mmap.c 98.3 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5
/*
 * mm/mmap.c
 *
 * Written by obz.
 *
Alan Cox's avatar
Alan Cox committed
6
 * Address space accounting code	<alan@lxorguk.ukuu.org.uk>
Linus Torvalds's avatar
Linus Torvalds committed
7 8
 */

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

11
#include <linux/kernel.h>
Linus Torvalds's avatar
Linus Torvalds committed
12
#include <linux/slab.h>
13
#include <linux/backing-dev.h>
Linus Torvalds's avatar
Linus Torvalds committed
14
#include <linux/mm.h>
Davidlohr Bueso's avatar
Davidlohr Bueso committed
15
#include <linux/vmacache.h>
Linus Torvalds's avatar
Linus Torvalds committed
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>
Linus Torvalds's avatar
Linus Torvalds committed
22 23 24 25 26 27
#include <linux/init.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/personality.h>
#include <linux/security.h>
#include <linux/hugetlb.h>
28
#include <linux/shmem_fs.h>
Linus Torvalds's avatar
Linus Torvalds committed
29
#include <linux/profile.h>
30
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
31 32 33
#include <linux/mount.h>
#include <linux/mempolicy.h>
#include <linux/rmap.h>
Andrea Arcangeli's avatar
Andrea Arcangeli committed
34
#include <linux/mmu_notifier.h>
35
#include <linux/mmdebug.h>
36
#include <linux/perf_event.h>
Al Viro's avatar
Al Viro committed
37
#include <linux/audit.h>
Andrea Arcangeli's avatar
Andrea Arcangeli committed
38
#include <linux/khugepaged.h>
39
#include <linux/uprobes.h>
40
#include <linux/rbtree_augmented.h>
41 42
#include <linux/notifier.h>
#include <linux/memory.h>
43
#include <linux/printk.h>
44
#include <linux/userfaultfd_k.h>
45
#include <linux/moduleparam.h>
46
#include <linux/pkeys.h>
47
#include <linux/oom.h>
Linus Torvalds's avatar
Linus Torvalds committed
48

49
#include <linux/uaccess.h>
Linus Torvalds's avatar
Linus Torvalds committed
50 51
#include <asm/cacheflush.h>
#include <asm/tlb.h>
52
#include <asm/mmu_context.h>
Linus Torvalds's avatar
Linus Torvalds committed
53

54 55
#include "internal.h"

56 57 58 59
#ifndef arch_mmap_check
#define arch_mmap_check(addr, len, flags)	(0)
#endif

60 61 62 63 64 65 66 67 68 69 70
#ifdef CONFIG_HAVE_ARCH_MMAP_RND_BITS
const int mmap_rnd_bits_min = CONFIG_ARCH_MMAP_RND_BITS_MIN;
const int mmap_rnd_bits_max = CONFIG_ARCH_MMAP_RND_BITS_MAX;
int mmap_rnd_bits __read_mostly = CONFIG_ARCH_MMAP_RND_BITS;
#endif
#ifdef CONFIG_HAVE_ARCH_MMAP_RND_COMPAT_BITS
const int mmap_rnd_compat_bits_min = CONFIG_ARCH_MMAP_RND_COMPAT_BITS_MIN;
const int mmap_rnd_compat_bits_max = CONFIG_ARCH_MMAP_RND_COMPAT_BITS_MAX;
int mmap_rnd_compat_bits __read_mostly = CONFIG_ARCH_MMAP_RND_COMPAT_BITS;
#endif

71
static bool ignore_rlimit_data;
72
core_param(ignore_rlimit_data, ignore_rlimit_data, bool, 0644);
73

74 75 76 77
static void unmap_region(struct mm_struct *mm,
		struct vm_area_struct *vma, struct vm_area_struct *prev,
		unsigned long start, unsigned long end);

Linus Torvalds's avatar
Linus Torvalds committed
78 79 80 81 82 83 84 85 86
/* 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
vishnu.ps's avatar
vishnu.ps committed
87
 *
Linus Torvalds's avatar
Linus Torvalds committed
88 89 90 91
 * 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
 *
92 93 94 95 96
 * On arm64, PROT_EXEC has the following behaviour for both MAP_SHARED and
 * MAP_PRIVATE:
 *								r: (no) no
 *								w: (no) no
 *								x: (yes) yes
Linus Torvalds's avatar
Linus Torvalds committed
97
 */
98
pgprot_t protection_map[16] __ro_after_init = {
Linus Torvalds's avatar
Linus Torvalds committed
99 100 101 102
	__P000, __P001, __P010, __P011, __P100, __P101, __P110, __P111,
	__S000, __S001, __S010, __S011, __S100, __S101, __S110, __S111
};

103 104
pgprot_t vm_get_page_prot(unsigned long vm_flags)
{
105 106 107
	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)));
108 109 110
}
EXPORT_SYMBOL(vm_get_page_prot);

111 112 113 114 115 116 117 118 119
static pgprot_t vm_pgprot_modify(pgprot_t oldprot, unsigned long vm_flags)
{
	return pgprot_modify(oldprot, vm_get_page_prot(vm_flags));
}

/* Update vma->vm_page_prot to reflect vma->vm_flags. */
void vma_set_page_prot(struct vm_area_struct *vma)
{
	unsigned long vm_flags = vma->vm_flags;
120
	pgprot_t vm_page_prot;
121

