nommu.c 48.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8
/*
 *  linux/mm/nommu.c
 *
 *  Replacement code for mm functions to support CPU's that don't
 *  have any form of memory management unit (thus no virtual memory).
 *
 *  See Documentation/nommu-mmap.txt
 *
9
 *  Copyright (c) 2004-2008 David Howells <dhowells@redhat.com>
L
Linus Torvalds 已提交
10 11 12
 *  Copyright (c) 2000-2003 David McCullough <davidm@snapgear.com>
 *  Copyright (c) 2000-2001 D Jeff Dionne <jeff@uClinux.org>
 *  Copyright (c) 2002      Greg Ungerer <gerg@snapgear.com>
13
 *  Copyright (c) 2007-2010 Paul Mundt <lethal@linux-sh.org>
L
Linus Torvalds 已提交
14 15
 */

16 17
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

18
#include <linux/export.h>
L
Linus Torvalds 已提交
19
#include <linux/mm.h>
D
Davidlohr Bueso 已提交
20
#include <linux/vmacache.h>
L
Linus Torvalds 已提交
21 22 23 24 25 26 27 28 29
#include <linux/mman.h>
#include <linux/swap.h>
#include <linux/file.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/blkdev.h>
#include <linux/backing-dev.h>
30
#include <linux/compiler.h>
L
Linus Torvalds 已提交
31 32 33 34
#include <linux/mount.h>
#include <linux/personality.h>
#include <linux/security.h>
#include <linux/syscalls.h>
A
Al Viro 已提交
35
#include <linux/audit.h>
36
#include <linux/printk.h>
L
Linus Torvalds 已提交
37 38 39 40

#include <asm/uaccess.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>
41
#include <asm/mmu_context.h>
42 43
#include "internal.h"

L
Linus Torvalds 已提交
44
void *high_memory;
45
EXPORT_SYMBOL(high_memory);
L
Linus Torvalds 已提交
46 47
struct page *mem_map;
unsigned long max_mapnr;
48
EXPORT_SYMBOL(max_mapnr);
H
Hugh Dickins 已提交
49
unsigned long highest_memmap_pfn;
50
int sysctl_nr_trim_pages = CONFIG_NOMMU_INITIAL_TRIM_EXCESS;
L
Linus Torvalds 已提交
51 52
int heap_stack_gap = 0;

53
atomic_long_t mmap_pages_allocated;
54

L
Linus Torvalds 已提交
55 56
EXPORT_SYMBOL(mem_map);

57 58 59 60
/* list of mapped, potentially shareable regions */
static struct kmem_cache *vm_region_jar;
struct rb_root nommu_region_tree = RB_ROOT;
DECLARE_RWSEM(nommu_region_sem);
L
Linus Torvalds 已提交
61

62
const struct vm_operations_struct generic_file_vm_ops = {
L
Linus Torvalds 已提交
63 64 65 66 67 68 69 70 71 72 73 74
};

/*
 * Return the total memory allocated for this pointer, not
 * just what the caller asked for.
 *
 * Doesn't have to be accurate, i.e. may have races.
 */
unsigned int kobjsize(const void *objp)
{
	struct page *page;

75 76 77 78
	/*
	 * If the object we have should not have ksize performed on it,
	 * return size of 0
	 */
79
	if (!objp || !virt_addr_valid(objp))
80 81 82 83 84 85 86 87
		return 0;

	page = virt_to_head_page(objp);

	/*
	 * If the allocator sets PageSlab, we know the pointer came from
	 * kmalloc().
	 */
L
Linus Torvalds 已提交
88 89 90
	if (PageSlab(page))
		return ksize(objp);

91 92 93 94 95 96 97 98 99 100 101 102 103 104
	/*
	 * If it's not a compound page, see if we have a matching VMA
	 * region. This test is intentionally done in reverse order,
	 * so if there's no VMA, we still fall through and hand back
	 * PAGE_SIZE for 0-order pages.
	 */
	if (!PageCompound(page)) {
		struct vm_area_struct *vma;

		vma = find_vma(current->mm, (unsigned long)objp);
		if (vma)
			return vma->vm_end - vma->vm_start;
	}

105 106
	/*
	 * The ksize() function is only guaranteed to work for pointers
107
	 * returned by kmalloc(). So handle arbitrary pointers here.
108
	 */
109
	return PAGE_SIZE << compound_order(page);
L
Linus Torvalds 已提交
110 111
}

112 113 114 115
long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
		      unsigned long start, unsigned long nr_pages,
		      unsigned int foll_flags, struct page **pages,
		      struct vm_area_struct **vmas, int *nonblocking)
L
Linus Torvalds 已提交
116
{
117
	struct vm_area_struct *vma;
118 119 120 121
	unsigned long vm_flags;
	int i;

	/* calculate required read or write permissions.
H
Hugh Dickins 已提交
122
	 * If FOLL_FORCE is set, we only require the "MAY" flags.
123
	 */
H
Hugh Dickins 已提交
124 125 126 127
	vm_flags  = (foll_flags & FOLL_WRITE) ?
			(VM_WRITE | VM_MAYWRITE) : (VM_READ | VM_MAYREAD);
	vm_flags &= (foll_flags & FOLL_FORCE) ?
			(VM_MAYREAD | VM_MAYWRITE) : (VM_READ | VM_WRITE);
L
Linus Torvalds 已提交
128

129
	for (i = 0; i < nr_pages; i++) {
130
		vma = find_vma(mm, start);
131 132 133 134
		if (!vma)
			goto finish_or_fault;

		/* protect what we can, including chardevs */
H
Hugh Dickins 已提交
135 136
		if ((vma->vm_flags & (VM_IO | VM_PFNMAP)) ||
		    !(vm_flags & vma->vm_flags))
137
			goto finish_or_fault;
138

L
Linus Torvalds 已提交
139 140 141 142 143 144
		if (pages) {
			pages[i] = virt_to_page(start);
			if (pages[i])
				page_cache_get(pages[i]);
		}
		if (vmas)
145
			vmas[i] = vma;
146
		start = (start + PAGE_SIZE) & PAGE_MASK;
L
Linus Torvalds 已提交
147
	}
148 149 150 151 152

	return i;

finish_or_fault:
	return i ? : -EFAULT;
L
Linus Torvalds 已提交
153
}
N
Nick Piggin 已提交
154 155 156 157 158 159 160 161

/*
 * get a list of pages in an address range belonging to the specified process
 * and indicate the VMA that covers each page
 * - this is potentially dodgy as we may end incrementing the page count of a
 *   slab page or a secondary page from a compound page
 * - don't permit access to VMAs that don't support it, such as I/O mappings
 */
162 163 164 165
long get_user_pages(struct task_struct *tsk, struct mm_struct *mm,
		    unsigned long start, unsigned long nr_pages,
		    int write, int force, struct page **pages,
		    struct vm_area_struct **vmas)
N
Nick Piggin 已提交
166 167 168 169
{
	int flags = 0;

	if (write)
H
Hugh Dickins 已提交
170
		flags |= FOLL_WRITE;
N
Nick Piggin 已提交
171
	if (force)
H
Hugh Dickins 已提交
172
		flags |= FOLL_FORCE;
N
Nick Piggin 已提交
173

174 175
	return __get_user_pages(tsk, mm, start, nr_pages, flags, pages, vmas,
				NULL);
N
Nick Piggin 已提交
176
}
177 178
EXPORT_SYMBOL(get_user_pages);

179 180 181 182 183 184 185 186 187 188
long get_user_pages_locked(struct task_struct *tsk, struct mm_struct *mm,
			   unsigned long start, unsigned long nr_pages,
			   int write, int force, struct page **pages,
			   int *locked)
{
	return get_user_pages(tsk, mm, start, nr_pages, write, force,
			      pages, NULL);
}
EXPORT_SYMBOL(get_user_pages_locked);

189 190 191 192
long __get_user_pages_unlocked(struct task_struct *tsk, struct mm_struct *mm,
			       unsigned long start, unsigned long nr_pages,
			       int write, int force, struct page **pages,
			       unsigned int gup_flags)
193 194 195 196 197 198 199 200
{
	long ret;
	down_read(&mm->mmap_sem);
	ret = get_user_pages(tsk, mm, start, nr_pages, write, force,
			     pages, NULL);
	up_read(&mm->mmap_sem);
	return ret;
}
201 202 203 204 205 206 207 208 209
EXPORT_SYMBOL(__get_user_pages_unlocked);

long get_user_pages_unlocked(struct task_struct *tsk, struct mm_struct *mm,
			     unsigned long start, unsigned long nr_pages,
			     int write, int force, struct page **pages)
{
	return __get_user_pages_unlocked(tsk, mm, start, nr_pages, write,
					 force, pages, 0);
}
210 211
EXPORT_SYMBOL(get_user_pages_unlocked);

P
Paul Mundt 已提交
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
/**
 * follow_pfn - look up PFN at a user virtual address
 * @vma: memory mapping
 * @address: user virtual address
 * @pfn: location to store found PFN
 *
 * Only IO mappings and raw PFN mappings are allowed.
 *
 * Returns zero and the pfn at @pfn on success, -ve otherwise.
 */
int follow_pfn(struct vm_area_struct *vma, unsigned long address,
	unsigned long *pfn)
{
	if (!(vma->vm_flags & (VM_IO | VM_PFNMAP)))
		return -EINVAL;

	*pfn = address >> PAGE_SHIFT;
	return 0;
}
EXPORT_SYMBOL(follow_pfn);

