truncate.c 16.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5
/*
 * mm/truncate.c - code for taking down pages from address_spaces
 *
 * Copyright (C) 2002, Linus Torvalds
 *
6
 * 10Sep2002	Andrew Morton
L
Linus Torvalds 已提交
7 8 9 10
 *		Initial version.
 */

#include <linux/kernel.h>
A
Alexey Dobriyan 已提交
11
#include <linux/backing-dev.h>
12
#include <linux/gfp.h>
L
Linus Torvalds 已提交
13
#include <linux/mm.h>
N
Nick Piggin 已提交
14
#include <linux/swap.h>
L
Linus Torvalds 已提交
15 16
#include <linux/module.h>
#include <linux/pagemap.h>
17
#include <linux/highmem.h>
L
Linus Torvalds 已提交
18
#include <linux/pagevec.h>
19
#include <linux/task_io_accounting_ops.h>
L
Linus Torvalds 已提交
20
#include <linux/buffer_head.h>	/* grr. try_to_release_page,
21
				   do_invalidatepage */
22
#include "internal.h"
L
Linus Torvalds 已提交
23 24


25
/**
26
 * do_invalidatepage - invalidate part or all of a page
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
 * @page: the page which is affected
 * @offset: the index of the truncation point
 *
 * do_invalidatepage() is called when all or part of the page has become
 * invalidated by a truncate operation.
 *
 * do_invalidatepage() does not have to release all buffers, but it must
 * ensure that no dirty buffer is left outside @offset and that no I/O
 * is underway against any of the blocks which are outside the truncation
 * point.  Because the caller is about to free (and possibly reuse) those
 * blocks on-disk.
 */
void do_invalidatepage(struct page *page, unsigned long offset)
{
	void (*invalidatepage)(struct page *, unsigned long);
	invalidatepage = page->mapping->a_ops->invalidatepage;
43
#ifdef CONFIG_BLOCK
44 45
	if (!invalidatepage)
		invalidatepage = block_invalidatepage;
46
#endif
47 48 49 50
	if (invalidatepage)
		(*invalidatepage)(page, offset);
}

L
Linus Torvalds 已提交
51 52
static inline void truncate_partial_page(struct page *page, unsigned partial)
{
53
	zero_user_segment(page, partial, PAGE_CACHE_SIZE);
54
	if (page_has_private(page))
L
Linus Torvalds 已提交
55 56 57
		do_invalidatepage(page, partial);
}

58 59 60 61 62 63 64 65 66 67 68 69 70 71
/*
 * This cancels just the dirty bit on the kernel page itself, it
 * does NOT actually remove dirty bits on any mmap's that may be
 * around. It also leaves the page tagged dirty, so any sync
 * activity will still find it on the dirty lists, and in particular,
 * clear_page_dirty_for_io() will still look at the dirty bits in
 * the VM.
 *
 * Doing this should *normally* only ever be done when a page
 * is truncated, and is not actually mapped anywhere at all. However,
 * fs/buffer.c does this when it notices that somebody has cleaned
 * out all the buffers on a page without actually doing it through
 * the VM. Can you say "ext3 is horribly ugly"? Tought you could.
 */
72 73
void cancel_dirty_page(struct page *page, unsigned int account_size)
{
74 75 76 77
	if (TestClearPageDirty(page)) {
		struct address_space *mapping = page->mapping;
		if (mapping && mapping_cap_account_dirty(mapping)) {
			dec_zone_page_state(page, NR_FILE_DIRTY);
78 79
			dec_bdi_stat(mapping->backing_dev_info,
					BDI_RECLAIMABLE);
80 81 82
			if (account_size)
				task_io_account_cancelled_write(account_size);
		}
83
	}
84
}
85
EXPORT_SYMBOL(cancel_dirty_page);
86