122 123
	vm_page_prot = vm_pgprot_modify(vma->vm_page_prot, vm_flags);
	if (vma_wants_writenotify(vma, vm_page_prot)) {
124
		vm_flags &= ~VM_SHARED;
125
		vm_page_prot = vm_pgprot_modify(vm_page_prot, vm_flags);
126
	}
127 128
	/* remove_protection_ptes reads vma->vm_page_prot without mmap_sem */
	WRITE_ONCE(vma->vm_page_prot, vm_page_prot);
129 130
}

Linus Torvalds's avatar
Linus Torvalds committed
131
/*
132
 * Requires inode->i_mapping->i_mmap_rwsem
Linus Torvalds's avatar
Linus Torvalds committed
133 134 135 136 137
 */
static void __remove_shared_vm_struct(struct vm_area_struct *vma,
		struct file *file, struct address_space *mapping)
{
	if (vma->vm_flags & VM_DENYWRITE)
Al Viro's avatar
Al Viro committed
138
		atomic_inc(&file_inode(file)->i_writecount);
Linus Torvalds's avatar
Linus Torvalds committed
139
	if (vma->vm_flags & VM_SHARED)
140
		mapping_unmap_writable(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
141 142

	flush_dcache_mmap_lock(mapping);
143
	vma_interval_tree_remove(vma, &mapping->i_mmap);
Linus Torvalds's avatar
Linus Torvalds committed
144 145 146 147
	flush_dcache_mmap_unlock(mapping);
}

/*
148
 * Unlink a file-based vm structure from its interval tree, to hide
149
 * vma from rmap and vmtruncate before freeing its page tables.
Linus Torvalds's avatar
Linus Torvalds committed
150
 */
151
void unlink_file_vma(struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
152 153 154 155 156
{
	struct file *file = vma->vm_file;

	if (file) {
		struct address_space *mapping = file->f_mapping;
157
		i_mmap_lock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
158
		__remove_shared_vm_struct(vma, file, mapping);
159
		i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
160
	}
161 162 163 164 165 166 167 168 169 170
}

/*
 * 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();
Linus Torvalds's avatar
Linus Torvalds committed
171 172
	if (vma->vm_ops && vma->vm_ops->close)
		vma->vm_ops->close(vma);
173
	if (vma->vm_file)
174
		fput(vma->vm_file);
175
	mpol_put(vma_policy(vma));
Linus Torvalds's avatar
Linus Torvalds committed
176
	kmem_cache_free(vm_area_cachep, vma);
177
	return next;
Linus Torvalds's avatar
Linus Torvalds committed
178 179
}

180
static int do_brk(unsigned long addr, unsigned long len, struct list_head *uf);
181

182
SYSCALL_DEFINE1(brk, unsigned long, brk)
Linus Torvalds's avatar
Linus Torvalds committed
183
{
184
	unsigned long retval;
Linus Torvalds's avatar
Linus Torvalds committed
185 186
	unsigned long newbrk, oldbrk;
	struct mm_struct *mm = current->mm;
187
	struct vm_area_struct *next;
188
	unsigned long min_brk;
189
	bool populate;
190
	LIST_HEAD(uf);
Linus Torvalds's avatar
Linus Torvalds committed
191

192 193
	if (down_write_killable(&mm->mmap_sem))
		return -EINTR;
Linus Torvalds's avatar
Linus Torvalds committed
194

195
#ifdef CONFIG_COMPAT_BRK
196 197 198 199 200
	/*
	 * 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
	 */
201
	if (current->brk_randomized)
202 203 204
		min_brk = mm->start_brk;
	else
		min_brk = mm->end_data;
205 206 207 208
#else
	min_brk = mm->start_brk;
#endif
	if (brk < min_brk)
Linus Torvalds's avatar
Linus Torvalds committed
209
		goto out;
Ram Gupta's avatar
Ram Gupta committed
210 211 212 213 214 215 216

	/*
	 * 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
	 */
217 218
	if (check_data_rlimit(rlimit(RLIMIT_DATA), brk, mm->start_brk,
			      mm->end_data, mm->start_data))
Ram Gupta's avatar
Ram Gupta committed
219 220
		goto out;

Linus Torvalds's avatar
Linus Torvalds committed
221 222 223 224 225 226 227
	newbrk = PAGE_ALIGN(brk);
	oldbrk = PAGE_ALIGN(mm->brk);
	if (oldbrk == newbrk)
		goto set_brk;

	/* Always allow shrinking brk. */
	if (brk <= mm->brk) {
228
		if (!do_munmap(mm, newbrk, oldbrk-newbrk, &uf))
Linus Torvalds's avatar
Linus Torvalds committed
229 230 231 232 233
			goto set_brk;
		goto out;
	}

	/* Check against existing mmap mappings. */
234 235
	next = find_vma(mm, oldbrk);
	if (next && newbrk + PAGE_SIZE > vm_start_gap(next))
Linus Torvalds's avatar
Linus Torvalds committed
236 237 238
		goto out;

	/* Ok, looks good - let it rip. */
239
	if (do_brk(oldbrk, newbrk-oldbrk, &uf) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
240
		goto out;
241

Linus Torvalds's avatar
Linus Torvalds committed
242 243
set_brk:
	mm->brk = brk;
244 245
	populate = newbrk > oldbrk && (mm->def_flags & VM_LOCKED) != 0;
	up_write(&mm->mmap_sem);
246
	userfaultfd_unmap_complete(mm, &uf);
247 248 249 250
	if (populate)
		mm_populate(oldbrk, newbrk - oldbrk);
	return brk;