233
LIST_HEAD(vmap_area_list);
L
Linus Torvalds 已提交
234

235
void vfree(const void *addr)
L
Linus Torvalds 已提交
236 237 238
{
	kfree(addr);
}
239
EXPORT_SYMBOL(vfree);
L
Linus Torvalds 已提交
240

A
Al Viro 已提交
241
void *__vmalloc(unsigned long size, gfp_t gfp_mask, pgprot_t prot)
L
Linus Torvalds 已提交
242 243
{
	/*
244 245
	 *  You can't specify __GFP_HIGHMEM with kmalloc() since kmalloc()
	 * returns only a logical address.
L
Linus Torvalds 已提交
246
	 */
247
	return kmalloc(size, (gfp_mask | __GFP_COMP) & ~__GFP_HIGHMEM);
L
Linus Torvalds 已提交
248
}
249
EXPORT_SYMBOL(__vmalloc);
L
Linus Torvalds 已提交
250

251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
void *vmalloc_user(unsigned long size)
{
	void *ret;

	ret = __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO,
			PAGE_KERNEL);
	if (ret) {
		struct vm_area_struct *vma;

		down_write(&current->mm->mmap_sem);
		vma = find_vma(current->mm, (unsigned long)ret);
		if (vma)
			vma->vm_flags |= VM_USERMAP;
		up_write(&current->mm->mmap_sem);
	}

	return ret;
}
EXPORT_SYMBOL(vmalloc_user);

271
struct page *vmalloc_to_page(const void *addr)
L
Linus Torvalds 已提交
272 273 274
{
	return virt_to_page(addr);
}
275
EXPORT_SYMBOL(vmalloc_to_page);
L
Linus Torvalds 已提交
276

277
unsigned long vmalloc_to_pfn(const void *addr)
L
Linus Torvalds 已提交
278 279 280
{
	return page_to_pfn(virt_to_page(addr));
}
281
EXPORT_SYMBOL(vmalloc_to_pfn);
L
Linus Torvalds 已提交
282 283 284

long vread(char *buf, char *addr, unsigned long count)
{
285 286 287 288
	/* Don't allow overflow */
	if ((unsigned long) buf + count < count)
		count = -(unsigned long) buf;

L
Linus Torvalds 已提交
289 290 291 292 293 294 295 296 297 298 299
	memcpy(buf, addr, count);
	return count;
}

long vwrite(char *buf, char *addr, unsigned long count)
{
	/* Don't allow overflow */
	if ((unsigned long) addr + count < count)
		count = -(unsigned long) addr;

	memcpy(addr, buf, count);
300
	return count;
L
Linus Torvalds 已提交
301 302 303
}

/*
304
 *	vmalloc  -  allocate virtually contiguous memory
L
Linus Torvalds 已提交
305 306 307 308
 *
 *	@size:		allocation size
 *
 *	Allocate enough pages to cover @size from the page level
309
 *	allocator and map them into contiguous kernel virtual space.
L
Linus Torvalds 已提交
310
 *
311
 *	For tight control over page level allocator and protection flags
L
Linus Torvalds 已提交
312 313 314 315 316 317
 *	use __vmalloc() instead.
 */
void *vmalloc(unsigned long size)
{
       return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
}
318 319
EXPORT_SYMBOL(vmalloc);

320
/*
321
 *	vzalloc - allocate virtually contiguous memory with zero fill
322 323 324 325
 *
 *	@size:		allocation size
 *
 *	Allocate enough pages to cover @size from the page level
326
 *	allocator and map them into contiguous kernel virtual space.
327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349
 *	The memory allocated is set to zero.
 *
 *	For tight control over page level allocator and protection flags
 *	use __vmalloc() instead.
 */
void *vzalloc(unsigned long size)
{
	return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM | __GFP_ZERO,
			PAGE_KERNEL);
}
EXPORT_SYMBOL(vzalloc);

/**
 * vmalloc_node - allocate memory on a specific node
 * @size:	allocation size
 * @node:	numa node
 *
 * Allocate enough pages to cover @size from the page level
 * allocator and map them into contiguous kernel virtual space.
 *
 * For tight control over page level allocator and protection flags
 * use __vmalloc() instead.
 */
350 351 352 353
void *vmalloc_node(unsigned long size, int node)
{
	return vmalloc(size);
}
354
EXPORT_SYMBOL(vmalloc_node);
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372

/**
 * vzalloc_node - allocate memory on a specific node with zero fill
 * @size:	allocation size
 * @node:	numa node
 *
 * Allocate enough pages to cover @size from the page level
 * allocator and map them into contiguous kernel virtual space.
 * The memory allocated is set to zero.
 *
 * For tight control over page level allocator and protection flags
 * use __vmalloc() instead.
 */
void *vzalloc_node(unsigned long size, int node)
{
	return vzalloc(size);
}
EXPORT_SYMBOL(vzalloc_node);
L
Linus Torvalds 已提交
373

P
Paul Mundt 已提交
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
#ifndef PAGE_KERNEL_EXEC
# define PAGE_KERNEL_EXEC PAGE_KERNEL
#endif

/**
 *	vmalloc_exec  -  allocate virtually contiguous, executable memory
 *	@size:		allocation size
 *
 *	Kernel-internal function to allocate enough pages to cover @size
 *	the page level allocator and map them into contiguous and
 *	executable kernel virtual space.
 *
 *	For tight control over page level allocator and protection flags
 *	use __vmalloc() instead.
 */

void *vmalloc_exec(unsigned long size)
{
	return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL_EXEC);
}

395 396
/**
 * vmalloc_32  -  allocate virtually contiguous memory (32bit addressable)
L
Linus Torvalds 已提交
397 398 399
 *	@size:		allocation size
 *
 *	Allocate enough 32bit PA addressable pages to cover @size from the
400
 *	page level allocator and map them into contiguous kernel virtual space.
L
Linus Torvalds 已提交
401 402 403 404 405
 */
void *vmalloc_32(unsigned long size)
{
	return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL);
}
406 407 408 409 410 411 412 413
EXPORT_SYMBOL(vmalloc_32);

/**
 * vmalloc_32_user - allocate zeroed virtually contiguous 32bit memory
 *	@size:		allocation size
 *
 * The resulting memory area is 32bit addressable and zeroed so it can be
 * mapped to userspace without leaking data.
414 415 416
 *
 * VM_USERMAP is set on the corresponding VMA so that subsequent calls to
 * remap_vmalloc_range() are permissible.
417 418 419
 */
void *vmalloc_32_user(unsigned long size)
{
420 421 422 423 424
	/*
	 * We'll have to sort out the ZONE_DMA bits for 64-bit,
	 * but for now this can simply use vmalloc_user() directly.
	 */
	return vmalloc_user(size);
425 426
}
EXPORT_SYMBOL(vmalloc_32_user);
L
Linus Torvalds 已提交
427 428 429 430 431 432

void *vmap(struct page **pages, unsigned int count, unsigned long flags, pgprot_t prot)
{
	BUG();
	return NULL;
}
433
EXPORT_SYMBOL(vmap);
L
Linus Torvalds 已提交
434

435
void vunmap(const void *addr)
L
Linus Torvalds 已提交
436 437 438
{
	BUG();
}
439
EXPORT_SYMBOL(vunmap);
L
Linus Torvalds 已提交
440

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
void *vm_map_ram(struct page **pages, unsigned int count, int node, pgprot_t prot)
{
	BUG();
	return NULL;
}
EXPORT_SYMBOL(vm_map_ram);

void vm_unmap_ram(const void *mem, unsigned int count)
{
	BUG();
}
EXPORT_SYMBOL(vm_unmap_ram);

void vm_unmap_aliases(void)
{
}
EXPORT_SYMBOL_GPL(vm_unmap_aliases);

459 460 461 462
/*
 * Implement a stub for vmalloc_sync_all() if the architecture chose not to
 * have one.
 */
463
void __weak vmalloc_sync_all(void)
464 465 466
{
}

467 468 469 470 471 472 473 474 475 476 477 478
/**
 *	alloc_vm_area - allocate a range of kernel address space
 *	@size:		size of the area
 *
 *	Returns:	NULL on failure, vm_struct on success
 *
 *	This function reserves a range of kernel address space, and
 *	allocates pagetables to map that range.  No actual mappings
 *	are created.  If the kernel address space is not shared
 *	between processes, it syncs the pagetable across all
 *	processes.
 */
479
struct vm_struct *alloc_vm_area(size_t size, pte_t **ptes)
480 481 482 483 484 485 486 487 488 489 490 491
{
	BUG();
	return NULL;
}
EXPORT_SYMBOL_GPL(alloc_vm_area);

void free_vm_area(struct vm_struct *area)
{
	BUG();
}
EXPORT_SYMBOL_GPL(free_vm_area);

492 493 494 495 496 497 498
int vm_insert_page(struct vm_area_struct *vma, unsigned long addr,
		   struct page *page)
{
	return -EINVAL;
}
EXPORT_SYMBOL(vm_insert_page);

L
Linus Torvalds 已提交
499 500 501 502 503 504 505
/*
 *  sys_brk() for the most part doesn't need the global kernel
 *  lock, except when an application is doing something nasty
 *  like trying to un-brk an area that has already been mapped
 *  to a regular file.  in this case, the unmapping will need
 *  to invoke file system routines that need the global lock.
 */
506
SYSCALL_DEFINE1(brk, unsigned long, brk)
L
Linus Torvalds 已提交
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526
{
	struct mm_struct *mm = current->mm;

	if (brk < mm->start_brk || brk > mm->context.end_brk)
		return mm->brk;

	if (mm->brk == brk)
		return mm->brk;