L
Linus Torvalds 已提交
87 88
/*
 * If truncate cannot remove the fs-private metadata from the page, the page
89
 * becomes orphaned.  It will be left on the LRU and may even be mapped into
90
 * user pagetables if we're racing with filemap_fault().
L
Linus Torvalds 已提交
91 92 93
 *
 * We need to bale out if page->mapping is no longer equal to the original
 * mapping.  This happens a) when the VM reclaimed the page while we waited on
94
 * its lock, b) when a concurrent invalidate_mapping_pages got there first and
L
Linus Torvalds 已提交
95 96
 * c) when tmpfs swizzles a page between a tmpfs inode and swapper_space.
 */
97
static int
L
Linus Torvalds 已提交
98 99 100
truncate_complete_page(struct address_space *mapping, struct page *page)
{
	if (page->mapping != mapping)
101
		return -EIO;
L
Linus Torvalds 已提交
102

103
	if (page_has_private(page))
L
Linus Torvalds 已提交
104 105
		do_invalidatepage(page, 0);

106 107
	cancel_dirty_page(page, PAGE_CACHE_SIZE);

108
	clear_page_mlock(page);
109
	remove_from_page_cache(page);
L
Linus Torvalds 已提交
110 111
	ClearPageMappedToDisk(page);
	page_cache_release(page);	/* pagecache ref */
112
	return 0;
L
Linus Torvalds 已提交
113 114 115
}

/*
116
 * This is for invalidate_mapping_pages().  That function can be called at
L
Linus Torvalds 已提交
117
 * any time, and is not supposed to throw away dirty pages.  But pages can
N
Nick Piggin 已提交
118 119
 * be marked dirty at any time too, so use remove_mapping which safely
 * discards clean, unused pages.
L
Linus Torvalds 已提交
120 121 122 123 124 125
 *
 * Returns non-zero if the page was successfully invalidated.
 */
static int
invalidate_complete_page(struct address_space *mapping, struct page *page)
{
N
Nick Piggin 已提交
126 127
	int ret;

L
Linus Torvalds 已提交
128 129 130
	if (page->mapping != mapping)
		return 0;

131
	if (page_has_private(page) && !try_to_release_page(page, 0))
L
Linus Torvalds 已提交
132 133
		return 0;

134
	clear_page_mlock(page);
N
Nick Piggin 已提交
135 136 137
	ret = remove_mapping(mapping, page);

	return ret;
L
Linus Torvalds 已提交
138 139
}

140 141 142 143 144 145 146 147 148 149
int truncate_inode_page(struct address_space *mapping, struct page *page)
{
	if (page_mapped(page)) {
		unmap_mapping_range(mapping,
				   (loff_t)page->index << PAGE_CACHE_SHIFT,
				   PAGE_CACHE_SIZE, 0);
	}
	return truncate_complete_page(mapping, page);
}

150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166
/*
 * Used to get rid of pages on hardware memory corruption.
 */
int generic_error_remove_page(struct address_space *mapping, struct page *page)
{
	if (!mapping)
		return -EINVAL;
	/*
	 * Only punch for normal data pages for now.
	 * Handling other types like directories would need more auditing.
	 */
	if (!S_ISREG(mapping->host->i_mode))
		return -EIO;
	return truncate_inode_page(mapping, page);
}
EXPORT_SYMBOL(generic_error_remove_page);

167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
/*
 * Safely invalidate one page from its pagecache mapping.
 * It only drops clean, unused pages. The page must be locked.
 *
 * Returns 1 if the page is successfully invalidated, otherwise 0.
 */
int invalidate_inode_page(struct page *page)
{
	struct address_space *mapping = page_mapping(page);
	if (!mapping)
		return 0;
	if (PageDirty(page) || PageWriteback(page))
		return 0;
	if (page_mapped(page))
		return 0;
	return invalidate_complete_page(mapping, page);
}