Linus Torvalds's avatar
Linus Torvalds committed
251 252 253 254 255 256
out:
	retval = mm->brk;
	up_write(&mm->mmap_sem);
	return retval;
}

257 258
static long vma_compute_subtree_gap(struct vm_area_struct *vma)
{
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
	unsigned long max, prev_end, subtree_gap;

	/*
	 * Note: in the rare case of a VM_GROWSDOWN above a VM_GROWSUP, we
	 * allow two stack_guard_gaps between them here, and when choosing
	 * an unmapped area; whereas when expanding we only require one.
	 * That's a little inconsistent, but keeps the code here simpler.
	 */
	max = vm_start_gap(vma);
	if (vma->vm_prev) {
		prev_end = vm_end_gap(vma->vm_prev);
		if (max > prev_end)
			max -= prev_end;
		else
			max = 0;
	}
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
	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;
}

290
#ifdef CONFIG_DEBUG_VM_RB
291
static int browse_rb(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
292
{
293
	struct rb_root *root = &mm->mm_rb;
294
	int i = 0, j, bug = 0;
Linus Torvalds's avatar
Linus Torvalds committed
295 296 297 298 299 300
	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);
301
		if (vma->vm_start < prev) {
302 303
			pr_emerg("vm_start %lx < prev %lx\n",
				  vma->vm_start, prev);
304 305 306
			bug = 1;
		}
		if (vma->vm_start < pend) {
307 308
			pr_emerg("vm_start %lx < pend %lx\n",
				  vma->vm_start, pend);
309 310 311
			bug = 1;
		}
		if (vma->vm_start > vma->vm_end) {
312 313
			pr_emerg("vm_start %lx > vm_end %lx\n",
				  vma->vm_start, vma->vm_end);
314 315
			bug = 1;
		}
316
		spin_lock(&mm->page_table_lock);
317
		if (vma->rb_subtree_gap != vma_compute_subtree_gap(vma)) {
318
			pr_emerg("free gap %lx, correct %lx\n",
319 320 321 322
			       vma->rb_subtree_gap,
			       vma_compute_subtree_gap(vma));
			bug = 1;
		}
323
		spin_unlock(&mm->page_table_lock);
Linus Torvalds's avatar
Linus Torvalds committed
324 325
		i++;
		pn = nd;
326 327
		prev = vma->vm_start;
		pend = vma->vm_end;
Linus Torvalds's avatar
Linus Torvalds committed
328 329
	}
	j = 0;
330
	for (nd = pn; nd; nd = rb_prev(nd))
Linus Torvalds's avatar
Linus Torvalds committed
331
		j++;
332
	if (i != j) {
333
		pr_emerg("backwards %d, forwards %d\n", j, i);
334
		bug = 1;
Linus Torvalds's avatar
Linus Torvalds committed
335
	}
336
	return bug ? -1 : i;
Linus Torvalds's avatar
Linus Torvalds committed
337 338
}

339 340 341 342 343 344 345
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);
346 347 348
		VM_BUG_ON_VMA(vma != ignore &&
			vma->rb_subtree_gap != vma_compute_subtree_gap(vma),
			vma);
Linus Torvalds's avatar
Linus Torvalds committed
349 350 351
	}
}

352
static void validate_mm(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
353 354 355
{
	int bug = 0;
	int i = 0;
356
	unsigned long highest_address = 0;
357
	struct vm_area_struct *vma = mm->mmap;
358

359
	while (vma) {
360
		struct anon_vma *anon_vma = vma->anon_vma;
361
		struct anon_vma_chain *avc;
362

363 364 365 366 367 368 369
		if (anon_vma) {
			anon_vma_lock_read(anon_vma);
			list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
				anon_vma_interval_tree_verify(avc);
			anon_vma_unlock_read(anon_vma);
		}

370
		highest_address = vm_end_gap(vma);
371
		vma = vma->vm_next;
Linus Torvalds's avatar
Linus Torvalds committed
372 373
		i++;
	}
374
	if (i != mm->map_count) {
375
		pr_emerg("map_count %d vm_next %d\n", mm->map_count, i);
376 377 378
		bug = 1;
	}
	if (highest_address != mm->highest_vm_end) {
379
		pr_emerg("mm->highest_vm_end %lx, found %lx\n",
380
			  mm->highest_vm_end, highest_address);
381 382
		bug = 1;
	}
383
	i = browse_rb(mm);
384
	if (i != mm->map_count) {
385 386
		if (i != -1)
			pr_emerg("map_count %d rb %d\n", mm->map_count, i);
387 388
		bug = 1;
	}
389
	VM_BUG_ON_MM(bug, mm);
Linus Torvalds's avatar
Linus Torvalds committed
390 391
}
#else
392
#define validate_mm_rb(root, ignore) do { } while (0)
Linus Torvalds's avatar
Linus Torvalds committed
393 394 395
#define validate_mm(mm) do { } while (0)
#endif

396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
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);
}

422
static void __vma_rb_erase(struct vm_area_struct *vma, struct rb_root *root)
423 424 425 426 427 428 429 430 431
{
	/*
	 * 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);
}

432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
static __always_inline void vma_rb_erase_ignore(struct vm_area_struct *vma,
						struct rb_root *root,
						struct vm_area_struct *ignore)
{
	/*
	 * All rb_subtree_gap values must be consistent prior to erase,
	 * with the possible exception of the "next" vma being erased if
	 * next->vm_start was reduced.
	 */
	validate_mm_rb(root, ignore);