	/*
	 * Always allow shrinking brk
	 */
	if (brk <= mm->brk) {
		mm->brk = brk;
		return brk;
	}

	/*
	 * Ok, looks good - let it rip.
	 */
527
	flush_icache_range(mm->brk, brk);
L
Linus Torvalds 已提交
528 529 530
	return mm->brk = brk;
}

531 532 533 534
/*
 * initialise the VMA and region record slabs
 */
void __init mmap_init(void)
L
Linus Torvalds 已提交
535
{
536 537
	int ret;

538
	ret = percpu_counter_init(&vm_committed_as, 0, GFP_KERNEL);
539
	VM_BUG_ON(ret);
540
	vm_region_jar = KMEM_CACHE(vm_region, SLAB_PANIC|SLAB_ACCOUNT);
L
Linus Torvalds 已提交
541 542
}

543
/*
544 545
 * validate the region tree
 * - the caller must hold the region lock
546
 */
547 548
#ifdef CONFIG_DEBUG_NOMMU_REGIONS
static noinline void validate_nommu_regions(void)
549
{
550 551
	struct vm_region *region, *last;
	struct rb_node *p, *lastp;
552

553 554 555 556 557
	lastp = rb_first(&nommu_region_tree);
	if (!lastp)
		return;

	last = rb_entry(lastp, struct vm_region, vm_rb);
558 559
	BUG_ON(last->vm_end <= last->vm_start);
	BUG_ON(last->vm_top < last->vm_end);
560 561 562 563 564

	while ((p = rb_next(lastp))) {
		region = rb_entry(p, struct vm_region, vm_rb);
		last = rb_entry(lastp, struct vm_region, vm_rb);

565 566 567
		BUG_ON(region->vm_end <= region->vm_start);
		BUG_ON(region->vm_top < region->vm_end);
		BUG_ON(region->vm_start < last->vm_top);
568

569 570
		lastp = p;
	}
571
}
572
#else
573 574 575
static void validate_nommu_regions(void)
{
}
576
#endif
577 578

/*
579
 * add a region into the global tree
580
 */
581
static void add_nommu_region(struct vm_region *region)
582
{
583 584
	struct vm_region *pregion;
	struct rb_node **p, *parent;
585

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600
	validate_nommu_regions();

	parent = NULL;
	p = &nommu_region_tree.rb_node;
	while (*p) {
		parent = *p;
		pregion = rb_entry(parent, struct vm_region, vm_rb);
		if (region->vm_start < pregion->vm_start)
			p = &(*p)->rb_left;
		else if (region->vm_start > pregion->vm_start)
			p = &(*p)->rb_right;
		else if (pregion == region)
			return;
		else
			BUG();
601 602
	}

603 604
	rb_link_node(&region->vm_rb, parent, p);
	rb_insert_color(&region->vm_rb, &nommu_region_tree);
605

606
	validate_nommu_regions();
607 608
}

609
/*
610
 * delete a region from the global tree
611
 */
612
static void delete_nommu_region(struct vm_region *region)
613
{
614
	BUG_ON(!nommu_region_tree.rb_node);
615

616 617 618
	validate_nommu_regions();
	rb_erase(&region->vm_rb, &nommu_region_tree);
	validate_nommu_regions();
619 620
}

621
/*
622
 * free a contiguous series of pages
623
 */
624
static void free_page_series(unsigned long from, unsigned long to)
625
{
626 627 628
	for (; from < to; from += PAGE_SIZE) {
		struct page *page = virt_to_page(from);

629
		atomic_long_dec(&mmap_pages_allocated);
630
		put_page(page);
631 632 633
	}
}

634
/*
635
 * release a reference to a region
636
 * - the caller must hold the region semaphore for writing, which this releases
637
 * - the region may not have been added to the tree yet, in which case vm_top
638
 *   will equal vm_start
639
 */
640 641
static void __put_nommu_region(struct vm_region *region)
	__releases(nommu_region_sem)
L
Linus Torvalds 已提交
642
{
643
	BUG_ON(!nommu_region_tree.rb_node);
L
Linus Torvalds 已提交
644

645
	if (--region->vm_usage == 0) {
646
		if (region->vm_top > region->vm_start)
647 648 649 650 651 652 653 654
			delete_nommu_region(region);
		up_write(&nommu_region_sem);

		if (region->vm_file)
			fput(region->vm_file);

		/* IO memory and memory shared directly out of the pagecache
		 * from ramfs/tmpfs mustn't be released here */
655
		if (region->vm_flags & VM_MAPPED_COPY)
656
			free_page_series(region->vm_start, region->vm_top);
657 658 659
		kmem_cache_free(vm_region_jar, region);
	} else {
		up_write(&nommu_region_sem);
L
Linus Torvalds 已提交
660
	}
661
}
L
Linus Torvalds 已提交
662

663 664 665 666 667 668 669
/*
 * release a reference to a region
 */
static void put_nommu_region(struct vm_region *region)
{
	down_write(&nommu_region_sem);
	__put_nommu_region(region);
L
Linus Torvalds 已提交
670 671
}

672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687
/*
 * update protection on a vma
 */
static void protect_vma(struct vm_area_struct *vma, unsigned long flags)
{
#ifdef CONFIG_MPU
	struct mm_struct *mm = vma->vm_mm;
	long start = vma->vm_start & PAGE_MASK;
	while (start < vma->vm_end) {
		protect_page(mm, start, flags);
		start += PAGE_SIZE;
	}
	update_protections(mm);
#endif
}

688
/*
689 690 691 692
 * add a VMA into a process's mm_struct in the appropriate place in the list
 * and tree and add to the address space's page tree also if not an anonymous
 * page
 * - should be called with mm->mmap_sem held writelocked
693
 */
694
static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma)
L
Linus Torvalds 已提交
695
{
696
	struct vm_area_struct *pvma, *prev;
L
Linus Torvalds 已提交
697
	struct address_space *mapping;
698
	struct rb_node **p, *parent, *rb_prev;
699 700 701 702 703

	BUG_ON(!vma->vm_region);

	mm->map_count++;
	vma->vm_mm = mm;
L
Linus Torvalds 已提交
704

705 706
	protect_vma(vma, vma->vm_flags);

L
Linus Torvalds 已提交
707 708 709 710
	/* add the VMA to the mapping */
	if (vma->vm_file) {
		mapping = vma->vm_file->f_mapping;

711
		i_mmap_lock_write(mapping);
L
Linus Torvalds 已提交
712
		flush_dcache_mmap_lock(mapping);
713
		vma_interval_tree_insert(vma, &mapping->i_mmap);
L
Linus Torvalds 已提交
714
		flush_dcache_mmap_unlock(mapping);
715
		i_mmap_unlock_write(mapping);
L
Linus Torvalds 已提交
716 717
	}

718
	/* add the VMA to the tree */
719
	parent = rb_prev = NULL;
720
	p = &mm->mm_rb.rb_node;
L
Linus Torvalds 已提交
721 722 723 724
	while (*p) {
		parent = *p;
		pvma = rb_entry(parent, struct vm_area_struct, vm_rb);

725 726 727
		/* sort by: start addr, end addr, VMA struct addr in that order
		 * (the latter is necessary as we may get identical VMAs) */
		if (vma->vm_start < pvma->vm_start)
L
Linus Torvalds 已提交
728
			p = &(*p)->rb_left;
729 730
		else if (vma->vm_start > pvma->vm_start) {
			rb_prev = parent;
L
Linus Torvalds 已提交
731
			p = &(*p)->rb_right;
732
		} else if (vma->vm_end < pvma->vm_end)
733
			p = &(*p)->rb_left;
734 735
		else if (vma->vm_end > pvma->vm_end) {
			rb_prev = parent;
736
			p = &(*p)->rb_right;
737
		} else if (vma < pvma)
738
			p = &(*p)->rb_left;
739 740
		else if (vma > pvma) {
			rb_prev = parent;
741
			p = &(*p)->rb_right;
742
		} else
743
			BUG();
L
Linus Torvalds 已提交
744 745 746
	}

	rb_link_node(&vma->vm_rb, parent, p);
747 748 749
	rb_insert_color(&vma->vm_rb, &mm->mm_rb);

	/* add VMA to the VMA list also */
750 751 752
	prev = NULL;
	if (rb_prev)
		prev = rb_entry(rb_prev, struct vm_area_struct, vm_rb);
753

754
	__vma_link_list(mm, vma, prev, parent);
L
Linus Torvalds 已提交
755 756
}

757
/*
758
 * delete a VMA from its owning mm_struct and address space
759
 */
760
static void delete_vma_from_mm(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
761
{
D
Davidlohr Bueso 已提交
762
	int i;
L
Linus Torvalds 已提交
763
	struct address_space *mapping;
764
	struct mm_struct *mm = vma->vm_mm;
D
Davidlohr Bueso 已提交
765
	struct task_struct *curr = current;
766

767 768
	protect_vma(vma, 0);

769
	mm->map_count--;
D
Davidlohr Bueso 已提交
770 771 772
	for (i = 0; i < VMACACHE_SIZE; i++) {
		/* if the vma is cached, invalidate the entire cache */
		if (curr->vmacache[i] == vma) {
773
			vmacache_invalidate(mm);
D
Davidlohr Bueso 已提交
774 775 776
			break;
		}
	}
L
Linus Torvalds 已提交
777 778 779 780 781

