nommu.c 45.4 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-only
Linus Torvalds's avatar
Linus Torvalds committed
2 3 4 5 6 7 8 9
/*
 *  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
 *
10
 *  Copyright (c) 2004-2008 David Howells <dhowells@redhat.com>
Linus Torvalds's avatar
Linus Torvalds committed
11 12 13
 *  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>
14
 *  Copyright (c) 2007-2010 Paul Mundt <lethal@linux-sh.org>
Linus Torvalds's avatar
Linus Torvalds committed
15 16
 */

17 18
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

19
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
20
#include <linux/mm.h>
21
#include <linux/sched/mm.h>
Davidlohr Bueso's avatar
Davidlohr Bueso committed
22
#include <linux/vmacache.h>
Linus Torvalds's avatar
Linus Torvalds committed
23 24 25 26 27 28 29 30 31
#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>
32
#include <linux/compiler.h>
Linus Torvalds's avatar
Linus Torvalds committed
33 34 35 36
#include <linux/mount.h>
#include <linux/personality.h>
#include <linux/security.h>
#include <linux/syscalls.h>
Al Viro's avatar
Al Viro committed
37
#include <linux/audit.h>
38
#include <linux/printk.h>
Linus Torvalds's avatar
Linus Torvalds committed
39

40
#include <linux/uaccess.h>
Linus Torvalds's avatar
Linus Torvalds committed
41 42
#include <asm/tlb.h>
#include <asm/tlbflush.h>
43
#include <asm/mmu_context.h>
44 45
#include "internal.h"

Linus Torvalds's avatar
Linus Torvalds committed
46
void *high_memory;
47
EXPORT_SYMBOL(high_memory);
Linus Torvalds's avatar
Linus Torvalds committed
48 49
struct page *mem_map;
unsigned long max_mapnr;
50
EXPORT_SYMBOL(max_mapnr);
Hugh Dickins's avatar
Hugh Dickins committed
51
unsigned long highest_memmap_pfn;
52
int sysctl_nr_trim_pages = CONFIG_NOMMU_INITIAL_TRIM_EXCESS;
Linus Torvalds's avatar
Linus Torvalds committed
53 54
int heap_stack_gap = 0;

55
atomic_long_t mmap_pages_allocated;
56

Linus Torvalds's avatar
Linus Torvalds committed
57 58
EXPORT_SYMBOL(mem_map);

59 60 61 62
/* 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);
Linus Torvalds's avatar
Linus Torvalds committed
63

64
const struct vm_operations_struct generic_file_vm_ops = {
Linus Torvalds's avatar
Linus Torvalds committed
65 66 67 68 69 70 71 72 73 74 75 76
};

/*
 * 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;

77 78 79 80
	/*
	 * If the object we have should not have ksize performed on it,
	 * return size of 0
	 */
81
	if (!objp || !virt_addr_valid(objp))
82 83 84 85 86 87 88 89
		return 0;

	page = virt_to_head_page(objp);

	/*
	 * If the allocator sets PageSlab, we know the pointer came from
	 * kmalloc().
	 */
Linus Torvalds's avatar
Linus Torvalds committed
90 91 92
	if (PageSlab(page))
		return ksize(objp);

93 94 95 96 97 98 99 100 101 102 103 104 105 106
	/*
	 * 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;
	}

107 108
	/*
	 * The ksize() function is only guaranteed to work for pointers
109
	 * returned by kmalloc(). So handle arbitrary pointers here.
110
	 */
111
	return PAGE_SIZE << compound_order(page);
Linus Torvalds's avatar
Linus Torvalds committed
112 113
}

Paul Mundt's avatar
Paul Mundt committed
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134
/**
 * 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);

135
LIST_HEAD(vmap_area_list);
Linus Torvalds's avatar
Linus Torvalds committed
136

137
void vfree(const void *addr)
Linus Torvalds's avatar
Linus Torvalds committed
138 139 140
{
	kfree(addr);
}
141
EXPORT_SYMBOL(vfree);
Linus Torvalds's avatar
Linus Torvalds committed
142