L
Linus Torvalds 已提交
185
/**
186
 * truncate_inode_pages - truncate range of pages specified by start & end byte offsets
L
Linus Torvalds 已提交
187 188
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
189
 * @lend: offset to which to truncate
L
Linus Torvalds 已提交
190
 *
191 192 193
 * Truncate the page cache, removing the pages that are between
 * specified offsets (and zeroing out partial page
 * (if lstart is not page aligned)).
L
Linus Torvalds 已提交
194 195 196 197 198 199 200 201 202 203 204 205 206 207
 *
 * Truncate takes two passes - the first pass is nonblocking.  It will not
 * block on page locks and it will not block on writeback.  The second pass
 * will wait.  This is to prevent as much IO as possible in the affected region.
 * The first pass will remove most pages, so the search cost of the second pass
 * is low.
 *
 * When looking at page->index outside the page lock we need to be careful to
 * copy it into a local to avoid races (it could change at any time).
 *
 * We pass down the cache-hot hint to the page freeing code.  Even if the
 * mapping is large, it is probably the case that the final pages are the most
 * recently touched, and freeing happens in ascending file offset order.
 */
208 209
void truncate_inode_pages_range(struct address_space *mapping,
				loff_t lstart, loff_t lend)
L
Linus Torvalds 已提交
210 211
{
	const pgoff_t start = (lstart + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
212
	pgoff_t end;
L
Linus Torvalds 已提交
213 214 215 216 217 218 219 220
	const unsigned partial = lstart & (PAGE_CACHE_SIZE - 1);
	struct pagevec pvec;
	pgoff_t next;
	int i;

	if (mapping->nrpages == 0)
		return;

221 222 223
	BUG_ON((lend & (PAGE_CACHE_SIZE - 1)) != (PAGE_CACHE_SIZE - 1));
	end = (lend >> PAGE_CACHE_SHIFT);

L
Linus Torvalds 已提交
224 225
	pagevec_init(&pvec, 0);
	next = start;
226 227
	while (next <= end &&
	       pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
228
		mem_cgroup_uncharge_start();
L
Linus Torvalds 已提交
229 230 231 232
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
			pgoff_t page_index = page->index;

233 234 235 236 237
			if (page_index > end) {
				next = page_index;
				break;
			}

L
Linus Torvalds 已提交
238 239 240
			if (page_index > next)
				next = page_index;
			next++;
N
Nick Piggin 已提交
241
			if (!trylock_page(page))
L
Linus Torvalds 已提交
242 243 244 245 246
				continue;
			if (PageWriteback(page)) {
				unlock_page(page);
				continue;
			}
247
			truncate_inode_page(mapping, page);
L
Linus Torvalds 已提交
248 249 250
			unlock_page(page);
		}
		pagevec_release(&pvec);
251
		mem_cgroup_uncharge_end();
L
Linus Torvalds 已提交
252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
		cond_resched();
	}

	if (partial) {
		struct page *page = find_lock_page(mapping, start - 1);
		if (page) {
			wait_on_page_writeback(page);
			truncate_partial_page(page, partial);
			unlock_page(page);
			page_cache_release(page);
		}
	}

	next = start;
	for ( ; ; ) {
		cond_resched();
		if (!pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
			if (next == start)
				break;
			next = start;
			continue;
		}
274 275 276 277
		if (pvec.pages[0]->index > end) {
			pagevec_release(&pvec);
			break;
		}
278
		mem_cgroup_uncharge_start();
L
Linus Torvalds 已提交
279 280 281
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];

282 283
			if (page->index > end)
				break;
L
Linus Torvalds 已提交
284 285
			lock_page(page);
			wait_on_page_writeback(page);
286
			truncate_inode_page(mapping, page);
L
Linus Torvalds 已提交
287 288 289 290 291 292
			if (page->index > next)
				next = page->index;
			next++;
			unlock_page(page);
		}
		pagevec_release(&pvec);
293
		mem_cgroup_uncharge_end();
L
Linus Torvalds 已提交
294 295
	}
}
296
EXPORT_SYMBOL(truncate_inode_pages_range);
L
Linus Torvalds 已提交
297