	/* remove the VMA from the mapping */
	if (vma->vm_file) {
		mapping = vma->vm_file->f_mapping;

782
		i_mmap_lock_write(mapping);
L
Linus Torvalds 已提交
783
		flush_dcache_mmap_lock(mapping);
784
		vma_interval_tree_remove(vma, &mapping->i_mmap);
L
Linus Torvalds 已提交
785
		flush_dcache_mmap_unlock(mapping);
786
		i_mmap_unlock_write(mapping);
L
Linus Torvalds 已提交
787 788
	}

789 790
	/* remove from the MM's tree and list */
	rb_erase(&vma->vm_rb, &mm->mm_rb);
791 792 793 794 795 796 797 798

	if (vma->vm_prev)
		vma->vm_prev->vm_next = vma->vm_next;
	else
		mm->mmap = vma->vm_next;

	if (vma->vm_next)
		vma->vm_next->vm_prev = vma->vm_prev;
799 800 801 802 803 804 805 806 807
}

/*
 * destroy a VMA record
 */
static void delete_vma(struct mm_struct *mm, struct vm_area_struct *vma)
{
	if (vma->vm_ops && vma->vm_ops->close)
		vma->vm_ops->close(vma);
808
	if (vma->vm_file)
809 810 811 812 813 814 815 816 817 818 819 820 821 822
		fput(vma->vm_file);
	put_nommu_region(vma->vm_region);
	kmem_cache_free(vm_area_cachep, vma);
}

/*
 * look up the first VMA in which addr resides, NULL if none
 * - should be called with mm->mmap_sem at least held readlocked
 */
struct vm_area_struct *find_vma(struct mm_struct *mm, unsigned long addr)
{
	struct vm_area_struct *vma;

	/* check the cache first */
D
Davidlohr Bueso 已提交
823 824
	vma = vmacache_find(mm, addr);
	if (likely(vma))
825 826
		return vma;

827
	/* trawl the list (there may be multiple mappings in which addr
828
	 * resides) */
829
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
830 831 832
		if (vma->vm_start > addr)
			return NULL;
		if (vma->vm_end > addr) {
D
Davidlohr Bueso 已提交
833
			vmacache_update(addr, vma);
834 835 836 837 838 839 840 841 842 843 844 845 846 847
			return vma;
		}
	}

	return NULL;
}
EXPORT_SYMBOL(find_vma);

/*
 * find a VMA
 * - we don't extend stack VMAs under NOMMU conditions
 */
struct vm_area_struct *find_extend_vma(struct mm_struct *mm, unsigned long addr)
{
848
	return find_vma(mm, addr);
849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
}

/*
 * expand a stack to a given address
 * - not supported under NOMMU conditions
 */
int expand_stack(struct vm_area_struct *vma, unsigned long address)
{
	return -ENOMEM;
}

/*
 * look up the first VMA exactly that exactly matches addr
 * - should be called with mm->mmap_sem at least held readlocked
 */
static struct vm_area_struct *find_vma_exact(struct mm_struct *mm,
					     unsigned long addr,
					     unsigned long len)
{
	struct vm_area_struct *vma;
	unsigned long end = addr + len;

	/* check the cache first */
D
Davidlohr Bueso 已提交
872 873
	vma = vmacache_find_exact(mm, addr, end);
	if (vma)
874 875
		return vma;

876
	/* trawl the list (there may be multiple mappings in which addr
877
	 * resides) */
878
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
879 880 881 882 883
		if (vma->vm_start < addr)
			continue;
		if (vma->vm_start > addr)
			return NULL;
		if (vma->vm_end == end) {
D
Davidlohr Bueso 已提交
884
			vmacache_update(addr, vma);
885 886 887 888 889
			return vma;
		}
	}

	return NULL;
L
Linus Torvalds 已提交
890 891 892 893 894 895 896 897 898 899 900 901 902 903
}

/*
 * determine whether a mapping should be permitted and, if so, what sort of
 * mapping we're capable of supporting
 */
static int validate_mmap_request(struct file *file,
				 unsigned long addr,
				 unsigned long len,
				 unsigned long prot,
				 unsigned long flags,
				 unsigned long pgoff,
				 unsigned long *_capabilities)
{
904
	unsigned long capabilities, rlen;
L
Linus Torvalds 已提交
905 906 907
	int ret;

	/* do the simple checks first */
908
	if (flags & MAP_FIXED)
L
Linus Torvalds 已提交
909 910 911 912 913 914
		return -EINVAL;

	if ((flags & MAP_TYPE) != MAP_PRIVATE &&
	    (flags & MAP_TYPE) != MAP_SHARED)
		return -EINVAL;

915
	if (!len)
L
Linus Torvalds 已提交
916 917
		return -EINVAL;

918
	/* Careful about overflows.. */
919 920
	rlen = PAGE_ALIGN(len);
	if (!rlen || rlen > TASK_SIZE)
921 922
		return -ENOMEM;

L
Linus Torvalds 已提交
923
	/* offset overflow? */
924
	if ((pgoff + (rlen >> PAGE_SHIFT)) < pgoff)
925
		return -EOVERFLOW;
L
Linus Torvalds 已提交
926 927 928

	if (file) {
		/* files must support mmap */
A
Al Viro 已提交
929
		if (!file->f_op->mmap)
L
Linus Torvalds 已提交
930 931 932 933 934 935
			return -ENODEV;

		/* work out if what we've got could possibly be shared
		 * - we support chardevs that provide their own "memory"
		 * - we support files/blockdevs that are memory backed
		 */
936 937 938
		if (file->f_op->mmap_capabilities) {
			capabilities = file->f_op->mmap_capabilities(file);
		} else {
L
Linus Torvalds 已提交
939 940
			/* no explicit capabilities set, so assume some
			 * defaults */
A
Al Viro 已提交
941
			switch (file_inode(file)->i_mode & S_IFMT) {
L
Linus Torvalds 已提交
942 943
			case S_IFREG:
			case S_IFBLK:
944
				capabilities = NOMMU_MAP_COPY;
L
Linus Torvalds 已提交
945 946 947 948
				break;

			case S_IFCHR:
				capabilities =
949 950 951
					NOMMU_MAP_DIRECT |
					NOMMU_MAP_READ |
					NOMMU_MAP_WRITE;
L
Linus Torvalds 已提交
952 953 954 955 956 957 958 959 960 961
				break;

			default:
				return -EINVAL;
			}
		}

		/* eliminate any capabilities that we can't support on this
		 * device */
		if (!file->f_op->get_unmapped_area)
962
			capabilities &= ~NOMMU_MAP_DIRECT;
A
Al Viro 已提交
963
		if (!(file->f_mode & FMODE_CAN_READ))
964
			capabilities &= ~NOMMU_MAP_COPY;
L
Linus Torvalds 已提交
965

966 967 968 969
		/* The file shall have been opened with read permission. */
		if (!(file->f_mode & FMODE_READ))
			return -EACCES;

L
Linus Torvalds 已提交
970 971 972 973 974 975
		if (flags & MAP_SHARED) {
			/* do checks for writing, appending and locking */
			if ((prot & PROT_WRITE) &&
			    !(file->f_mode & FMODE_WRITE))
				return -EACCES;

A
Al Viro 已提交
976
			if (IS_APPEND(file_inode(file)) &&
L
Linus Torvalds 已提交
977 978 979
			    (file->f_mode & FMODE_WRITE))
				return -EACCES;

980
			if (locks_verify_locked(file))
L
Linus Torvalds 已提交
981 982
				return -EAGAIN;

983
			if (!(capabilities & NOMMU_MAP_DIRECT))
L
Linus Torvalds 已提交
984 985 986
				return -ENODEV;

			/* we mustn't privatise shared mappings */
987
			capabilities &= ~NOMMU_MAP_COPY;
988
		} else {
L
Linus Torvalds 已提交
989 990
			/* we're going to read the file into private memory we
			 * allocate */
991
			if (!(capabilities & NOMMU_MAP_COPY))
L
Linus Torvalds 已提交
992 993 994 995 996
				return -ENODEV;

			/* we don't permit a private writable mapping to be
			 * shared with the backing device */
			if (prot & PROT_WRITE)
997
				capabilities &= ~NOMMU_MAP_DIRECT;
L
Linus Torvalds 已提交
998 999
		}

1000 1001 1002 1003
		if (capabilities & NOMMU_MAP_DIRECT) {
			if (((prot & PROT_READ)  && !(capabilities & NOMMU_MAP_READ))  ||
			    ((prot & PROT_WRITE) && !(capabilities & NOMMU_MAP_WRITE)) ||
			    ((prot & PROT_EXEC)  && !(capabilities & NOMMU_MAP_EXEC))
1004
			    ) {
1005
				capabilities &= ~NOMMU_MAP_DIRECT;
1006
				if (flags & MAP_SHARED) {
1007
					pr_warn("MAP_SHARED not completely supported on !MMU\n");
1008 1009 1010 1011 1012
					return -EINVAL;
				}
			}
		}