143
void *__vmalloc(unsigned long size, gfp_t gfp_mask, pgprot_t prot)
Linus Torvalds's avatar
Linus Torvalds committed
144 145
{
	/*
146 147
	 *  You can't specify __GFP_HIGHMEM with kmalloc() since kmalloc()
	 * returns only a logical address.
Linus Torvalds's avatar
Linus Torvalds committed
148
	 */
149
	return kmalloc(size, (gfp_mask | __GFP_COMP) & ~__GFP_HIGHMEM);
Linus Torvalds's avatar
Linus Torvalds committed
150
}
151
EXPORT_SYMBOL(__vmalloc);
Linus Torvalds's avatar
Linus Torvalds committed
152

153 154 155 156 157
void *__vmalloc_node_flags(unsigned long size, int node, gfp_t flags)
{
	return __vmalloc(size, flags, PAGE_KERNEL);
}

158 159 160 161
void *vmalloc_user(unsigned long size)
{
	void *ret;

162
	ret = __vmalloc(size, GFP_KERNEL | __GFP_ZERO, PAGE_KERNEL);
163 164 165 166 167 168 169 170 171 172 173 174 175 176
	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);

177
struct page *vmalloc_to_page(const void *addr)
Linus Torvalds's avatar
Linus Torvalds committed
178 179 180
{
	return virt_to_page(addr);
}
181
EXPORT_SYMBOL(vmalloc_to_page);
Linus Torvalds's avatar
Linus Torvalds committed
182

183
unsigned long vmalloc_to_pfn(const void *addr)
Linus Torvalds's avatar
Linus Torvalds committed
184 185 186
{
	return page_to_pfn(virt_to_page(addr));
}
187
EXPORT_SYMBOL(vmalloc_to_pfn);
Linus Torvalds's avatar
Linus Torvalds committed
188 189 190

long vread(char *buf, char *addr, unsigned long count)
{
191 192 193 194
	/* Don't allow overflow */
	if ((unsigned long) buf + count < count)
		count = -(unsigned long) buf;

Linus Torvalds's avatar
Linus Torvalds committed
195 196 197 198 199 200 201 202 203 204 205
	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);
206
	return count;
Linus Torvalds's avatar
Linus Torvalds committed
207 208 209
}

/*
210
 *	vmalloc  -  allocate virtually contiguous memory
Linus Torvalds's avatar
Linus Torvalds committed
211 212 213 214
 *
 *	@size:		allocation size
 *
 *	Allocate enough pages to cover @size from the page level
215
 *	allocator and map them into contiguous kernel virtual space.
Linus Torvalds's avatar
Linus Torvalds committed
216
 *
217
 *	For tight control over page level allocator and protection flags
Linus Torvalds's avatar
Linus Torvalds committed
218 219 220 221 222 223
 *	use __vmalloc() instead.
 */
void *vmalloc(unsigned long size)
{
       return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
}
224 225
EXPORT_SYMBOL(vmalloc);

226
/*
227
 *	vzalloc - allocate virtually contiguous memory with zero fill
228 229 230 231
 *
 *	@size:		allocation size
 *
 *	Allocate enough pages to cover @size from the page level
232
 *	allocator and map them into contiguous kernel virtual space.
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
 *	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.
 */
256 257 258 259
void *vmalloc_node(unsigned long size, int node)
{
	return vmalloc(size);
}
260
EXPORT_SYMBOL(vmalloc_node);
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278

/**
 * 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);
Linus Torvalds's avatar
Linus Torvalds committed
279

Paul Mundt's avatar
Paul Mundt committed
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
/**
 *	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);
}