298 299 300 301 302
/**
 * truncate_inode_pages - truncate *all* the pages from an offset
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
 *
303
 * Called under (and serialised by) inode->i_mutex.
304 305 306 307 308
 */
void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
{
	truncate_inode_pages_range(mapping, lstart, (loff_t)-1);
}
L
Linus Torvalds 已提交
309 310
EXPORT_SYMBOL(truncate_inode_pages);

311 312 313 314 315 316 317 318 319 320 321 322 323 324 325
/**
 * invalidate_mapping_pages - Invalidate all the unlocked pages of one inode
 * @mapping: the address_space which holds the pages to invalidate
 * @start: the offset 'from' which to invalidate
 * @end: the offset 'to' which to invalidate (inclusive)
 *
 * This function only removes the unlocked pages, if you want to
 * remove all the pages of one inode, you must call truncate_inode_pages.
 *
 * invalidate_mapping_pages() will not block on IO activity. It will not
 * invalidate pages which are dirty, locked, under writeback or mapped into
 * pagetables.
 */
unsigned long invalidate_mapping_pages(struct address_space *mapping,
				       pgoff_t start, pgoff_t end)
L
Linus Torvalds 已提交
326 327 328 329 330 331 332 333 334
{
	struct pagevec pvec;
	pgoff_t next = start;
	unsigned long ret = 0;
	int i;

	pagevec_init(&pvec, 0);
	while (next <= end &&
			pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
335
		mem_cgroup_uncharge_start();
L
Linus Torvalds 已提交
336 337
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
338 339
			pgoff_t index;
			int lock_failed;
L
Linus Torvalds 已提交
340

N
Nick Piggin 已提交
341
			lock_failed = !trylock_page(page);
342 343 344 345 346 347 348 349 350 351

			/*
			 * We really shouldn't be looking at the ->index of an
			 * unlocked page.  But we're not allowed to lock these
			 * pages.  So we rely upon nobody altering the ->index
			 * of this (pinned-by-us) page.
			 */
			index = page->index;
			if (index > next)
				next = index;
L
Linus Torvalds 已提交
352
			next++;
353 354 355
			if (lock_failed)
				continue;

356 357
			ret += invalidate_inode_page(page);

L
Linus Torvalds 已提交
358 359 360 361 362
			unlock_page(page);
			if (next > end)
				break;
		}
		pagevec_release(&pvec);
363
		mem_cgroup_uncharge_end();
364
		cond_resched();
L
Linus Torvalds 已提交
365 366 367
	}
	return ret;
}
368
EXPORT_SYMBOL(invalidate_mapping_pages);
L
Linus Torvalds 已提交
369

370 371 372 373
/*
 * This is like invalidate_complete_page(), except it ignores the page's
 * refcount.  We do this because invalidate_inode_pages2() needs stronger
 * invalidation guarantees, and cannot afford to leave pages behind because
374 375
 * shrink_page_list() has a temp ref on them, or because they're transiently
 * sitting in the lru_cache_add() pagevecs.
376 377 378 379 380 381 382
 */
static int
invalidate_complete_page2(struct address_space *mapping, struct page *page)
{
	if (page->mapping != mapping)
		return 0;

383
	if (page_has_private(page) && !try_to_release_page(page, GFP_KERNEL))
384 385
		return 0;

N
Nick Piggin 已提交
386
	spin_lock_irq(&mapping->tree_lock);
387 388 389
	if (PageDirty(page))
		goto failed;

390
	clear_page_mlock(page);
391
	BUG_ON(page_has_private(page));
392
	__remove_from_page_cache(page);
N
Nick Piggin 已提交
393
	spin_unlock_irq(&mapping->tree_lock);
394
	mem_cgroup_uncharge_cache_page(page);
395 396 397 398

	if (mapping->a_ops->freepage)
		mapping->a_ops->freepage(page);