L
Linus Torvalds 已提交
1013 1014
		/* handle executable mappings and implied executable
		 * mappings */
1015
		if (path_noexec(&file->f_path)) {
L
Linus Torvalds 已提交
1016 1017
			if (prot & PROT_EXEC)
				return -EPERM;
1018
		} else if ((prot & PROT_READ) && !(prot & PROT_EXEC)) {
L
Linus Torvalds 已提交
1019 1020
			/* handle implication of PROT_EXEC by PROT_READ */
			if (current->personality & READ_IMPLIES_EXEC) {
1021
				if (capabilities & NOMMU_MAP_EXEC)
L
Linus Torvalds 已提交
1022 1023
					prot |= PROT_EXEC;
			}
1024
		} else if ((prot & PROT_READ) &&
L
Linus Torvalds 已提交
1025
			 (prot & PROT_EXEC) &&
1026
			 !(capabilities & NOMMU_MAP_EXEC)
L
Linus Torvalds 已提交
1027 1028
			 ) {
			/* backing file is not executable, try to copy */
1029
			capabilities &= ~NOMMU_MAP_DIRECT;
L
Linus Torvalds 已提交
1030
		}
1031
	} else {
L
Linus Torvalds 已提交
1032 1033 1034
		/* anonymous mappings are always memory backed and can be
		 * privately mapped
		 */
1035
		capabilities = NOMMU_MAP_COPY;
L
Linus Torvalds 已提交
1036 1037 1038 1039 1040 1041 1042 1043

		/* handle PROT_EXEC implication by PROT_READ */
		if ((prot & PROT_READ) &&
		    (current->personality & READ_IMPLIES_EXEC))
			prot |= PROT_EXEC;
	}

	/* allow the security API to have its say */
1044
	ret = security_mmap_addr(addr);
L
Linus Torvalds 已提交
1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
	if (ret < 0)
		return ret;

	/* looks okay */
	*_capabilities = capabilities;
	return 0;
}

/*
 * we've determined that we can make the mapping, now translate what we
 * now know into VMA flags
 */
static unsigned long determine_vm_flags(struct file *file,
					unsigned long prot,
					unsigned long flags,
					unsigned long capabilities)
{
	unsigned long vm_flags;

	vm_flags = calc_vm_prot_bits(prot) | calc_vm_flag_bits(flags);
	/* vm_flags |= mm->def_flags; */

1067
	if (!(capabilities & NOMMU_MAP_DIRECT)) {
L
Linus Torvalds 已提交
1068
		/* attempt to share read-only copies of mapped file chunks */
1069
		vm_flags |= VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC;
L
Linus Torvalds 已提交
1070 1071
		if (file && !(prot & PROT_WRITE))
			vm_flags |= VM_MAYSHARE;
1072
	} else {
L
Linus Torvalds 已提交
1073 1074 1075
		/* overlay a shareable mapping on the backing device or inode
		 * if possible - used for chardevs, ramfs/tmpfs/shmfs and
		 * romfs/cramfs */
1076
		vm_flags |= VM_MAYSHARE | (capabilities & NOMMU_VMFLAGS);
L
Linus Torvalds 已提交
1077
		if (flags & MAP_SHARED)
1078
			vm_flags |= VM_SHARED;
L
Linus Torvalds 已提交
1079 1080 1081 1082 1083 1084
	}

	/* refuse to let anyone share private mappings with this process if
	 * it's being traced - otherwise breakpoints set in it may interfere
	 * with another untraced process
	 */
T
Tejun Heo 已提交
1085
	if ((flags & MAP_PRIVATE) && current->ptrace)
L
Linus Torvalds 已提交
1086 1087 1088 1089 1090 1091
		vm_flags &= ~VM_MAYSHARE;

	return vm_flags;
}

/*
1092 1093
 * set up a shared mapping on a file (the driver or filesystem provides and
 * pins the storage)
L
Linus Torvalds 已提交
1094
 */
1095
static int do_mmap_shared_file(struct vm_area_struct *vma)
L
Linus Torvalds 已提交
1096 1097 1098 1099
{
	int ret;

	ret = vma->vm_file->f_op->mmap(vma->vm_file, vma);
1100 1101
	if (ret == 0) {
		vma->vm_region->vm_top = vma->vm_region->vm_end;
1102
		return 0;
1103
	}
L
Linus Torvalds 已提交
1104 1105 1106
	if (ret != -ENOSYS)
		return ret;

1107 1108 1109
	/* getting -ENOSYS indicates that direct mmap isn't possible (as
	 * opposed to tried but failed) so we can only give a suitable error as
	 * it's not possible to make a private copy if MAP_SHARED was given */
L
Linus Torvalds 已提交
1110 1111 1112 1113 1114 1115
	return -ENODEV;
}

/*
 * set up a private mapping or an anonymous shared mapping
 */
1116 1117
static int do_mmap_private(struct vm_area_struct *vma,
			   struct vm_region *region,
1118 1119
			   unsigned long len,
			   unsigned long capabilities)
L
Linus Torvalds 已提交
1120
{
1121
	unsigned long total, point;
L
Linus Torvalds 已提交
1122
	void *base;
1123
	int ret, order;
L
Linus Torvalds 已提交
1124 1125 1126 1127 1128

	/* invoke the file's mapping function so that it can keep track of
	 * shared mappings on devices or memory
	 * - VM_MAYSHARE will be set if it may attempt to share
	 */
1129
	if (capabilities & NOMMU_MAP_DIRECT) {
L
Linus Torvalds 已提交
1130
		ret = vma->vm_file->f_op->mmap(vma->vm_file, vma);
1131
		if (ret == 0) {
L
Linus Torvalds 已提交
1132
			/* shouldn't return success if we're not sharing */
1133 1134
			BUG_ON(!(vma->vm_flags & VM_MAYSHARE));
			vma->vm_region->vm_top = vma->vm_region->vm_end;
1135
			return 0;
L
Linus Torvalds 已提交
1136
		}
1137 1138
		if (ret != -ENOSYS)
			return ret;
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143 1144

		/* getting an ENOSYS error indicates that direct mmap isn't
		 * possible (as opposed to tried but failed) so we'll try to
		 * make a private copy of the data and map that instead */
	}

1145

L
Linus Torvalds 已提交
1146 1147 1148 1149
	/* allocate some memory to hold the mapping
	 * - note that this may not return a page-aligned address if the object
	 *   we're allocating is smaller than a page
	 */
B
Bob Liu 已提交
1150
	order = get_order(len);
1151
	total = 1 << order;
B
Bob Liu 已提交
1152
	point = len >> PAGE_SHIFT;
1153

1154
	/* we don't want to allocate a power-of-2 sized page set */
1155
	if (sysctl_nr_trim_pages && total - point >= sysctl_nr_trim_pages)
1156
		total = point;
1157

J
Joonsoo Kim 已提交
1158
	base = alloc_pages_exact(total << PAGE_SHIFT, GFP_KERNEL);
1159 1160 1161 1162
	if (!base)
		goto enomem;

	atomic_long_add(total, &mmap_pages_allocated);
L
Linus Torvalds 已提交
1163

1164 1165
	region->vm_flags = vma->vm_flags |= VM_MAPPED_COPY;
	region->vm_start = (unsigned long) base;
B
Bob Liu 已提交
1166
	region->vm_end   = region->vm_start + len;
1167
	region->vm_top   = region->vm_start + (total << PAGE_SHIFT);
1168 1169 1170

	vma->vm_start = region->vm_start;
	vma->vm_end   = region->vm_start + len;
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181

	if (vma->vm_file) {
		/* read the contents of a file into the copy */
		mm_segment_t old_fs;
		loff_t fpos;

		fpos = vma->vm_pgoff;
		fpos <<= PAGE_SHIFT;

		old_fs = get_fs();
		set_fs(KERNEL_DS);
A
Al Viro 已提交
1182
		ret = __vfs_read(vma->vm_file, base, len, &fpos);
L
Linus Torvalds 已提交
1183 1184 1185 1186 1187 1188
		set_fs(old_fs);

		if (ret < 0)
			goto error_free;

		/* clear the last little bit */
B
Bob Liu 已提交
1189 1190
		if (ret < len)
			memset(base + ret, 0, len - ret);
L
Linus Torvalds 已提交
1191 1192 1193 1194 1195 1196

	}

	return 0;

error_free:
1197
	free_page_series(region->vm_start, region->vm_top);
1198 1199
	region->vm_start = vma->vm_start = 0;
	region->vm_end   = vma->vm_end = 0;
1200
	region->vm_top   = 0;
L
Linus Torvalds 已提交
1201 1202 1203
	return ret;

enomem:
1204
	pr_err("Allocation of length %lu from process %d (%s) failed\n",
1205
	       len, current->pid, current->comm);
1206
	show_free_areas(0);
L
Linus Torvalds 已提交
1207 1208 1209 1210 1211 1212
	return -ENOMEM;
}

/*
 * handle mapping creation for uClinux
 */
1213 1214 1215 1216 1217 1218 1219 1220
unsigned long do_mmap(struct file *file,
			unsigned long addr,
			unsigned long len,
			unsigned long prot,
			unsigned long flags,
			vm_flags_t vm_flags,
			unsigned long pgoff,
			unsigned long *populate)
L
Linus Torvalds 已提交
1221
{
1222 1223
	struct vm_area_struct *vma;
	struct vm_region *region;
L
Linus Torvalds 已提交
1224
	struct rb_node *rb;
1225
	unsigned long capabilities, result;
L
Linus Torvalds 已提交
1226 1227
	int ret;

1228
	*populate = 0;
1229

L
Linus Torvalds 已提交
1230 1231 1232 1233
	/* decide whether we should attempt the mapping, and if so what sort of
	 * mapping */
	ret = validate_mmap_request(file, addr, len, prot, flags, pgoff,
				    &capabilities);
1234
	if (ret < 0)
L
Linus Torvalds 已提交
1235 1236
		return ret;