	__vma_rb_erase(vma, root);
}

static __always_inline 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);

	__vma_rb_erase(vma, root);
}

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
/*
 * 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);
}

490 491 492
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)
Linus Torvalds's avatar
Linus Torvalds committed
493
{
494
	struct rb_node **__rb_link, *__rb_parent, *rb_prev;
Linus Torvalds's avatar
Linus Torvalds committed
495 496 497 498 499 500 501 502 503 504 505

	__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) {
506 507 508
			/* Fail if an existing vma overlaps the area */
			if (vma_tmp->vm_start < end)
				return -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
509 510 511 512 513 514 515 516 517 518 519 520
			__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;
521
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
522 523
}

524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551
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;
}

Linus Torvalds's avatar
Linus Torvalds committed
552 553 554
void __vma_link_rb(struct mm_struct *mm, struct vm_area_struct *vma,
		struct rb_node **rb_link, struct rb_node *rb_parent)
{
555 556 557 558
	/* Update tracking information for the gap following the new vma. */
	if (vma->vm_next)
		vma_gap_update(vma->vm_next);
	else
559
		mm->highest_vm_end = vm_end_gap(vma);
560 561 562 563 564 565 566 567 568 569

	/*
	 * 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.
	 */
Linus Torvalds's avatar
Linus Torvalds committed
570
	rb_link_node(&vma->vm_rb, rb_parent, rb_link);
571 572 573
	vma->rb_subtree_gap = 0;
	vma_gap_update(vma);
	vma_rb_insert(vma, &mm->mm_rb);
Linus Torvalds's avatar
Linus Torvalds committed
574 575
}

576
static void __vma_link_file(struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
577
{
ZhenwenXu's avatar
ZhenwenXu committed
578
	struct file *file;
Linus Torvalds's avatar
Linus Torvalds committed
579 580 581 582 583 584

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

		if (vma->vm_flags & VM_DENYWRITE)
Al Viro's avatar
Al Viro committed
585
			atomic_dec(&file_inode(file)->i_writecount);
Linus Torvalds's avatar
Linus Torvalds committed
586
		if (vma->vm_flags & VM_SHARED)
587
			atomic_inc(&mapping->i_mmap_writable);
Linus Torvalds's avatar
Linus Torvalds committed
588 589

		flush_dcache_mmap_lock(mapping);
590
		vma_interval_tree_insert(vma, &mapping->i_mmap);
Linus Torvalds's avatar
Linus Torvalds committed
591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
		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;

610
	if (vma->vm_file) {
Linus Torvalds's avatar
Linus Torvalds committed
611
		mapping = vma->vm_file->f_mapping;
612
		i_mmap_lock_write(mapping);
613
	}
Linus Torvalds's avatar
Linus Torvalds committed
614 615 616 617 618

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

	if (mapping)
619
		i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
620 621 622 623 624 625

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

/*
626
 * Helper for vma_adjust() in the split_vma insert case: insert a vma into the
627
 * mm's list and rbtree.  It has already been inserted into the interval tree.
Linus Torvalds's avatar
Linus Torvalds committed
628
 */
ZhenwenXu's avatar
ZhenwenXu committed
629
static void __insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
630
{
631
	struct vm_area_struct *prev;
ZhenwenXu's avatar
ZhenwenXu committed
632
	struct rb_node **rb_link, *rb_parent;
Linus Torvalds's avatar
Linus Torvalds committed
633

634 635 636
	if (find_vma_links(mm, vma->vm_start, vma->vm_end,
			   &prev, &rb_link, &rb_parent))
		BUG();
Linus Torvalds's avatar
Linus Torvalds committed
637 638 639 640
	__vma_link(mm, vma, prev, rb_link, rb_parent);
	mm->map_count++;
}

641 642 643
static __always_inline void __vma_unlink_common(struct mm_struct *mm,
						struct vm_area_struct *vma,
						struct vm_area_struct *prev,
644 645
						bool has_prev,
						struct vm_area_struct *ignore)
Linus Torvalds's avatar
Linus Torvalds committed
646
{
647
	struct vm_area_struct *next;
648

649
	vma_rb_erase_ignore(vma, &mm->mm_rb, ignore);
650 651 652 653 654 655 656 657 658 659
	next = vma->vm_next;
	if (has_prev)
		prev->vm_next = next;
	else {
		prev = vma->vm_prev;
		if (prev)
			prev->vm_next = next;
		else
			mm->mmap = next;
	}
660 661
	if (next)
		next->vm_prev = prev;
Davidlohr Bueso's avatar
Davidlohr Bueso committed
662 663 664

	/* Kill the cache */
	vmacache_invalidate(mm);
Linus Torvalds's avatar
Linus Torvalds committed
665 666
}

667 668 669 670
static inline void __vma_unlink_prev(struct mm_struct *mm,
				     struct vm_area_struct *vma,
				     struct vm_area_struct *prev)
{
671
	__vma_unlink_common(mm, vma, prev, true, vma);
672 673
}

