_lock_nest_lock(&mapping->i_mmap_mutex, &mm->mmap_sem);
3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088
	}
}

/*
 * This operation locks against the VM for all pte/vma/mm related
 * operations that could ever happen on a certain mm. This includes
 * vmtruncate, try_to_unmap, and all page faults.
 *
 * The caller must take the mmap_sem in write mode before calling
 * mm_take_all_locks(). The caller isn't allowed to release the
 * mmap_sem until mm_drop_all_locks() returns.
 *
 * mmap_sem in write mode is required in order to block all operations
 * that could modify pagetables and free pages without need of
 * altering the vma layout (for example populate_range() with
 * nonlinear vmas). It's also needed in write mode to avoid new
 * anon_vmas to be associated with existing vmas.
 *
 * A single task can't take more than one mm_take_all_locks() in a row
 * or it would deadlock.
 *
3089
 * The LSB in anon_vma->rb_root.rb_node and the AS_MM_ALL_LOCKS bitflag in
3090 3091 3092 3093
 * mapping->flags avoid to take the same lock twice, if more than one
 * vma in this mm is backed by the same anon_vma or address_space.
 *
 * We can take all the locks in random order because the VM code
3094
 * taking i_mmap_mutex or anon_vma->rwsem outside the mmap_sem never
3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105
 * takes more than one of them in a row. Secondly we're protected
 * against a concurrent mm_take_all_locks() by the mm_all_locks_mutex.
 *
 * mm_take_all_locks() and mm_drop_all_locks are expensive operations
 * that may have to take thousand of locks.
 *
 * mm_take_all_locks() can fail if it's interrupted by signals.
 */
int mm_take_all_locks(struct mm_struct *mm)
{
	struct vm_area_struct *vma;
3106
	struct anon_vma_chain *avc;
3107 3108 3109 3110 3111 3112 3113 3114 3115

	BUG_ON(down_read_trylock(&mm->mmap_sem));

	mutex_lock(&mm_all_locks_mutex);

	for (vma = mm->mmap; vma; vma = vma->vm_next) {
		if (signal_pending(current))
			goto out_unlock;
		if (vma->vm_file && vma->vm_file->f_mapping)
3116
			vm_lock_mapping(mm, vma->vm_file->f_mapping);
3117
	}
3118 3119 3120 3121 3122

	for (vma = mm->mmap; vma; vma = vma->vm_next) {
		if (signal_pending(current))
			goto out_unlock;
		if (vma->anon_vma)
3123 3124
			list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
				vm_lock_anon_vma(mm, avc->anon_vma);
3125
	}
3126

3127
	return 0;
3128 3129

out_unlock:
3130 3131
	mm_drop_all_locks(mm);
	return -EINTR;
3132 3133 3134 3135
}

static void vm_unlock_anon_vma(struct anon_vma *anon_vma)
{
3136
	if (test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) {
3137 3138 3139 3140 3141
		/*
		 * The LSB of head.next can't change to 0 from under
		 * us because we hold the mm_all_locks_mutex.
		 *
		 * We must however clear the bitflag before unlocking
3142
		 * the vma so the users using the anon_vma->rb_root will
3143 3144 3145 3146
		 * never see our bitflag.
		 *
		 * No need of atomic instructions here, head.next
		 * can't change from under us until we release the
3147
		 * anon_vma->root->rwsem.
3148 3149
		 */
		if (!__test_and_clear_bit(0, (unsigned long *)
3150
					  &anon_vma->root->rb_root.rb_node))
3151
			BUG();
3152
		anon_vma_unlock_write(anon_vma);
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162
	}
}

static void vm_unlock_mapping(struct address_space *mapping)
{
	if (test_bit(AS_MM_ALL_LOCKS, &mapping->flags)) {
		/*
		 * AS_MM_ALL_LOCKS can't change to 0 from under us
		 * because we hold the mm_all_locks_mutex.
		 */
3163
		mutex_unlock(&mapping->i_mmap_mutex);
3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
		if (!test_and_clear_bit(AS_MM_ALL_LOCKS,
					&mapping->flags))
			BUG();
	}
}

/*
 * The mmap_sem cannot be released by the caller until
 * mm_drop_all_locks() returns.
 */
void mm_drop_all_locks(struct mm_struct *mm)
{
	struct vm_area_struct *vma;
3177
	struct anon_vma_chain *avc;
3178 3179 3180 3181 3182 3183