1237 1238
	/* we ignore the address hint */
	addr = 0;
B
Bob Liu 已提交
1239
	len = PAGE_ALIGN(len);
1240

L
Linus Torvalds 已提交
1241 1242
	/* we've determined that we can make the mapping, now translate what we
	 * now know into VMA flags */
1243
	vm_flags |= determine_vm_flags(file, prot, flags, capabilities);
L
Linus Torvalds 已提交
1244

1245 1246 1247 1248 1249 1250 1251 1252
	/* we're going to need to record the mapping */
	region = kmem_cache_zalloc(vm_region_jar, GFP_KERNEL);
	if (!region)
		goto error_getting_region;

	vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
	if (!vma)
		goto error_getting_vma;
L
Linus Torvalds 已提交
1253

1254
	region->vm_usage = 1;
1255 1256 1257
	region->vm_flags = vm_flags;
	region->vm_pgoff = pgoff;

1258
	INIT_LIST_HEAD(&vma->anon_vma_chain);
1259 1260
	vma->vm_flags = vm_flags;
	vma->vm_pgoff = pgoff;
L
Linus Torvalds 已提交
1261

1262
	if (file) {
A
Al Viro 已提交
1263 1264
		region->vm_file = get_file(file);
		vma->vm_file = get_file(file);
1265 1266 1267 1268 1269
	}

	down_write(&nommu_region_sem);

	/* if we want to share, we need to check for regions created by other
L
Linus Torvalds 已提交
1270
	 * mmap() calls that overlap with our proposed mapping
1271
	 * - we can only share with a superset match on most regular files
L
Linus Torvalds 已提交
1272 1273 1274 1275 1276 1277
	 * - shared mappings on character devices and memory backed files are
	 *   permitted to overlap inexactly as far as we are concerned for in
	 *   these cases, sharing is handled in the driver or filesystem rather
	 *   than here
	 */
	if (vm_flags & VM_MAYSHARE) {
1278 1279
		struct vm_region *pregion;
		unsigned long pglen, rpglen, pgend, rpgend, start;
L
Linus Torvalds 已提交
1280

1281 1282
		pglen = (len + PAGE_SIZE - 1) >> PAGE_SHIFT;
		pgend = pgoff + pglen;
1283

1284 1285
		for (rb = rb_first(&nommu_region_tree); rb; rb = rb_next(rb)) {
			pregion = rb_entry(rb, struct vm_region, vm_rb);
L
Linus Torvalds 已提交
1286

1287
			if (!(pregion->vm_flags & VM_MAYSHARE))
L
Linus Torvalds 已提交
1288 1289 1290
				continue;

			/* search for overlapping mappings on the same file */
A
Al Viro 已提交
1291 1292
			if (file_inode(pregion->vm_file) !=
			    file_inode(file))
L
Linus Torvalds 已提交
1293 1294
				continue;

1295
			if (pregion->vm_pgoff >= pgend)
L
Linus Torvalds 已提交
1296 1297
				continue;

1298 1299 1300 1301
			rpglen = pregion->vm_end - pregion->vm_start;
			rpglen = (rpglen + PAGE_SIZE - 1) >> PAGE_SHIFT;
			rpgend = pregion->vm_pgoff + rpglen;
			if (pgoff >= rpgend)
L
Linus Torvalds 已提交
1302 1303
				continue;

1304 1305 1306 1307 1308
			/* handle inexactly overlapping matches between
			 * mappings */
			if ((pregion->vm_pgoff != pgoff || rpglen != pglen) &&
			    !(pgoff >= pregion->vm_pgoff && pgend <= rpgend)) {
				/* new mapping is not a subset of the region */
1309
				if (!(capabilities & NOMMU_MAP_DIRECT))
L
Linus Torvalds 已提交
1310 1311 1312 1313
					goto sharing_violation;
				continue;
			}

1314
			/* we've found a region we can share */
1315
			pregion->vm_usage++;
1316 1317 1318 1319 1320 1321
			vma->vm_region = pregion;
			start = pregion->vm_start;
			start += (pgoff - pregion->vm_pgoff) << PAGE_SHIFT;
			vma->vm_start = start;
			vma->vm_end = start + len;

1322
			if (pregion->vm_flags & VM_MAPPED_COPY)
1323
				vma->vm_flags |= VM_MAPPED_COPY;
1324
			else {
1325 1326 1327 1328 1329
				ret = do_mmap_shared_file(vma);
				if (ret < 0) {
					vma->vm_region = NULL;
					vma->vm_start = 0;
					vma->vm_end = 0;
1330
					pregion->vm_usage--;
1331 1332 1333 1334 1335 1336 1337 1338 1339
					pregion = NULL;
					goto error_just_free;
				}
			}
			fput(region->vm_file);
			kmem_cache_free(vm_region_jar, region);
			region = pregion;
			result = start;
			goto share;
L
Linus Torvalds 已提交
1340 1341 1342 1343 1344 1345
		}

		/* obtain the address at which to make a shared mapping
		 * - this is the hook for quasi-memory character devices to
		 *   tell us the location of a shared mapping
		 */
1346
		if (capabilities & NOMMU_MAP_DIRECT) {
L
Linus Torvalds 已提交
1347 1348
			addr = file->f_op->get_unmapped_area(file, addr, len,
							     pgoff, flags);
1349
			if (IS_ERR_VALUE(addr)) {
L
Linus Torvalds 已提交
1350
				ret = addr;
1351
				if (ret != -ENOSYS)
1352
					goto error_just_free;
L
Linus Torvalds 已提交
1353 1354 1355 1356

				/* the driver refused to tell us where to site
				 * the mapping so we'll have to attempt to copy
				 * it */
1357
				ret = -ENODEV;
1358
				if (!(capabilities & NOMMU_MAP_COPY))
1359
					goto error_just_free;
L
Linus Torvalds 已提交
1360

1361
				capabilities &= ~NOMMU_MAP_DIRECT;
1362 1363 1364
			} else {
				vma->vm_start = region->vm_start = addr;
				vma->vm_end = region->vm_end = addr + len;
L
Linus Torvalds 已提交
1365 1366 1367 1368
			}
		}
	}

1369
	vma->vm_region = region;
L
Linus Torvalds 已提交
1370

1371
	/* set up the mapping
1372
	 * - the region is filled in if NOMMU_MAP_DIRECT is still set
1373
	 */
L
Linus Torvalds 已提交
1374
	if (file && vma->vm_flags & VM_SHARED)
1375
		ret = do_mmap_shared_file(vma);
L
Linus Torvalds 已提交
1376
	else
1377
		ret = do_mmap_private(vma, region, len, capabilities);
L
Linus Torvalds 已提交
1378
	if (ret < 0)
1379 1380
		goto error_just_free;
	add_nommu_region(region);
1381

1382 1383 1384 1385 1386
	/* clear anonymous mappings that don't ask for uninitialized data */
	if (!vma->vm_file && !(flags & MAP_UNINITIALIZED))
		memset((void *)region->vm_start, 0,
		       region->vm_end - region->vm_start);

L
Linus Torvalds 已提交
1387
	/* okay... we have a mapping; now we have to register it */
1388
	result = vma->vm_start;
L
Linus Torvalds 已提交
1389 1390 1391

	current->mm->total_vm += len >> PAGE_SHIFT;

1392 1393
share:
	add_vma_to_mm(current->mm, vma);
L
Linus Torvalds 已提交
1394

1395 1396 1397 1398 1399 1400
	/* we flush the region from the icache only when the first executable
	 * mapping of it is made  */
	if (vma->vm_flags & VM_EXEC && !region->vm_icache_flushed) {
		flush_icache_range(region->vm_start, region->vm_end);
		region->vm_icache_flushed = true;
	}
L
Linus Torvalds 已提交
1401

1402
	up_write(&nommu_region_sem);
L
Linus Torvalds 已提交
1403

1404
	return result;
L
Linus Torvalds 已提交
1405

1406 1407 1408
error_just_free:
	up_write(&nommu_region_sem);
error:
1409 1410
	if (region->vm_file)
		fput(region->vm_file);
1411
	kmem_cache_free(vm_region_jar, region);
1412 1413
	if (vma->vm_file)
		fput(vma->vm_file);
1414 1415 1416 1417 1418
	kmem_cache_free(vm_area_cachep, vma);
	return ret;

sharing_violation:
	up_write(&nommu_region_sem);
1419
	pr_warn("Attempt to share mismatched mappings\n");
1420 1421
	ret = -EINVAL;
	goto error;
L
Linus Torvalds 已提交
1422

1423 1424
error_getting_vma:
	kmem_cache_free(vm_region_jar, region);
1425 1426
	pr_warn("Allocation of vma for %lu byte allocation from process %d failed\n",
			len, current->pid);
1427
	show_free_areas(0);
L
Linus Torvalds 已提交
1428 1429
	return -ENOMEM;

1430
error_getting_region:
1431 1432
	pr_warn("Allocation of vm region for %lu byte allocation from process %d failed\n",
			len, current->pid);
1433
	show_free_areas(0);
L
Linus Torvalds 已提交
1434 1435
	return -ENOMEM;
}
1436

H
Hugh Dickins 已提交
1437 1438 1439 1440 1441 1442 1443
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;

A
Al Viro 已提交
1444
	audit_mmap_fd(fd, flags);
H
Hugh Dickins 已提交
1445 1446 1447 1448 1449 1450 1451 1452
	if (!(flags & MAP_ANONYMOUS)) {
		file = fget(fd);
		if (!file)
			goto out;
	}

	flags &= ~(MAP_EXECUTABLE | MAP_DENYWRITE);

