truncate.c 23.6 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
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
Linus Torvalds's avatar
Linus Torvalds committed
7 8 9 10
 *		Initial version.
 */

#include <linux/kernel.h>
11
#include <linux/backing-dev.h>
12
#include <linux/dax.h>
13
#include <linux/gfp.h>
Linus Torvalds's avatar
Linus Torvalds committed
14
#include <linux/mm.h>
15
#include <linux/swap.h>
16
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
17
#include <linux/pagemap.h>
18
#include <linux/highmem.h>
Linus Torvalds's avatar
Linus Torvalds committed
19
#include <linux/pagevec.h>
20
#include <linux/task_io_accounting_ops.h>
Linus Torvalds's avatar
Linus Torvalds committed
21
#include <linux/buffer_head.h>	/* grr. try_to_release_page,
22
				   do_invalidatepage */
23
#include <linux/cleancache.h>
24
#include <linux/rmap.h>
25
#include "internal.h"
Linus Torvalds's avatar
Linus Torvalds committed
26

27 28 29
static void clear_exceptional_entry(struct address_space *mapping,
				    pgoff_t index, void *entry)
{
30 31 32
	struct radix_tree_node *node;
	void **slot;

33 34 35 36
	/* Handled by shmem itself */
	if (shmem_mapping(mapping))
		return;

37
	if (dax_mapping(mapping)) {
Jan Kara's avatar
Jan Kara committed
38 39
		dax_delete_mapping_entry(mapping, index);
		return;
40
	}
Jan Kara's avatar
Jan Kara committed
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68
	spin_lock_irq(&mapping->tree_lock);
	/*
	 * Regular page slots are stabilized by the page lock even
	 * without the tree itself locked.  These unlocked entries
	 * need verification under the tree lock.
	 */
	if (!__radix_tree_lookup(&mapping->page_tree, index, &node,
				&slot))
		goto unlock;
	if (*slot != entry)
		goto unlock;
	radix_tree_replace_slot(slot, NULL);
	mapping->nrexceptional--;
	if (!node)
		goto unlock;
	workingset_node_shadows_dec(node);
	/*
	 * Don't track node without shadow entries.
	 *
	 * Avoid acquiring the list_lru lock if already untracked.
	 * The list_empty() test is safe as node->private_list is
	 * protected by mapping->tree_lock.
	 */
	if (!workingset_node_shadows(node) &&
	    !list_empty(&node->private_list))
		list_lru_del(&workingset_shadow_nodes,
				&node->private_list);
	__radix_tree_delete_node(&mapping->page_tree, node);
69
unlock:
70 71
	spin_unlock_irq(&mapping->tree_lock);
}
Linus Torvalds's avatar
Linus Torvalds committed
72

73
/**
74
 * do_invalidatepage - invalidate part or all of a page
75
 * @page: the page which is affected
76 77
 * @offset: start of the range to invalidate
 * @length: length of the range to invalidate
78 79 80 81 82 83 84 85 86 87
 *
 * 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.
 */
88 89
void do_invalidatepage(struct page *page, unsigned int offset,
		       unsigned int length)
90
{
91 92
	void (*invalidatepage)(struct page *, unsigned int, unsigned int);

93
	invalidatepage = page->mapping->a_ops->invalidatepage;
94
#ifdef CONFIG_BLOCK
95 96
	if (!invalidatepage)
		invalidatepage = block_invalidatepage;
97
#endif
98
	if (invalidatepage)
99
		(*invalidatepage)(page, offset, length);
100 101
}

Linus Torvalds's avatar
Linus Torvalds committed
102 103
/*
 * If truncate cannot remove the fs-private metadata from the page, the page
104
 * becomes orphaned.  It will be left on the LRU and may even be mapped into
105
 * user pagetables if we're racing with filemap_fault().
Linus Torvalds's avatar
Linus Torvalds committed
106 107 108
 *
 * 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
109
 * its lock, b) when a concurrent invalidate_mapping_pages got there first and
Linus Torvalds's avatar
Linus Torvalds committed
110 111
 * c) when tmpfs swizzles a page between a tmpfs inode and swapper_space.
 */
112
static int
Linus Torvalds's avatar
Linus Torvalds committed
113 114 115
truncate_complete_page(struct address_space *mapping, struct page *page)
{
	if (page->mapping != mapping)
116
		return -EIO;
Linus Torvalds's avatar
Linus Torvalds committed
117

118
	if (page_has_private(page))
119
		do_invalidatepage(page, 0, PAGE_SIZE);
Linus Torvalds's avatar
Linus Torvalds committed
120

121 122 123 124 125
	/*
	 * Some filesystems seem to re-dirty the page even after
	 * the VM has canceled the dirty bit (eg ext3 journaling).
	 * Hence dirty accounting check is placed after invalidation.
	 */
126
	cancel_dirty_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
127
	ClearPageMappedToDisk(page);
128
	delete_from_page_cache(page);
129
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
130 131 132
}