297 298
/**
 * vmalloc_32  -  allocate virtually contiguous memory (32bit addressable)
Linus Torvalds's avatar
Linus Torvalds committed
299 300 301
 *	@size:		allocation size
 *
 *	Allocate enough 32bit PA addressable pages to cover @size from the
302
 *	page level allocator and map them into contiguous kernel virtual space.
Linus Torvalds's avatar
Linus Torvalds committed
303 304 305 306 307
 */
void *vmalloc_32(unsigned long size)
{
	return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL);
}
308 309 310 311 312 313 314 315
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.
316 317 318
 *
 * VM_USERMAP is set on the corresponding VMA so that subsequent calls to
 * remap_vmalloc_range() are permissible.
319 320 321
 */
void *vmalloc_32_user(unsigned long size)
{
322 323 324 325 326
	/*
	 * 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);
327 328
}
EXPORT_SYMBOL(vmalloc_32_user);
Linus Torvalds's avatar
Linus Torvalds committed
329 330 331 332 333 334

void *vmap(struct page **pages, unsigned int count, unsigned long flags, pgprot_t prot)
{
	BUG();
	return NULL;
}
335
EXPORT_SYMBOL(vmap);
Linus Torvalds's avatar
Linus Torvalds committed
336

337
void vunmap(const void *addr)
Linus Torvalds's avatar
Linus Torvalds committed
338 339 340
{
	BUG();
}
341
EXPORT_SYMBOL(vunmap);
Linus Torvalds's avatar
Linus Torvalds committed
342

343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
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);

361 362 363 364
/*
 * Implement a stub for vmalloc_sync_all() if the architecture chose not to
 * have one.
 */
365
void __weak vmalloc_sync_all(void)
366 367 368
{
}

369
struct vm_struct *alloc_vm_area(size_t size, pte_t **ptes)
370 371 372 373 374 375 376 377 378 379 380 381
{
	BUG();
	return NULL;
}
EXPORT_SYMBOL_GPL(alloc_vm_area);

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

382 383 384 385 386 387 388
int vm_insert_page(struct vm_area_struct *vma, unsigned long addr,
		   struct page *page)
{
	return -EINVAL;
}
EXPORT_SYMBOL(vm_insert_page);

389 390 391 392 393 394 395 396 397 398 399 400 401 402
int vm_map_pages(struct vm_area_struct *vma, struct page **pages,
			unsigned long num)
{
	return -EINVAL;
}
EXPORT_SYMBOL(vm_map_pages);

int vm_map_pages_zero(struct vm_area_struct *vma, struct page **pages,
				unsigned long num)
{
	return -EINVAL;
}
EXPORT_SYMBOL(vm_map_pages_zero);

Linus Torvalds's avatar
Linus Torvalds committed
403 404 405 406 407 408 409
/*
 *  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.
 */
410
SYSCALL_DEFINE1(brk, unsigned long, brk)
Linus Torvalds's avatar
Linus Torvalds committed
411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430
{
	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.
	 */
431
	flush_icache_range(mm->brk, brk);
Linus Torvalds's avatar
Linus Torvalds committed
432 433 434
	return mm->brk = brk;
}

435
/*
436
 * initialise the percpu counter for VM and region record slabs
437 438
 */
void __init mmap_init(void)
Linus Torvalds's avatar
Linus Torvalds committed
439
{
440 441
	int ret;

442
	ret = percpu_counter_init(&vm_committed_as, 0, GFP_KERNEL);
443
	VM_BUG_ON(ret);
444
	vm_region_jar = KMEM_CACHE(vm_region, SLAB_PANIC|SLAB_ACCOUNT);
Linus Torvalds's avatar
Linus Torvalds committed
445 446
}

447
/*
448 449
 * validate the region tree
 * - the caller must hold the region lock
450
 */
451 452
#ifdef CONFIG_DEBUG_NOMMU_REGIONS
static noinline void validate_nommu_regions(void)
453
{
454 455
	struct vm_region *region, *last;
	struct rb_node *p, *lastp;
456