Linus Torvalds's avatar
Linus Torvalds committed
674 675 676 677 678 679 680
/*
 * 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.
 */
681 682 683
int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
	unsigned long end, pgoff_t pgoff, struct vm_area_struct *insert,
	struct vm_area_struct *expand)
Linus Torvalds's avatar
Linus Torvalds committed
684 685
{
	struct mm_struct *mm = vma->vm_mm;
686
	struct vm_area_struct *next = vma->vm_next, *orig_vma = vma;
Linus Torvalds's avatar
Linus Torvalds committed
687
	struct address_space *mapping = NULL;
688
	struct rb_root_cached *root = NULL;
689
	struct anon_vma *anon_vma = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
690
	struct file *file = vma->vm_file;
691
	bool start_changed = false, end_changed = false;
Linus Torvalds's avatar
Linus Torvalds committed
692 693 694 695
	long adjust_next = 0;
	int remove_next = 0;

	if (next && !insert) {
696
		struct vm_area_struct *exporter = NULL, *importer = NULL;
697

Linus Torvalds's avatar
Linus Torvalds committed
698 699 700 701
		if (end >= next->vm_end) {
			/*
			 * vma expands, overlapping all the next, and
			 * perhaps the one after too (mprotect case 6).
702
			 * The only other cases that gets here are
703
			 * case 1, case 7 and case 8.
Linus Torvalds's avatar
Linus Torvalds committed
704
			 */
705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
			if (next == expand) {
				/*
				 * The only case where we don't expand "vma"
				 * and we expand "next" instead is case 8.
				 */
				VM_WARN_ON(end != next->vm_end);
				/*
				 * remove_next == 3 means we're
				 * removing "vma" and that to do so we
				 * swapped "vma" and "next".
				 */
				remove_next = 3;
				VM_WARN_ON(file != next->vm_file);
				swap(vma, next);
			} else {
				VM_WARN_ON(expand != vma);
				/*
				 * case 1, 6, 7, remove_next == 2 is case 6,
				 * remove_next == 1 is case 1 or 7.
				 */
				remove_next = 1 + (end > next->vm_end);
				VM_WARN_ON(remove_next == 2 &&
					   end != next->vm_next->vm_end);
				VM_WARN_ON(remove_next == 1 &&
					   end != next->vm_end);
				/* trim end to next, for case 6 first pass */
				end = next->vm_end;
			}

734
			exporter = next;
Linus Torvalds's avatar
Linus Torvalds committed
735
			importer = vma;
736 737 738 739 740

			/*
			 * If next doesn't have anon_vma, import from vma after
			 * next, if the vma overlaps with it.
			 */
741
			if (remove_next == 2 && !next->anon_vma)
742 743
				exporter = next->vm_next;

Linus Torvalds's avatar
Linus Torvalds committed
744 745 746 747 748 749
		} 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;
750
			exporter = next;
Linus Torvalds's avatar
Linus Torvalds committed
751
			importer = vma;
752
			VM_WARN_ON(expand != importer);
Linus Torvalds's avatar
Linus Torvalds committed
753 754 755 756 757 758
		} 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.
			 */
vishnu.ps's avatar
vishnu.ps committed
759
			adjust_next = -((vma->vm_end - end) >> PAGE_SHIFT);
760
			exporter = vma;
Linus Torvalds's avatar
Linus Torvalds committed
761
			importer = next;
762
			VM_WARN_ON(expand != importer);
Linus Torvalds's avatar
Linus Torvalds committed
763 764
		}

765 766 767 768 769
		/*
		 * 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.
		 */
770
		if (exporter && exporter->anon_vma && !importer->anon_vma) {
771 772
			int error;

773
			importer->anon_vma = exporter->anon_vma;
774
			error = anon_vma_clone(importer, exporter);
775
			if (error)
776
				return error;
777 778
		}
	}
779
again:
780
	vma_adjust_trans_huge(orig_vma, start, end, adjust_next);
781

Linus Torvalds's avatar
Linus Torvalds committed
782 783
	if (file) {
		mapping = file->f_mapping;
784 785
		root = &mapping->i_mmap;
		uprobe_munmap(vma, vma->vm_start, vma->vm_end);
786

787 788
		if (adjust_next)
			uprobe_munmap(next, next->vm_start, next->vm_end);
789

790
		i_mmap_lock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
791 792
		if (insert) {
			/*
793
			 * Put into interval tree now, so instantiated pages
Linus Torvalds's avatar
Linus Torvalds committed
794 795 796 797 798 799 800 801
			 * 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);
		}
	}

802 803 804 805
	anon_vma = vma->anon_vma;
	if (!anon_vma && adjust_next)
		anon_vma = next->anon_vma;
	if (anon_vma) {
806 807
		VM_WARN_ON(adjust_next && next->anon_vma &&
			   anon_vma != next->anon_vma);
808
		anon_vma_lock_write(anon_vma);
809 810 811 812
		anon_vma_interval_tree_pre_update_vma(vma);
		if (adjust_next)
			anon_vma_interval_tree_pre_update_vma(next);
	}
813

Linus Torvalds's avatar
Linus Torvalds committed
814 815
	if (root) {
		flush_dcache_mmap_lock(mapping);
816
		vma_interval_tree_remove(vma, root);
Linus Torvalds's avatar
Linus Torvalds committed
817
		if (adjust_next)
818
			vma_interval_tree_remove(next, root);
Linus Torvalds's avatar
Linus Torvalds committed
819 820
	}