/*
133
 * This is for invalidate_mapping_pages().  That function can be called at
Linus Torvalds's avatar
Linus Torvalds committed
134
 * any time, and is not supposed to throw away dirty pages.  But pages can
135 136
 * be marked dirty at any time too, so use remove_mapping which safely
 * discards clean, unused pages.
Linus Torvalds's avatar
Linus Torvalds committed
137 138 139 140 141 142
 *
 * Returns non-zero if the page was successfully invalidated.
 */
static int
invalidate_complete_page(struct address_space *mapping, struct page *page)
{
143 144
	int ret;

Linus Torvalds's avatar
Linus Torvalds committed
145 146 147
	if (page->mapping != mapping)
		return 0;

148
	if (page_has_private(page) && !try_to_release_page(page, 0))
Linus Torvalds's avatar
Linus Torvalds committed
149 150
		return 0;

151 152 153
	ret = remove_mapping(mapping, page);

	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
154 155
}

156 157 158 159
int truncate_inode_page(struct address_space *mapping, struct page *page)
{
	if (page_mapped(page)) {
		unmap_mapping_range(mapping,
160 161
				   (loff_t)page->index << PAGE_SHIFT,
				   PAGE_SIZE, 0);
162 163 164 165
	}
	return truncate_complete_page(mapping, page);
}

166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
/*
 * 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);

183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
/*
 * 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);
}

Linus Torvalds's avatar
Linus Torvalds committed
201
/**
202
 * truncate_inode_pages_range - truncate range of pages specified by start & end byte offsets
Linus Torvalds's avatar
Linus Torvalds committed
203 204
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
205
 * @lend: offset to which to truncate (inclusive)
Linus Torvalds's avatar
Linus Torvalds committed
206
 *
207
 * Truncate the page cache, removing the pages that are between
208 209
 * specified offsets (and zeroing out partial pages
 * if lstart or lend + 1 is not page aligned).
Linus Torvalds's avatar
Linus Torvalds committed
210 211 212 213 214 215 216 217 218 219
 *
 * 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.
 *
 * 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.
220 221 222 223
 *
 * Note that since ->invalidatepage() accepts range to invalidate
 * truncate_inode_pages_range is able to handle cases where lend + 1 is not
 * page aligned properly.
Linus Torvalds's avatar
Linus Torvalds committed
224
 */
225 226
void truncate_inode_pages_range(struct address_space *mapping,
				loff_t lstart, loff_t lend)
Linus Torvalds's avatar
Linus Torvalds committed
227
{
228 229 230 231 232
	pgoff_t		start;		/* inclusive */
	pgoff_t		end;		/* exclusive */
	unsigned int	partial_start;	/* inclusive */
	unsigned int	partial_end;	/* exclusive */
	struct pagevec	pvec;
233
	pgoff_t		indices[PAGEVEC_SIZE];
234 235
	pgoff_t		index;
	int		i;
Linus Torvalds's avatar
Linus Torvalds committed
236

237
	cleancache_invalidate_inode(mapping);
238
	if (mapping->nrpages == 0 && mapping->nrexceptional == 0)
Linus Torvalds's avatar
Linus Torvalds committed
239 240
		return;

241
	/* Offsets within partial pages */
242 243
	partial_start = lstart & (PAGE_SIZE - 1);
	partial_end = (lend + 1) & (PAGE_SIZE - 1);
244 245 246 247 248 249 250

	/*
	 * 'start' and 'end' always covers the range of pages to be fully
	 * truncated. Partial pages are covered with 'partial_start' at the
	 * start of the range and 'partial_end' at the end of the range.
	 * Note that 'end' is exclusive while 'lend' is inclusive.
	 */
251
	start = (lstart + PAGE_SIZE - 1) >> PAGE_SHIFT;
252 253 254 255 256 257 258 259
	if (lend == -1)
		/*
		 * lend == -1 indicates end-of-file so we have to set 'end'
		 * to the highest possible pgoff_t and since the type is
		 * unsigned we're using -1.
		 */
		end = -1;
	else
260
		end = (lend + 1) >> PAGE_SHIFT;
261

Linus Torvalds's avatar
Linus Torvalds committed
262
	pagevec_init(&pvec, 0);
263
	index = start;
264 265 266
	while (index < end && pagevec_lookup_entries(&pvec, mapping, index,
			min(end - index, (pgoff_t)PAGEVEC_SIZE),
			indices)) {
Linus Torvalds's avatar
Linus Torvalds committed
267 268 269
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];

270
			/* We rely upon deletion not changing page->index */
271
			index = indices[i];
272
			if (index >= end)
273 274
				break;