457 458 459 460 461
	lastp = rb_first(&nommu_region_tree);
	if (!lastp)
		return;

	last = rb_entry(lastp, struct vm_region, vm_rb);
462 463
	BUG_ON(last->vm_end <= last->vm_start);
	BUG_ON(last->vm_top < last->vm_end);
464 465 466 467 468

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

469 470 471
		BUG_ON(region->vm_end <= region->vm_start);
		BUG_ON(region->vm_top < region->vm_end);
		BUG_ON(region->vm_start < last->vm_top);
472

473 474
		lastp = p;
	}
475
}
476
#else
477 478 479
static void validate_nommu_regions(void)
{
}
480
#endif
481 482

/*
483
 * add a region into the global tree
484
 */
485
static void add_nommu_region(struct vm_region *region)
486
{
487 488
	struct vm_region *pregion;
	struct rb_node **p, *parent;
489

490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
	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();
505 506
	}

507 508
	rb_link_node(&region->vm_rb, parent, p);
	rb_insert_color(&region->vm_rb, &nommu_region_tree);
509

510
	validate_nommu_regions();
511 512
}

513
/*
514
 * delete a region from the global tree
515
 */
516
static void delete_nommu_region(struct vm_region *region)
517
{
518
	BUG_ON(!nommu_region_tree.rb_node);
519

520 521 522
	validate_nommu_regions();
	rb_erase(&region->vm_rb, &nommu_region_tree);
	validate_nommu_regions();
523 524
}

525
/*
526
 * free a contiguous series of pages
527
 */
528
static void free_page_series(unsigned long from, unsigned long to)
529
{
530 531 532
	for (; from < to; from += PAGE_SIZE) {
		struct page *page = virt_to_page(from);

533
		atomic_long_dec(&mmap_pages_allocated);
534
		put_page(page);
535 536 537
	}
}

538
/*
539
 * release a reference to a region
540
 * - the caller must hold the region semaphore for writing, which this releases
541
 * - the region may not have been added to the tree yet, in which case vm_top
542
 *   will equal vm_start
543
 */
544 545
static void __put_nommu_region(struct vm_region *region)
	__releases(nommu_region_sem)
Linus Torvalds's avatar
Linus Torvalds committed
546
{
547
	BUG_ON(!nommu_region_tree.rb_node);
Linus Torvalds's avatar
Linus Torvalds committed
548

549
	if (--region->vm_usage == 0) {
550
		if (region->vm_top > region->vm_start)
551 552 553 554 555 556 557 558
			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 */
559
		if (region->vm_flags & VM_MAPPED_COPY)
560
			free_page_series(region->vm_start, region->vm_top);
561 562 563
		kmem_cache_free(vm_region_jar, region);
	} else {
		up_write(&nommu_region_sem);
Linus Torvalds's avatar
Linus Torvalds committed
564
	}
565
}
Linus Torvalds's avatar
Linus Torvalds committed
566

567 568 569 570 571 572 573
/*
 * release a reference to a region
 */
static void put_nommu_region(struct vm_region *region)
{
	down_write(&nommu_region_sem);
	__put_nommu_region(region);
Linus Torvalds's avatar
Linus Torvalds committed
574 575
}

576
/*
577 578 579 580
 * 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
581
 */
582
static void add_vma_to_mm(struct mm_struct *mm, struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
583
{
584
	struct vm_area_struct *pvma, *prev;
Linus Torvalds's avatar
Linus Torvalds committed
585
	struct address_space *mapping;
586
	struct rb_node **p, *parent, *rb_prev;
587 588 589 590 591

	BUG_ON(!vma->vm_region);

	mm->map_count++;
	vma->vm_mm = mm;
Linus Torvalds's avatar
Linus Torvalds committed
592 593 594 595 596