821 822 823 824 825 826 827 828
	if (start != vma->vm_start) {
		vma->vm_start = start;
		start_changed = true;
	}
	if (end != vma->vm_end) {
		vma->vm_end = end;
		end_changed = true;
	}
Linus Torvalds's avatar
Linus Torvalds committed
829 830 831 832 833 834 835 836
	vma->vm_pgoff = pgoff;
	if (adjust_next) {
		next->vm_start += adjust_next << PAGE_SHIFT;
		next->vm_pgoff += adjust_next;
	}

	if (root) {
		if (adjust_next)
837 838
			vma_interval_tree_insert(next, root);
		vma_interval_tree_insert(vma, root);
Linus Torvalds's avatar
Linus Torvalds committed
839 840 841 842 843 844 845 846
		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.
		 */
847 848 849
		if (remove_next != 3)
			__vma_unlink_prev(mm, next, vma);
		else
850 851 852 853 854 855 856 857 858 859
			/*
			 * vma is not before next if they've been
			 * swapped.
			 *
			 * pre-swap() next->vm_start was reduced so
			 * tell validate_mm_rb to ignore pre-swap()
			 * "next" (which is stored in post-swap()
			 * "vma").
			 */
			__vma_unlink_common(mm, next, NULL, false, vma);
Linus Torvalds's avatar
Linus Torvalds committed
860 861 862 863 864 865 866 867 868
		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);
869 870 871 872 873
	} else {
		if (start_changed)
			vma_gap_update(vma);
		if (end_changed) {
			if (!next)
874
				mm->highest_vm_end = vm_end_gap(vma);
875 876 877
			else if (!adjust_next)
				vma_gap_update(next);
		}
Linus Torvalds's avatar
Linus Torvalds committed
878 879
	}

880 881 882 883
	if (anon_vma) {
		anon_vma_interval_tree_post_update_vma(vma);
		if (adjust_next)
			anon_vma_interval_tree_post_update_vma(next);
884
		anon_vma_unlock_write(anon_vma);
885
	}
Linus Torvalds's avatar
Linus Torvalds committed
886
	if (mapping)
887
		i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
888

889
	if (root) {
890
		uprobe_mmap(vma);
891 892

		if (adjust_next)
893
			uprobe_mmap(next);
894 895
	}

Linus Torvalds's avatar
Linus Torvalds committed
896
	if (remove_next) {
Matt Helsley's avatar
Matt Helsley committed
897
		if (file) {
898
			uprobe_munmap(next, next->vm_start, next->vm_end);
Linus Torvalds's avatar
Linus Torvalds committed
899
			fput(file);
Matt Helsley's avatar
Matt Helsley committed
900
		}
901 902
		if (next->anon_vma)
			anon_vma_merge(vma, next);
Linus Torvalds's avatar
Linus Torvalds committed
903
		mm->map_count--;
904
		mpol_put(vma_policy(next));
Linus Torvalds's avatar
Linus Torvalds committed
905 906 907 908 909 910
		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.
		 */
911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931
		if (remove_next != 3) {
			/*
			 * If "next" was removed and vma->vm_end was
			 * expanded (up) over it, in turn
			 * "next->vm_prev->vm_end" changed and the
			 * "vma->vm_next" gap must be updated.
			 */
			next = vma->vm_next;
		} else {
			/*
			 * For the scope of the comment "next" and
			 * "vma" considered pre-swap(): if "vma" was
			 * removed, next->vm_start was expanded (down)
			 * over it and the "next" gap must be updated.
			 * Because of the swap() the post-swap() "vma"
			 * actually points to pre-swap() "next"
			 * (post-swap() "next" as opposed is now a
			 * dangling pointer).
			 */
			next = vma;
		}
932 933 934
		if (remove_next == 2) {
			remove_next = 1;
			end = next->vm_end;
Linus Torvalds's avatar
Linus Torvalds committed
935
			goto again;
936
		}
937 938
		else if (next)
			vma_gap_update(next);
939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958
		else {
			/*
			 * If remove_next == 2 we obviously can't
			 * reach this path.
			 *
			 * If remove_next == 3 we can't reach this
			 * path because pre-swap() next is always not
			 * NULL. pre-swap() "next" is not being
			 * removed and its next->vm_end is not altered
			 * (and furthermore "end" already matches
			 * next->vm_end in remove_next == 3).
			 *
			 * We reach this only in the remove_next == 1
			 * case if the "next" vma that was removed was
			 * the highest vma of the mm. However in such
			 * case next->vm_end == "end" and the extended
			 * "vma" has vma->vm_end == next->vm_end so
			 * mm->highest_vm_end doesn't need any update
			 * in remove_next == 1 case.
			 */
959
			VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma));
960
		}
Linus Torvalds's avatar
Linus Torvalds committed
961
	}
962
	if (insert && file)
963
		uprobe_mmap(insert);
Linus Torvalds's avatar
Linus Torvalds committed
964 965

	validate_mm(mm);
966 967

	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
968 969 970 971 972 973 974
}

/*
 * 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,
975 976
				struct file *file, unsigned long vm_flags,
				struct vm_userfaultfd_ctx vm_userfaultfd_ctx)
Linus Torvalds's avatar
Linus Torvalds committed
977
{
978 979 980 981 982 983 984 985 986
	/*
	 * 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)
Linus Torvalds's avatar
Linus Torvalds committed
987 988 989 990 991
		return 0;
	if (vma->vm_file != file)
		return 0;
	if (vma->vm_ops && vma->vm_ops->close)
		return 0;
992 993
	if (!is_mergeable_vm_userfaultfd_ctx(vma, vm_userfaultfd_ctx))
		return 0;
Linus Torvalds's avatar
Linus Torvalds committed
994 995 996 997
	return 1;
}

static inline int is_mergeable_anon_vma(struct anon_vma *anon_vma1,
Shaohua Li's avatar
Shaohua Li committed
998 999
					struct anon_vma *anon_vma2,
					struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
1000
{
Shaohua Li's avatar
Shaohua Li committed
1001 1002 1003 1004 1005 1006 1007 1008
	/*
	 * 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;
Linus Torvalds's avatar
Linus Torvalds committed
1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
}