	BUG_ON(down_read_trylock(&mm->mmap_sem));
	BUG_ON(!mutex_is_locked(&mm_all_locks_mutex));

	for (vma = mm->mmap; vma; vma = vma->vm_next) {
		if (vma->anon_vma)
3184 3185
			list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
				vm_unlock_anon_vma(avc->anon_vma);
3186 3187 3188 3189 3190 3191
		if (vma->vm_file && vma->vm_file->f_mapping)
			vm_unlock_mapping(vma->vm_file->f_mapping);
	}

	mutex_unlock(&mm_all_locks_mutex);
}
3192 3193 3194 3195 3196 3197

/*
 * initialise the VMA slab
 */
void __init mmap_init(void)
{
3198 3199 3200 3201
	int ret;

	ret = percpu_counter_init(&vm_committed_as, 0);
	VM_BUG_ON(ret);
3202
}
3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213

/*
 * Initialise sysctl_user_reserve_kbytes.
 *
 * This is intended to prevent a user from starting a single memory hogging
 * process, such that they cannot recover (kill the hog) in OVERCOMMIT_NEVER
 * mode.
 *
 * The default value is min(3% of free memory, 128MB)
 * 128MB is enough to recover with sshd/login, bash, and top/kill.
 */
3214
static int init_user_reserve(void)
3215 3216 3217 3218 3219 3220 3221 3222
{
	unsigned long free_kbytes;

	free_kbytes = global_page_state(NR_FREE_PAGES) << (PAGE_SHIFT - 10);

	sysctl_user_reserve_kbytes = min(free_kbytes / 32, 1UL << 17);
	return 0;
}
3223
subsys_initcall(init_user_reserve);
3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234

/*
 * Initialise sysctl_admin_reserve_kbytes.
 *
 * The purpose of sysctl_admin_reserve_kbytes is to allow the sys admin
 * to log in and kill a memory hogging process.
 *
 * Systems with more than 256MB will reserve 8MB, enough to recover
 * with sshd, bash, and top in OVERCOMMIT_GUESS. Smaller systems will
 * only reserve 3% of free pages by default.
 */
3235
static int init_admin_reserve(void)
3236 3237 3238 3239 3240 3241 3242 3243
{
	unsigned long free_kbytes;

	free_kbytes = global_page_state(NR_FREE_PAGES) << (PAGE_SHIFT - 10);

	sysctl_admin_reserve_kbytes = min(free_kbytes / 32, 1UL << 13);
	return 0;
}
3244
subsys_initcall(init_admin_reserve);
3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309

/*
 * Reinititalise user and admin reserves if memory is added or removed.
 *
 * The default user reserve max is 128MB, and the default max for the
 * admin reserve is 8MB. These are usually, but not always, enough to
 * enable recovery from a memory hogging process using login/sshd, a shell,
 * and tools like top. It may make sense to increase or even disable the
 * reserve depending on the existence of swap or variations in the recovery
 * tools. So, the admin may have changed them.
 *
 * If memory is added and the reserves have been eliminated or increased above
 * the default max, then we'll trust the admin.
 *
 * If memory is removed and there isn't enough free memory, then we
 * need to reset the reserves.
 *
 * Otherwise keep the reserve set by the admin.
 */
static int reserve_mem_notifier(struct notifier_block *nb,
			     unsigned long action, void *data)
{
	unsigned long tmp, free_kbytes;

	switch (action) {
	case MEM_ONLINE:
		/* Default max is 128MB. Leave alone if modified by operator. */
		tmp = sysctl_user_reserve_kbytes;
		if (0 < tmp && tmp < (1UL << 17))
			init_user_reserve();

		/* Default max is 8MB.  Leave alone if modified by operator. */
		tmp = sysctl_admin_reserve_kbytes;
		if (0 < tmp && tmp < (1UL << 13))
			init_admin_reserve();

		break;
	case MEM_OFFLINE:
		free_kbytes = global_page_state(NR_FREE_PAGES) << (PAGE_SHIFT - 10);

		if (sysctl_user_reserve_kbytes > free_kbytes) {
			init_user_reserve();
			pr_info("vm.user_reserve_kbytes reset to %lu\n",
				sysctl_user_reserve_kbytes);
		}

		if (sysctl_admin_reserve_kbytes > free_kbytes) {
			init_admin_reserve();
			pr_info("vm.admin_reserve_kbytes reset to %lu\n",
				sysctl_admin_reserve_kbytes);
		}
		break;
	default:
		break;
	}
	return NOTIFY_OK;
}

static struct notifier_block reserve_mem_nb = {
	.notifier_call = reserve_mem_notifier,
};

static int __meminit init_reserve_notifier(void)
{
	if (register_hotmemory_notifier(&reserve_mem_nb))
3310
		pr_err("Failed registering memory add/remove notifier for admin reserve\n");
3311 3312 3313

	return 0;
}
3314
subsys_initcall(init_reserve_notifier);