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

597
		i_mmap_lock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
598
		flush_dcache_mmap_lock(mapping);
599
		vma_interval_tree_insert(vma, &mapping->i_mmap);
Linus Torvalds's avatar
Linus Torvalds committed
600
		flush_dcache_mmap_unlock(mapping);
601
		i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
602 603
	}

604
	/* add the VMA to the tree */
605
	parent = rb_prev = NULL;
606
	p = &mm->mm_rb.rb_node;
Linus Torvalds's avatar
Linus Torvalds committed
607 608 609 610
	while (*p) {
		parent = *p;
		pvma = rb_entry(parent, struct vm_area_struct, vm_rb);

611 612 613
		/* 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)
Linus Torvalds's avatar
Linus Torvalds committed
614
			p = &(*p)->rb_left;
615 616
		else if (vma->vm_start > pvma->vm_start) {
			rb_prev = parent;
Linus Torvalds's avatar
Linus Torvalds committed
617
			p = &(*p)->rb_right;
618
		} else if (vma->vm_end < pvma->vm_end)
619
			p = &(*p)->rb_left;
620 621
		else if (vma->vm_end > pvma->vm_end) {
			rb_prev = parent;
622
			p = &(*p)->rb_right;
623
		} else if (vma < pvma)
624
			p = &(*p)->rb_left;
625 626
		else if (vma > pvma) {
			rb_prev = parent;
627
			p = &(*p)->rb_right;
628
		} else
629
			BUG();
Linus Torvalds's avatar
Linus Torvalds committed
630 631 632
	}

	rb_link_node(&vma->vm_rb, parent, p);
633 634 635
	rb_insert_color(&vma->vm_rb, &mm->mm_rb);

	/* add VMA to the VMA list also */
636 637 638
	prev = NULL;
	if (rb_prev)
		prev = rb_entry(rb_prev, struct vm_area_struct, vm_rb);
639

640
	__vma_link_list(mm, vma, prev, parent);
Linus Torvalds's avatar
Linus Torvalds committed
641 642
}

643
/*
644
 * delete a VMA from its owning mm_struct and address space
645
 */
646
static void delete_vma_from_mm(struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
647
{
Davidlohr Bueso's avatar
Davidlohr Bueso committed
648
	int i;
Linus Torvalds's avatar
Linus Torvalds committed
649
	struct address_space *mapping;
650
	struct mm_struct *mm = vma->vm_mm;
Davidlohr Bueso's avatar
Davidlohr Bueso committed
651
	struct task_struct *curr = current;
652 653

	mm->map_count--;
Davidlohr Bueso's avatar
Davidlohr Bueso committed
654 655
	for (i = 0; i < VMACACHE_SIZE; i++) {
		/* if the vma is cached, invalidate the entire cache */
656
		if (curr->vmacache.vmas[i] == vma) {
657
			vmacache_invalidate(mm);
Davidlohr Bueso's avatar
Davidlohr Bueso committed
658 659 660
			break;
		}
	}
Linus Torvalds's avatar
Linus Torvalds committed
661 662 663 664 665

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

666
		i_mmap_lock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
667
		flush_dcache_mmap_lock(mapping);
668
		vma_interval_tree_remove(vma, &mapping->i_mmap);
Linus Torvalds's avatar
Linus Torvalds committed
669
		flush_dcache_mmap_unlock(mapping);
670
		i_mmap_unlock_write(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
671 672
	}

673 674
	/* remove from the MM's tree and list */
	rb_erase(&vma->vm_rb, &mm->mm_rb);
675 676 677 678 679 680 681 682

	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;
683 684 685 686 687 688 689 690 691
}

/*
 * 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);
692
	if (vma->vm_file)
693 694
		fput(vma->vm_file);
	put_nommu_region(vma->vm_region);
695
	vm_area_free(vma);
696 697 698 699 700 701 702 703 704 705 706
}