G
Greg Ungerer 已提交
1453
	retval = vm_mmap_pgoff(file, addr, len, prot, flags, pgoff);
H
Hugh Dickins 已提交
1454 1455 1456 1457 1458 1459 1460

	if (file)
		fput(file);
out:
	return retval;
}

C
Christoph Hellwig 已提交
1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476
#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;
1477
	if (offset_in_page(a.offset))
C
Christoph Hellwig 已提交
1478 1479 1480 1481 1482 1483 1484
		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 */

L
Linus Torvalds 已提交
1485
/*
1486 1487
 * split a vma into two pieces at address 'addr', a new vma is allocated either
 * for the first part or the tail.
L
Linus Torvalds 已提交
1488
 */
1489 1490
int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
	      unsigned long addr, int new_below)
L
Linus Torvalds 已提交
1491
{
1492 1493 1494
	struct vm_area_struct *new;
	struct vm_region *region;
	unsigned long npages;
L
Linus Torvalds 已提交
1495

1496 1497 1498
	/* we're only permitted to split anonymous regions (these should have
	 * only a single usage on the region) */
	if (vma->vm_file)
1499
		return -ENOMEM;
L
Linus Torvalds 已提交
1500

1501 1502
	if (mm->map_count >= sysctl_max_map_count)
		return -ENOMEM;
L
Linus Torvalds 已提交
1503

1504 1505 1506
	region = kmem_cache_alloc(vm_region_jar, GFP_KERNEL);
	if (!region)
		return -ENOMEM;
L
Linus Torvalds 已提交
1507

1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
	new = kmem_cache_alloc(vm_area_cachep, GFP_KERNEL);
	if (!new) {
		kmem_cache_free(vm_region_jar, region);
		return -ENOMEM;
	}

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

	npages = (addr - vma->vm_start) >> PAGE_SHIFT;

	if (new_below) {
1522
		region->vm_top = region->vm_end = new->vm_end = addr;
1523 1524 1525
	} else {
		region->vm_start = new->vm_start = addr;
		region->vm_pgoff = new->vm_pgoff += npages;
L
Linus Torvalds 已提交
1526
	}
1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538

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

	delete_vma_from_mm(vma);
	down_write(&nommu_region_sem);
	delete_nommu_region(vma->vm_region);
	if (new_below) {
		vma->vm_region->vm_start = vma->vm_start = addr;
		vma->vm_region->vm_pgoff = vma->vm_pgoff += npages;
	} else {
		vma->vm_region->vm_end = vma->vm_end = addr;
1539
		vma->vm_region->vm_top = addr;
1540 1541 1542 1543 1544 1545 1546
	}
	add_nommu_region(vma->vm_region);
	add_nommu_region(new->vm_region);
	up_write(&nommu_region_sem);
	add_vma_to_mm(mm, vma);
	add_vma_to_mm(mm, new);
	return 0;
L
Linus Torvalds 已提交
1547 1548
}

1549
/*
1550 1551
 * shrink a VMA by removing the specified chunk from either the beginning or
 * the end
1552
 */
1553 1554 1555
static int shrink_vma(struct mm_struct *mm,
		      struct vm_area_struct *vma,
		      unsigned long from, unsigned long to)
L
Linus Torvalds 已提交
1556
{
1557
	struct vm_region *region;
L
Linus Torvalds 已提交
1558

1559 1560 1561 1562 1563 1564 1565 1566
	/* adjust the VMA's pointers, which may reposition it in the MM's tree
	 * and list */
	delete_vma_from_mm(vma);
	if (from > vma->vm_start)
		vma->vm_end = from;
	else
		vma->vm_start = to;
	add_vma_to_mm(mm, vma);
L
Linus Torvalds 已提交
1567

1568 1569
	/* cut the backing region down to size */
	region = vma->vm_region;
1570
	BUG_ON(region->vm_usage != 1);
1571 1572 1573

	down_write(&nommu_region_sem);
	delete_nommu_region(region);
1574 1575 1576 1577
	if (from > region->vm_start) {
		to = region->vm_top;
		region->vm_top = region->vm_end = from;
	} else {
1578
		region->vm_start = to;
1579
	}
1580 1581 1582 1583 1584 1585
	add_nommu_region(region);
	up_write(&nommu_region_sem);

	free_page_series(from, to);
	return 0;
}
L
Linus Torvalds 已提交
1586

1587 1588 1589 1590 1591 1592 1593 1594
/*
 * release a mapping
 * - under NOMMU conditions the chunk to be unmapped must be backed by a single
 *   VMA, though it need not cover the whole VMA
 */
int do_munmap(struct mm_struct *mm, unsigned long start, size_t len)
{
	struct vm_area_struct *vma;
B
Bob Liu 已提交
1595
	unsigned long end;
1596
	int ret;
L
Linus Torvalds 已提交
1597

B
Bob Liu 已提交
1598
	len = PAGE_ALIGN(len);
1599 1600
	if (len == 0)
		return -EINVAL;
1601

B
Bob Liu 已提交
1602 1603
	end = start + len;

1604 1605 1606
	/* find the first potentially overlapping VMA */
	vma = find_vma(mm, start);
	if (!vma) {
1607
		static int limit;
1608
		if (limit < 5) {
1609 1610 1611
			pr_warn("munmap of memory not mmapped by process %d (%s): 0x%lx-0x%lx\n",
					current->pid, current->comm,
					start, start + len - 1);
1612 1613
			limit++;
		}
1614 1615
		return -EINVAL;
	}
L
Linus Torvalds 已提交
1616

1617 1618 1619
	/* we're allowed to split an anonymous VMA but not a file-backed one */
	if (vma->vm_file) {
		do {
1620
			if (start > vma->vm_start)
1621 1622 1623
				return -EINVAL;
			if (end == vma->vm_end)
				goto erase_whole_vma;
1624 1625
			vma = vma->vm_next;
		} while (vma);
1626 1627 1628 1629 1630
		return -EINVAL;
	} else {
		/* the chunk must be a subset of the VMA found */
		if (start == vma->vm_start && end == vma->vm_end)
			goto erase_whole_vma;
1631
		if (start < vma->vm_start || end > vma->vm_end)
1632
			return -EINVAL;
1633
		if (offset_in_page(start))
1634
			return -EINVAL;
1635
		if (end != vma->vm_end && offset_in_page(end))
1636 1637 1638
			return -EINVAL;
		if (start != vma->vm_start && end != vma->vm_end) {
			ret = split_vma(mm, vma, start, 1);
1639
			if (ret < 0)
1640 1641 1642 1643
				return ret;
		}
		return shrink_vma(mm, vma, start, end);
	}
L
Linus Torvalds 已提交
1644

1645 1646 1647
erase_whole_vma:
	delete_vma_from_mm(vma);
	delete_vma(mm, vma);
L
Linus Torvalds 已提交
1648 1649
	return 0;
}
1650
EXPORT_SYMBOL(do_munmap);
L
Linus Torvalds 已提交
1651

A
Al Viro 已提交
1652
int vm_munmap(unsigned long addr, size_t len)
1653
{
A
Al Viro 已提交
1654
	struct mm_struct *mm = current->mm;
1655 1656 1657 1658 1659 1660 1661
	int ret;

	down_write(&mm->mmap_sem);
	ret = do_munmap(mm, addr, len);
	up_write(&mm->mmap_sem);
	return ret;
}
1662 1663 1664 1665
EXPORT_SYMBOL(vm_munmap);

SYSCALL_DEFINE2(munmap, unsigned long, addr, size_t, len)
{
A
Al Viro 已提交
1666
	return vm_munmap(addr, len);
1667
}
1668 1669

/*
1670
 * release all the mappings made in a process's VM space
1671
 */
1672
void exit_mmap(struct mm_struct *mm)
L
Linus Torvalds 已提交
1673
{
1674
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
1675

1676 1677
	if (!mm)
		return;
L
Linus Torvalds 已提交
1678

1679
	mm->total_vm = 0;
L
Linus Torvalds 已提交
1680

1681 1682 1683 1684
	while ((vma = mm->mmap)) {
		mm->mmap = vma->vm_next;
		delete_vma_from_mm(vma);
		delete_vma(mm, vma);
1685
		cond_resched();
L
Linus Torvalds 已提交
1686 1687 1688
	}
}

1689
unsigned long vm_brk(unsigned long addr, unsigned long len)
L
Linus Torvalds 已提交
1690 1691 1692 1693 1694
{
	return -ENOMEM;
}

/*
1695 1696
 * expand (or shrink) an existing mapping, potentially moving it at the same
 * time (controlled by the MREMAP_MAYMOVE flag and available VM space)
L
Linus Torvalds 已提交
1697
 *
1698
 * under NOMMU conditions, we only permit changing a mapping's size, and only
1699 1700
 * as long as it stays within the region allocated by do_mmap_private() and the
 * block is not shareable
L
Linus Torvalds 已提交
1701
 *
1702
 * MREMAP_FIXED is not supported under NOMMU conditions
L
Linus Torvalds 已提交
1703
 */
A
Al Viro 已提交
1704
static unsigned long do_mremap(unsigned long addr,
L
Linus Torvalds 已提交
1705 1706 1707
			unsigned long old_len, unsigned long new_len,
			unsigned long flags, unsigned long new_addr)
{
1708
	struct vm_area_struct *vma;
L
Linus Torvalds 已提交
1709 1710