399 400 401
	page_cache_release(page);	/* pagecache ref */
	return 1;
failed:
N
Nick Piggin 已提交
402
	spin_unlock_irq(&mapping->tree_lock);
403 404 405
	return 0;
}

406 407 408 409 410 411 412 413 414
static int do_launder_page(struct address_space *mapping, struct page *page)
{
	if (!PageDirty(page))
		return 0;
	if (page->mapping != mapping || mapping->a_ops->launder_page == NULL)
		return 0;
	return mapping->a_ops->launder_page(page);
}

L
Linus Torvalds 已提交
415 416
/**
 * invalidate_inode_pages2_range - remove range of pages from an address_space
417
 * @mapping: the address_space
L
Linus Torvalds 已提交
418 419 420 421 422 423
 * @start: the page offset 'from' which to invalidate
 * @end: the page offset 'to' which to invalidate (inclusive)
 *
 * Any pages which are found to be mapped into pagetables are unmapped prior to
 * invalidation.
 *
424
 * Returns -EBUSY if any pages could not be invalidated.
L
Linus Torvalds 已提交
425 426 427 428 429 430 431 432
 */
int invalidate_inode_pages2_range(struct address_space *mapping,
				  pgoff_t start, pgoff_t end)
{
	struct pagevec pvec;
	pgoff_t next;
	int i;
	int ret = 0;
433
	int ret2 = 0;
L
Linus Torvalds 已提交
434 435 436 437 438
	int did_range_unmap = 0;
	int wrapped = 0;

	pagevec_init(&pvec, 0);
	next = start;
439
	while (next <= end && !wrapped &&
L
Linus Torvalds 已提交
440 441
		pagevec_lookup(&pvec, mapping, next,
			min(end - next, (pgoff_t)PAGEVEC_SIZE - 1) + 1)) {
442
		mem_cgroup_uncharge_start();
443
		for (i = 0; i < pagevec_count(&pvec); i++) {
L
Linus Torvalds 已提交
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460
			struct page *page = pvec.pages[i];
			pgoff_t page_index;

			lock_page(page);
			if (page->mapping != mapping) {
				unlock_page(page);
				continue;
			}
			page_index = page->index;
			next = page_index + 1;
			if (next == 0)
				wrapped = 1;
			if (page_index > end) {
				unlock_page(page);
				break;
			}
			wait_on_page_writeback(page);
461
			if (page_mapped(page)) {
L
Linus Torvalds 已提交
462 463 464 465 466
				if (!did_range_unmap) {
					/*
					 * Zap the rest of the file in one hit.
					 */
					unmap_mapping_range(mapping,
467 468
					   (loff_t)page_index<<PAGE_CACHE_SHIFT,
					   (loff_t)(end - page_index + 1)
L
Linus Torvalds 已提交
469 470 471 472 473 474 475 476
							<< PAGE_CACHE_SHIFT,
					    0);
					did_range_unmap = 1;
				} else {
					/*
					 * Just zap this page
					 */
					unmap_mapping_range(mapping,
477
					  (loff_t)page_index<<PAGE_CACHE_SHIFT,
L
Linus Torvalds 已提交
478 479 480
					  PAGE_CACHE_SIZE, 0);
				}
			}
481
			BUG_ON(page_mapped(page));
482 483 484
			ret2 = do_launder_page(mapping, page);
			if (ret2 == 0) {
				if (!invalidate_complete_page2(mapping, page))
485
					ret2 = -EBUSY;
486 487 488
			}
			if (ret2 < 0)
				ret = ret2;
L
Linus Torvalds 已提交
489 490 491
			unlock_page(page);
		}
		pagevec_release(&pvec);
492
		mem_cgroup_uncharge_end();
L
Linus Torvalds 已提交
493 494 495 496 497 498 499 500
		cond_resched();
	}
	return ret;
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2_range);