/*
 * 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 */
Davidlohr Bueso's avatar
Davidlohr Bueso committed
707 708
	vma = vmacache_find(mm, addr);
	if (likely(vma))
709 710
		return vma;

711
	/* trawl the list (there may be multiple mappings in which addr
712
	 * resides) */
713
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
714 715 716
		if (vma->vm_start > addr)
			return NULL;
		if (vma->vm_end > addr) {
Davidlohr Bueso's avatar
Davidlohr Bueso committed
717
			vmacache_update(addr, vma);
718 719 720 721 722 723 724 725 726 727 728 729 730 731
			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)
{
732
	return find_vma(mm, addr);
733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
}

/*
 * 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 */
Davidlohr Bueso's avatar
Davidlohr Bueso committed
756 757
	vma = vmacache_find_exact(mm, addr, end);
	if (vma)
758 759
		return vma;

760
	/* trawl the list (there may be multiple mappings in which addr
761
	 * resides) */
762
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
763 764 765 766 767
		if (vma->vm_start < addr)
			continue;
		if (vma->vm_start > addr)
			return NULL;
		if (vma->vm_end == end) {
Davidlohr Bueso's avatar
Davidlohr Bueso committed
768
			vmacache_update(addr, vma);
769 770 771 772 773
			return vma;
		}
	}

	return NULL;
Linus Torvalds's avatar
Linus Torvalds committed
774 775 776 777 778 779 780 781 782 783 784 785 786 787
}

/*
 * 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)
{
788
	unsigned long capabilities, rlen;
Linus Torvalds's avatar
Linus Torvalds committed
789 790 791
	int ret;

	/* do the simple checks first */
792
	if (flags & MAP_FIXED)
Linus Torvalds's avatar
Linus Torvalds committed
793 794 795 796 797 798
		return -EINVAL;

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

799
	if (!len)
Linus Torvalds's avatar
Linus Torvalds committed
800 801
		return -EINVAL;

802
	/* Careful about overflows.. */
803 804
	rlen = PAGE_ALIGN(len);
	if (!rlen || rlen > TASK_SIZE)
805 806
		return -ENOMEM;

Linus Torvalds's avatar
Linus Torvalds committed
807
	/* offset overflow? */
808
	if ((pgoff + (rlen >> PAGE_SHIFT)) < pgoff)
809
		return -EOVERFLOW;
Linus Torvalds's avatar
Linus Torvalds committed
810 811 812

	if (file) {
		/* files must support mmap */
Al Viro's avatar
Al Viro committed
813
		if (!file->f_op->mmap)
Linus Torvalds's avatar
Linus Torvalds committed
814 815 816 817 818 819
			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
		 */
820 821 822
		if (file->f_op->mmap_capabilities) {
			capabilities = file->f_op->mmap_capabilities(file);
		} else {
Linus Torvalds's avatar
Linus Torvalds committed
823 824
			/* no explicit capabilities set, so assume some
			 * defaults */
Al Viro's avatar
Al Viro committed
825
			switch (file_inode(file)->i_mode & S_IFMT) {
Linus Torvalds's avatar
Linus Torvalds committed
826 827
			case S_IFREG:
			case S_IFBLK:
828
				capabilities = NOMMU_MAP_COPY;
Linus Torvalds's avatar
Linus Torvalds committed
829 830 831 832
				break;

			case S_IFCHR:
				capabilities =
833 834 835
					NOMMU_MAP_DIRECT |
					NOMMU_MAP_READ |
					NOMMU_MAP_WRITE;
Linus Torvalds's avatar
Linus Torvalds committed
836 837 838 839 840 841 842 843 844 845
				break;

			default:
				return -EINVAL;
			}
		}

		/* eliminate any capabilities that we can't support on this
		 * device */
		if (!file->f_op->get_unmapped_area)
846
			capabilities &= ~NOMMU_MAP_DIRECT;
Al Viro's avatar
Al Viro committed
847
		if (!(file->f_mode & FMODE_CAN_READ))
848
			capabilities &= ~NOMMU_MAP_COPY;
Linus Torvalds's avatar
Linus Torvalds committed
849