/*
 * 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,
1024 1025 1026
		     struct anon_vma *anon_vma, struct file *file,
		     pgoff_t vm_pgoff,
		     struct vm_userfaultfd_ctx vm_userfaultfd_ctx)
Linus Torvalds's avatar
Linus Torvalds committed
1027
{
1028
	if (is_mergeable_vma(vma, file, vm_flags, vm_userfaultfd_ctx) &&
Shaohua Li's avatar
Shaohua Li committed
1029
	    is_mergeable_anon_vma(anon_vma, vma->anon_vma, vma)) {
Linus Torvalds's avatar
Linus Torvalds committed
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044
		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,
1045 1046 1047
		    struct anon_vma *anon_vma, struct file *file,
		    pgoff_t vm_pgoff,
		    struct vm_userfaultfd_ctx vm_userfaultfd_ctx)
Linus Torvalds's avatar
Linus Torvalds committed
1048
{
1049
	if (is_mergeable_vma(vma, file, vm_flags, vm_userfaultfd_ctx) &&
Shaohua Li's avatar
Shaohua Li committed
1050
	    is_mergeable_anon_vma(anon_vma, vma->anon_vma, vma)) {
Linus Torvalds's avatar
Linus Torvalds committed
1051
		pgoff_t vm_pglen;
1052
		vm_pglen = vma_pages(vma);
Linus Torvalds's avatar
Linus Torvalds committed
1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079
		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
1080
 *    mremap move:                                    PPPPXXXXXXXX 8
Linus Torvalds's avatar
Linus Torvalds committed
1081 1082 1083 1084
 *        AAAA
 *    PPPP    NNNN    PPPPPPPPPPPP    PPPPPPPPNNNN    PPPPNNNNNNNN
 *    might become    case 1 below    case 2 below    case 3 below
 *
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097
 * It is important for case 8 that the the vma NNNN overlapping the
 * region AAAA is never going to extended over XXXX. Instead XXXX must
 * be extended in region AAAA and NNNN must be removed. This way in
 * all cases where vma_merge succeeds, the moment vma_adjust drops the
 * rmap_locks, the properties of the merged vma will be already
 * correct for the whole merged range. Some of those properties like
 * vm_page_prot/vm_flags may be accessed by rmap_walks and they must
 * be correct for the whole merged range immediately after the
 * rmap_locks are released. Otherwise if XXXX would be removed and
 * NNNN would be extended over the XXXX range, remove_migration_ptes
 * or other rmap walkers (if working on addresses beyond the "end"
 * parameter) may establish ptes with the wrong permissions of NNNN
 * instead of the right permissions of XXXX.
Linus Torvalds's avatar
Linus Torvalds committed
1098 1099 1100 1101
 */
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,
vishnu.ps's avatar
vishnu.ps committed
1102
			struct anon_vma *anon_vma, struct file *file,
1103 1104
			pgoff_t pgoff, struct mempolicy *policy,
			struct vm_userfaultfd_ctx vm_userfaultfd_ctx)
Linus Torvalds's avatar
Linus Torvalds committed
1105 1106 1107
{
	pgoff_t pglen = (end - addr) >> PAGE_SHIFT;
	struct vm_area_struct *area, *next;
1108
	int err;
Linus Torvalds's avatar
Linus Torvalds committed
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121

	/*
	 * 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;
1122
	if (area && area->vm_end == end)		/* cases 6, 7, 8 */
Linus Torvalds's avatar
Linus Torvalds committed
1123 1124
		next = next->vm_next;

1125 1126 1127 1128 1129
	/* verify some invariant that must be enforced by the caller */
	VM_WARN_ON(prev && addr <= prev->vm_start);
	VM_WARN_ON(area && end > area->vm_end);
	VM_WARN_ON(addr >= end);

Linus Torvalds's avatar
Linus Torvalds committed
1130 1131 1132 1133
	/*
	 * Can it merge with the predecessor?
	 */
	if (prev && prev->vm_end == addr &&
vishnu.ps's avatar
vishnu.ps committed
1134
			mpol_equal(vma_policy(prev), policy) &&
Linus Torvalds's avatar
Linus Torvalds committed
1135
			can_vma_merge_after(prev, vm_flags,
1136 1137
					    anon_vma, file, pgoff,
					    vm_userfaultfd_ctx)) {
Linus Torvalds's avatar
Linus Torvalds committed
1138 1139 1140 1141 1142 1143
		/*
		 * 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,
1144 1145 1146
						     anon_vma, file,
						     pgoff+pglen,
						     vm_userfaultfd_ctx) &&
Linus Torvalds's avatar
Linus Torvalds committed
1147
				is_mergeable_anon_vma(prev->anon_vma,
Shaohua Li's avatar
Shaohua Li committed
1148
						      next->anon_vma, NULL)) {
Linus Torvalds's avatar
Linus Torvalds committed
1149
							/* cases 1, 6 */
1150 1151 1152
			err = __vma_adjust(prev, prev->vm_start,
					 next->vm_end, prev->vm_pgoff, NULL,
					 prev);
Linus Torvalds's avatar
Linus Torvalds committed
1153
		} else					/* cases 2, 5, 7 */
1154 1155
			err = __vma_adjust(prev, prev->vm_start,
					 end, prev->vm_pgoff, NULL, prev);
1156 1157
		if (err)
			return NULL;
1158
		khugepaged_enter_vma_merge(prev, vm_flags);
Linus Torvalds's avatar
Linus Torvalds committed
1159 1160 1161 1162 1163 1164 1165
		return prev;
	}