/**
 * invalidate_inode_pages2 - remove all pages from an address_space
501
 * @mapping: the address_space
L
Linus Torvalds 已提交
502 503 504 505
 *
 * Any pages which are found to be mapped into pagetables are unmapped prior to
 * invalidation.
 *
506
 * Returns -EBUSY if any pages could not be invalidated.
L
Linus Torvalds 已提交
507 508 509 510 511 512
 */
int invalidate_inode_pages2(struct address_space *mapping)
{
	return invalidate_inode_pages2_range(mapping, 0, -1);
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2);
N
npiggin@suse.de 已提交
513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531

/**
 * truncate_pagecache - unmap and remove pagecache that has been truncated
 * @inode: inode
 * @old: old file offset
 * @new: new file offset
 *
 * inode's new i_size must already be written before truncate_pagecache
 * is called.
 *
 * This function should typically be called before the filesystem
 * releases resources associated with the freed range (eg. deallocates
 * blocks). This way, pagecache will always stay logically coherent
 * with on-disk format, and the filesystem would not have to deal with
 * situations such as writepage being called for a page that has already
 * had its underlying blocks deallocated.
 */
void truncate_pagecache(struct inode *inode, loff_t old, loff_t new)
{
532 533 534 535 536 537 538 539 540 541 542 543 544 545
	struct address_space *mapping = inode->i_mapping;

	/*
	 * unmap_mapping_range is called twice, first simply for
	 * efficiency so that truncate_inode_pages does fewer
	 * single-page unmaps.  However after this first call, and
	 * before truncate_inode_pages finishes, it is possible for
	 * private pages to be COWed, which remain after
	 * truncate_inode_pages finishes, hence the second
	 * unmap_mapping_range call must be made for correctness.
	 */
	unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
	truncate_inode_pages(mapping, new);
	unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
N
npiggin@suse.de 已提交
546 547 548
}
EXPORT_SYMBOL(truncate_pagecache);

549 550 551 552 553
/**
 * truncate_setsize - update inode and pagecache for a new file size
 * @inode: inode
 * @newsize: new file size
 *
J
Jan Kara 已提交
554 555 556
 * truncate_setsize updates i_size and performs pagecache truncation (if
 * necessary) to @newsize. It will be typically be called from the filesystem's
 * setattr function when ATTR_SIZE is passed in.
557
 *
J
Jan Kara 已提交
558 559
 * Must be called with inode_mutex held and before all filesystem specific
 * block truncation has been performed.
560 561 562 563 564 565 566 567 568 569 570 571
 */
void truncate_setsize(struct inode *inode, loff_t newsize)
{
	loff_t oldsize;

	oldsize = inode->i_size;
	i_size_write(inode, newsize);

	truncate_pagecache(inode, oldsize, newsize);
}
EXPORT_SYMBOL(truncate_setsize);

N
npiggin@suse.de 已提交
572 573 574 575 576
/**
 * vmtruncate - unmap mappings "freed" by truncate() syscall
 * @inode: inode of the file used
 * @offset: file offset to start truncating
 *
577 578
 * This function is deprecated and truncate_setsize or truncate_pagecache
 * should be used instead, together with filesystem specific block truncation.
N
npiggin@suse.de 已提交
579 580 581 582 583
 */
int vmtruncate(struct inode *inode, loff_t offset)
{
	int error;

584
	error = inode_newsize_ok(inode, offset);
N
npiggin@suse.de 已提交
585 586
	if (error)
		return error;
587

588
	truncate_setsize(inode, offset);
N
npiggin@suse.de 已提交
589 590
	if (inode->i_op->truncate)
		inode->i_op->truncate(inode);
591
	return 0;
N
npiggin@suse.de 已提交
592 593
}
EXPORT_SYMBOL(vmtruncate);