850 851 852 853
		/* The file shall have been opened with read permission. */
		if (!(file->f_mode & FMODE_READ))
			return -EACCES;

Linus Torvalds's avatar
Linus Torvalds committed
854 855 856 857 858 859
		if (flags & MAP_SHARED) {
			/* do checks for writing, appending and locking */
			if ((prot & PROT_WRITE) &&
			    !(file->f_mode & FMODE_WRITE))
				return -EACCES;

Al Viro's avatar
Al Viro committed
860
			if (IS_APPEND(file_inode(file)) &&
Linus Torvalds's avatar
Linus Torvalds committed
861 862 863
			    (file->f_mode & FMODE_WRITE))
				return -EACCES;

864
			if (locks_verify_locked(file))
Linus Torvalds's avatar
Linus Torvalds committed
865 866
				return -EAGAIN;

867
			if (!(capabilities & NOMMU_MAP_DIRECT))
Linus Torvalds's avatar
Linus Torvalds committed
868 869 870
				return -ENODEV;

			/* we mustn't privatise shared mappings */
871
			capabilities &= ~NOMMU_MAP_COPY;
872
		} else {
Linus Torvalds's avatar
Linus Torvalds committed
873 874
			/* we're going to read the file into private memory we
			 * allocate */
875
			if (!(capabilities & NOMMU_MAP_COPY))
Linus Torvalds's avatar
Linus Torvalds committed
876 877 878 879 880
				return -ENODEV;

			/* we don't permit a private writable mapping to be
			 * shared with the backing device */
			if (prot & PROT_WRITE)
881
				capabilities &= ~NOMMU_MAP_DIRECT;
Linus Torvalds's avatar
Linus Torvalds committed
882 883
		}

884 885 886 887
		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))
888
			    ) {
889
				capabilities &= ~NOMMU_MAP_DIRECT;
890
				if (flags & MAP_SHARED) {
891
					pr_warn("MAP_SHARED not completely supported on !MMU\n");
892 893 894 895 896
					return -EINVAL;
				}
			}
		}

Linus Torvalds's avatar
Linus Torvalds committed
897 898
		/* handle executable mappings and implied executable
		 * mappings */
899
		if (path_noexec(&file->f_path)) {
Linus Torvalds's avatar
Linus Torvalds committed
900 901
			if (prot & PROT_EXEC)
				return -EPERM;
902
		} else if ((prot & PROT_READ) && !(prot & PROT_EXEC)) {
Linus Torvalds's avatar
Linus Torvalds committed
903 904
			/* handle implication of PROT_EXEC by PROT_READ */
			if (current->personality & READ_IMPLIES_EXEC) {
905
				if (capabilities & NOMMU_MAP_EXEC)
Linus Torvalds's avatar
Linus Torvalds committed
906 907
					prot |= PROT_EXEC;
			}
908
		} else if ((prot & PROT_READ) &&
Linus Torvalds's avatar
Linus Torvalds committed
909
			 (prot & PROT_EXEC) &&
910
			 !(capabilities & NOMMU_MAP_EXEC)
Linus Torvalds's avatar
Linus Torvalds committed
911 912
			 ) {
			/* backing file is not executable, try to copy */
913
			capabilities &= ~NOMMU_MAP_DIRECT;
Linus Torvalds's avatar
Linus Torvalds committed
914
		}
915
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
916 917 918
		/* anonymous mappings are always memory backed and can be
		 * privately mapped
		 */
919
		capabilities = NOMMU_MAP_COPY;
Linus Torvalds's avatar
Linus Torvalds committed
920 921 922 923 924 925 926 927

		/* 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 */
928
	ret = security_mmap_addr(addr);
Linus Torvalds's avatar
Linus Torvalds committed
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947
	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;