	/* insanity checks first */
B
Bob Liu 已提交
1711 1712
	old_len = PAGE_ALIGN(old_len);
	new_len = PAGE_ALIGN(new_len);
1713
	if (old_len == 0 || new_len == 0)
L
Linus Torvalds 已提交
1714 1715
		return (unsigned long) -EINVAL;

1716
	if (offset_in_page(addr))
1717 1718
		return -EINVAL;

L
Linus Torvalds 已提交
1719 1720 1721
	if (flags & MREMAP_FIXED && new_addr != addr)
		return (unsigned long) -EINVAL;

1722
	vma = find_vma_exact(current->mm, addr, old_len);
1723 1724
	if (!vma)
		return (unsigned long) -EINVAL;
L
Linus Torvalds 已提交
1725

1726
	if (vma->vm_end != vma->vm_start + old_len)
L
Linus Torvalds 已提交
1727 1728
		return (unsigned long) -EFAULT;

1729
	if (vma->vm_flags & VM_MAYSHARE)
L
Linus Torvalds 已提交
1730 1731
		return (unsigned long) -EPERM;

1732
	if (new_len > vma->vm_region->vm_end - vma->vm_region->vm_start)
L
Linus Torvalds 已提交
1733 1734 1735
		return (unsigned long) -ENOMEM;

	/* all checks complete - do it */
1736 1737 1738 1739
	vma->vm_end = vma->vm_start + new_len;
	return vma->vm_start;
}

1740 1741 1742
SYSCALL_DEFINE5(mremap, unsigned long, addr, unsigned long, old_len,
		unsigned long, new_len, unsigned long, flags,
		unsigned long, new_addr)
1743 1744 1745 1746 1747 1748 1749
{
	unsigned long ret;

	down_write(&current->mm->mmap_sem);
	ret = do_mremap(addr, old_len, new_len, flags, new_addr);
	up_write(&current->mm->mmap_sem);
	return ret;
L
Linus Torvalds 已提交
1750 1751
}

1752 1753 1754
struct page *follow_page_mask(struct vm_area_struct *vma,
			      unsigned long address, unsigned int flags,
			      unsigned int *page_mask)
L
Linus Torvalds 已提交
1755
{
1756
	*page_mask = 0;
L
Linus Torvalds 已提交
1757 1758 1759
	return NULL;
}

B
Bob Liu 已提交
1760 1761
int remap_pfn_range(struct vm_area_struct *vma, unsigned long addr,
		unsigned long pfn, unsigned long size, pgprot_t prot)
L
Linus Torvalds 已提交
1762
{
B
Bob Liu 已提交
1763 1764 1765
	if (addr != (pfn << PAGE_SHIFT))
		return -EINVAL;

1766
	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
1767
	return 0;
L
Linus Torvalds 已提交
1768
}
1769
EXPORT_SYMBOL(remap_pfn_range);
L
Linus Torvalds 已提交
1770

1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
int vm_iomap_memory(struct vm_area_struct *vma, phys_addr_t start, unsigned long len)
{
	unsigned long pfn = start >> PAGE_SHIFT;
	unsigned long vm_len = vma->vm_end - vma->vm_start;

	pfn += vma->vm_pgoff;
	return io_remap_pfn_range(vma, vma->vm_start, pfn, vm_len, vma->vm_page_prot);
}
EXPORT_SYMBOL(vm_iomap_memory);

1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795
int remap_vmalloc_range(struct vm_area_struct *vma, void *addr,
			unsigned long pgoff)
{
	unsigned int size = vma->vm_end - vma->vm_start;

	if (!(vma->vm_flags & VM_USERMAP))
		return -EINVAL;

	vma->vm_start = (unsigned long)(addr + (pgoff << PAGE_SHIFT));
	vma->vm_end = vma->vm_start + size;

	return 0;
}
EXPORT_SYMBOL(remap_vmalloc_range);

L
Linus Torvalds 已提交
1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806
unsigned long arch_get_unmapped_area(struct file *file, unsigned long addr,
	unsigned long len, unsigned long pgoff, unsigned long flags)
{
	return -ENOMEM;
}

void unmap_mapping_range(struct address_space *mapping,
			 loff_t const holebegin, loff_t const holelen,
			 int even_cows)
{
}
1807
EXPORT_SYMBOL(unmap_mapping_range);
L
Linus Torvalds 已提交
1808

N
Nick Piggin 已提交
1809
int filemap_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
1810 1811
{
	BUG();
N
Nick Piggin 已提交
1812
	return 0;
1813
}
1814
EXPORT_SYMBOL(filemap_fault);
1815

1816 1817 1818 1819 1820 1821
void filemap_map_pages(struct vm_area_struct *vma, struct vm_fault *vmf)
{
	BUG();
}
EXPORT_SYMBOL(filemap_map_pages);

1822 1823
static int __access_remote_vm(struct task_struct *tsk, struct mm_struct *mm,
		unsigned long addr, void *buf, int len, int write)
1824 1825 1826 1827 1828 1829
{
	struct vm_area_struct *vma;

	down_read(&mm->mmap_sem);

	/* the access must start within one of the target process's mappings */
1830 1831
	vma = find_vma(mm, addr);
	if (vma) {
1832 1833 1834 1835 1836
		/* don't overrun this mapping */
		if (addr + len >= vma->vm_end)
			len = vma->vm_end - addr;

		/* only read or write mappings where it is permitted */
1837
		if (write && vma->vm_flags & VM_MAYWRITE)
1838 1839
			copy_to_user_page(vma, NULL, addr,
					 (void *) addr, buf, len);
1840
		else if (!write && vma->vm_flags & VM_MAYREAD)
1841 1842
			copy_from_user_page(vma, NULL, addr,
					    buf, (void *) addr, len);
1843 1844 1845 1846 1847 1848 1849
		else
			len = 0;
	} else {
		len = 0;
	}

	up_read(&mm->mmap_sem);
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 1883 1884 1885 1886

	return len;
}

/**
 * @access_remote_vm - access another process' address space
 * @mm:		the mm_struct of the target address space
 * @addr:	start address to access
 * @buf:	source or destination buffer
 * @len:	number of bytes to transfer
 * @write:	whether the access is a write
 *
 * The caller must hold a reference on @mm.
 */
int access_remote_vm(struct mm_struct *mm, unsigned long addr,
		void *buf, int len, int write)
{
	return __access_remote_vm(NULL, mm, addr, buf, len, write);
}

/*
 * Access another process' address space.
 * - source/target buffer must be kernel space
 */
int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, int len, int write)
{
	struct mm_struct *mm;

	if (addr + len < addr)
		return 0;

	mm = get_task_mm(tsk);
	if (!mm)
		return 0;

	len = __access_remote_vm(tsk, mm, addr, buf, len, write);

1887 1888 1889
	mmput(mm);
	return len;
}
1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913

/**
 * nommu_shrink_inode_mappings - Shrink the shared mappings on an inode
 * @inode: The inode to check
 * @size: The current filesize of the inode
 * @newsize: The proposed filesize of the inode
 *
 * Check the shared mappings on an inode on behalf of a shrinking truncate to
 * make sure that that any outstanding VMAs aren't broken and then shrink the
 * vm_regions that extend that beyond so that do_mmap_pgoff() doesn't
 * automatically grant mappings that are too large.
 */
int nommu_shrink_inode_mappings(struct inode *inode, size_t size,
				size_t newsize)
{
	struct vm_area_struct *vma;
	struct vm_region *region;
	pgoff_t low, high;
	size_t r_size, r_top;

	low = newsize >> PAGE_SHIFT;
	high = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;

	down_write(&nommu_region_sem);
D
Davidlohr Bueso 已提交
1914
	i_mmap_lock_read(inode->i_mapping);
1915 1916

	/* search for VMAs that fall within the dead zone */
1917
	vma_interval_tree_foreach(vma, &inode->i_mapping->i_mmap, low, high) {
1918 1919 1920
		/* found one - only interested if it's shared out of the page
		 * cache */
		if (vma->vm_flags & VM_SHARED) {
D
Davidlohr Bueso 已提交
1921
			i_mmap_unlock_read(inode->i_mapping);
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932
			up_write(&nommu_region_sem);
			return -ETXTBSY; /* not quite true, but near enough */
		}
	}

	/* reduce any regions that overlap the dead zone - if in existence,
	 * these will be pointed to by VMAs that don't overlap the dead zone
	 *
	 * we don't check for any regions that start beyond the EOF as there
	 * shouldn't be any
	 */
D
Davidlohr Bueso 已提交
1933
	vma_interval_tree_foreach(vma, &inode->i_mapping->i_mmap, 0, ULONG_MAX) {
1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947
		if (!(vma->vm_flags & VM_SHARED))
			continue;

		region = vma->vm_region;
		r_size = region->vm_top - region->vm_start;
		r_top = (region->vm_pgoff << PAGE_SHIFT) + r_size;

		if (r_top > newsize) {
			region->vm_top -= r_top - newsize;
			if (region->vm_end > region->vm_top)
				region->vm_end = region->vm_top;
		}
	}

D
Davidlohr Bueso 已提交
1948
	i_mmap_unlock_read(inode->i_mapping);
1949 1950 1951
	up_write(&nommu_region_sem);
	return 0;
}
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971

/*
 * 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.
 */
static int __meminit init_user_reserve(void)
{
	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;
}
1972
subsys_initcall(init_user_reserve);
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992

/*
 * 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.
 */
static int __meminit init_admin_reserve(void)
{
	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;
}
1993
subsys_initcall(init_admin_reserve);