	/*
	 * Can this new request be merged in front of next?
	 */
	if (next && end == next->vm_start &&
vishnu.ps's avatar
vishnu.ps committed
1166
			mpol_equal(policy, vma_policy(next)) &&
Linus Torvalds's avatar
Linus Torvalds committed
1167
			can_vma_merge_before(next, vm_flags,
1168 1169
					     anon_vma, file, pgoff+pglen,
					     vm_userfaultfd_ctx)) {
Linus Torvalds's avatar
Linus Torvalds committed
1170
		if (prev && addr < prev->vm_end)	/* case 4 */
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
			err = __vma_adjust(prev, prev->vm_start,
					 addr, prev->vm_pgoff, NULL, next);
		else {					/* cases 3, 8 */
			err = __vma_adjust(area, addr, next->vm_end,
					 next->vm_pgoff - pglen, NULL, next);
			/*
			 * In case 3 area is already equal to next and
			 * this is a noop, but in case 8 "area" has
			 * been removed and next was expanded over it.
			 */
			area = next;
		}
1183 1184
		if (err)
			return NULL;
1185
		khugepaged_enter_vma_merge(area, vm_flags);
Linus Torvalds's avatar
Linus Torvalds committed
1186 1187 1188 1189 1190 1191
		return area;
	}

	return NULL;
}

1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
/*
 * 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 &&
1210
		!((a->vm_flags ^ b->vm_flags) & ~(VM_READ|VM_WRITE|VM_EXEC|VM_SOFTDIRTY)) &&
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224
		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
1225
 * we do that READ_ONCE() to make sure that we never re-load the pointer.
1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
 *
 * 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)) {
1239
		struct anon_vma *anon_vma = READ_ONCE(old->anon_vma);
1240 1241 1242 1243 1244 1245 1246

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

Linus Torvalds's avatar
Linus Torvalds committed
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256
/*
 * 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)
{
1257
	struct anon_vma *anon_vma;
Linus Torvalds's avatar
Linus Torvalds committed
1258 1259 1260 1261 1262 1263
	struct vm_area_struct *near;

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

1264 1265 1266
	anon_vma = reusable_anon_vma(near, vma, near);
	if (anon_vma)
		return anon_vma;
Linus Torvalds's avatar
Linus Torvalds committed
1267
try_prev:
1268
	near = vma->vm_prev;
Linus Torvalds's avatar
Linus Torvalds committed
1269 1270 1271
	if (!near)
		goto none;

1272 1273 1274
	anon_vma = reusable_anon_vma(near, near, vma);
	if (anon_vma)
		return anon_vma;
Linus Torvalds's avatar
Linus Torvalds committed
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286
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;
}

Al Viro's avatar
Al Viro committed
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299
/*
 * 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;
}

1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
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;
}

Linus Torvalds's avatar
Linus Torvalds committed
1318
/*
1319
 * The caller must hold down_write(&current->mm->mmap_sem).
Linus Torvalds's avatar
Linus Torvalds committed
1320
 */
1321
unsigned long do_mmap(struct file *file, unsigned long addr,
Linus Torvalds's avatar
Linus Torvalds committed
1322
			unsigned long len, unsigned long prot,
1323
			unsigned long flags, vm_flags_t vm_flags,
1324 1325
			unsigned long pgoff, unsigned long *populate,
			struct list_head *uf)
Linus Torvalds's avatar
Linus Torvalds committed
1326
{
vishnu.ps's avatar
vishnu.ps committed
1327
	struct mm_struct *mm = current->mm;
1328
	int pkey = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1329

1330
	*populate = 0;
1331

1332 1333 1334
	if (!len)
		return -EINVAL;

Linus Torvalds's avatar
Linus Torvalds committed
1335 1336 1337 1338 1339 1340 1341
	/*
	 * 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))
1342
		if (!(file && path_noexec(&file->f_path)))
Linus Torvalds's avatar
Linus Torvalds committed
1343 1344
			prot |= PROT_EXEC;

1345 1346 1347 1348
	/* force arch specific MAP_FIXED handling in get_unmapped_area */
	if (flags & MAP_FIXED_NOREPLACE)
		flags |= MAP_FIXED;

1349 1350 1351
	if (!(flags & MAP_FIXED))
		addr = round_hint_to_min(addr);

Linus Torvalds's avatar
Linus Torvalds committed
1352 1353
	/* Careful about overflows.. */
	len = PAGE_ALIGN(len);
1354
	if (!len)
Linus Torvalds's avatar
Linus Torvalds committed
1355 1356 1357 1358
		return -ENOMEM;

	/* offset overflow? */
	if ((pgoff + (len >> PAGE_SHIFT)) < pgoff)
vishnu.ps's avatar
vishnu.ps committed
1359
		return -EOVERFLOW;
Linus Torvalds's avatar
Linus Torvalds committed
1360 1361 1362 1363 1364 1365 1366 1367 1368

	/* 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);
1369
	if (offset_in_page(addr))
Linus Torvalds's avatar
Linus Torvalds committed
1370 1371
		return addr;

1372 1373 1374 1375 1376 1377 1378
	if (flags & MAP_FIXED_NOREPLACE)