948
	vm_flags = calc_vm_prot_bits(prot, 0) | calc_vm_flag_bits(flags);
Linus Torvalds's avatar
Linus Torvalds committed
949 950
	/* vm_flags |= mm->def_flags; */

951
	if (!(capabilities & NOMMU_MAP_DIRECT)) {
Linus Torvalds's avatar
Linus Torvalds committed
952
		/* attempt to share read-only copies of mapped file chunks */
953
		vm_flags |= VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC;
Linus Torvalds's avatar
Linus Torvalds committed
954 955
		if (file && !(prot & PROT_WRITE))
			vm_flags |= VM_MAYSHARE;
956
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
957 958 959
		/* overlay a shareable mapping on the backing device or inode
		 * if possible - used for chardevs, ramfs/tmpfs/shmfs and
		 * romfs/cramfs */
960
		vm_flags |= VM_MAYSHARE | (capabilities & NOMMU_VMFLAGS);
Linus Torvalds's avatar
Linus Torvalds committed
961
		if (flags & MAP_SHARED)
962
			vm_flags |= VM_SHARED;
Linus Torvalds's avatar
Linus Torvalds committed
963 964 965 966 967 968
	}

	/* 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
	 */
Tejun Heo's avatar
Tejun Heo committed
969
	if ((flags & MAP_PRIVATE) && current->ptrace)
Linus Torvalds's avatar
Linus Torvalds committed
970 971 972 973 974 975
		vm_flags &= ~VM_MAYSHARE;

	return vm_flags;
}

/*
976 977
 * set up a shared mapping on a file (the driver or filesystem provides and
 * pins the storage)
Linus Torvalds's avatar
Linus Torvalds committed
978
 */
979
static int do_mmap_shared_file(struct vm_area_struct *vma)
Linus Torvalds's avatar
Linus Torvalds committed
980 981 982
{
	int ret;

983
	ret = call_mmap(vma->vm_file, vma);
984 985
	if (ret == 0) {
		vma->vm_region->vm_top = vma->vm_region->vm_end;
986
		return 0;
987
	}
Linus Torvalds's avatar
Linus Torvalds committed
988 989 990
	if (ret != -ENOSYS)
		return ret;

991 992 993
	/* 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 */
Linus Torvalds's avatar
Linus Torvalds committed
994 995 996 997 998 999
	return -ENODEV;
}

/*
 * set up a private mapping or an anonymous shared mapping
 */
1000 1001
static int do_mmap_private(struct vm_area_struct *vma,
			   struct vm_region *region,
1002 1003
			   unsigned long len,
			   unsigned long capabilities)
Linus Torvalds's avatar
Linus Torvalds committed
1004
{
1005
	unsigned long total, point;
Linus Torvalds's avatar
Linus Torvalds committed
1006
	void *base;
1007
	int ret, order;
Linus Torvalds's avatar
Linus Torvalds committed
1008 1009 1010 1011 1012

	/* 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
	 */
1013
	if (capabilities & NOMMU_MAP_DIRECT) {
1014
		ret = call_mmap(vma->vm_file, vma);
1015
		if (ret == 0) {
Linus Torvalds's avatar
Linus Torvalds committed
1016
			/* shouldn't return success if we're not sharing */
1017 1018
			BUG_ON(!(vma->vm_flags & VM_MAYSHARE));
			vma->vm_region->vm_top = vma->vm_region->vm_end;
1019
			return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1020
		}
1021 1022
		if (ret != -ENOSYS)
			return ret;
Linus Torvalds's avatar
Linus Torvalds committed
1023 1024 1025 1026 1027 1028

		/* 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 */
	}

1029

Linus Torvalds's avatar
Linus Torvalds committed
1030 1031 1032 1033
	/* 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
	 */
Bob Liu's avatar
Bob Liu committed
1034
	order = get_order(len);
1035
	total = 1 << order;