275 276 277 278 279
			if (radix_tree_exceptional_entry(page)) {
				clear_exceptional_entry(mapping, index, page);
				continue;
			}

Nick Piggin's avatar
Nick Piggin committed
280
			if (!trylock_page(page))
Linus Torvalds's avatar
Linus Torvalds committed
281
				continue;
282
			WARN_ON(page->index != index);
Linus Torvalds's avatar
Linus Torvalds committed
283 284 285 286
			if (PageWriteback(page)) {
				unlock_page(page);
				continue;
			}
287
			truncate_inode_page(mapping, page);
Linus Torvalds's avatar
Linus Torvalds committed
288 289
			unlock_page(page);
		}
290
		pagevec_remove_exceptionals(&pvec);
Linus Torvalds's avatar
Linus Torvalds committed
291 292
		pagevec_release(&pvec);
		cond_resched();
293
		index++;
Linus Torvalds's avatar
Linus Torvalds committed
294 295
	}

296
	if (partial_start) {
Linus Torvalds's avatar
Linus Torvalds committed
297 298
		struct page *page = find_lock_page(mapping, start - 1);
		if (page) {
299
			unsigned int top = PAGE_SIZE;
300 301 302 303 304
			if (start > end) {
				/* Truncation within a single page */
				top = partial_end;
				partial_end = 0;
			}
Linus Torvalds's avatar
Linus Torvalds committed
305
			wait_on_page_writeback(page);
306 307 308 309 310
			zero_user_segment(page, partial_start, top);
			cleancache_invalidate_page(mapping, page);
			if (page_has_private(page))
				do_invalidatepage(page, partial_start,
						  top - partial_start);
Linus Torvalds's avatar
Linus Torvalds committed
311
			unlock_page(page);
312
			put_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
313 314
		}
	}
315 316 317 318 319 320 321 322 323 324
	if (partial_end) {
		struct page *page = find_lock_page(mapping, end);
		if (page) {
			wait_on_page_writeback(page);
			zero_user_segment(page, 0, partial_end);
			cleancache_invalidate_page(mapping, page);
			if (page_has_private(page))
				do_invalidatepage(page, 0,
						  partial_end);
			unlock_page(page);
325
			put_page(page);
326 327 328 329 330 331 332 333
		}
	}
	/*
	 * If the truncation happened within a single page no pages
	 * will be released, just zeroed, so we can bail out now.
	 */
	if (start >= end)
		return;
Linus Torvalds's avatar
Linus Torvalds committed
334

335
	index = start;
Linus Torvalds's avatar
Linus Torvalds committed
336 337
	for ( ; ; ) {
		cond_resched();
338
		if (!pagevec_lookup_entries(&pvec, mapping, index,
339 340
			min(end - index, (pgoff_t)PAGEVEC_SIZE), indices)) {
			/* If all gone from start onwards, we're done */
341
			if (index == start)
Linus Torvalds's avatar
Linus Torvalds committed
342
				break;
343
			/* Otherwise restart to make sure all gone */
344
			index = start;
Linus Torvalds's avatar
Linus Torvalds committed
345 346
			continue;
		}
347
		if (index == start && indices[0] >= end) {
348
			/* All gone out of hole to be punched, we're done */
349
			pagevec_remove_exceptionals(&pvec);
350 351 352
			pagevec_release(&pvec);
			break;
		}
Linus Torvalds's avatar
Linus Torvalds committed
353 354 355
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];

356
			/* We rely upon deletion not changing page->index */
357
			index = indices[i];
358 359 360
			if (index >= end) {
				/* Restart punch to make sure all gone */
				index = start - 1;
361
				break;
362
			}
363

364 365 366 367 368
			if (radix_tree_exceptional_entry(page)) {
				clear_exceptional_entry(mapping, index, page);
				continue;
			}

Linus Torvalds's avatar
Linus Torvalds committed
369
			lock_page(page);
370
			WARN_ON(page->index != index);
Linus Torvalds's avatar
Linus Torvalds committed
371
			wait_on_page_writeback(page);
372
			truncate_inode_page(mapping, page);
Linus Torvalds's avatar
Linus Torvalds committed
373 374
			unlock_page(page);
		}
375
		pagevec_remove_exceptionals(&pvec);
Linus Torvalds's avatar
Linus Torvalds committed
376
		pagevec_release(&pvec);
377
		index++;
Linus Torvalds's avatar
Linus Torvalds committed
378
	}
379
	cleancache_invalidate_inode(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
380
}
381
EXPORT_SYMBOL(truncate_inode_pages_range);
Linus Torvalds's avatar
Linus Torvalds committed
382

383 384 385 386 387
/**
 * truncate_inode_pages - truncate *all* the pages from an offset
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
 *
388
 * Called under (and serialised by) inode->i_mutex.
389 390 391 392 393
 *
 * Note: When this function returns, there can be a page in the process of
 * deletion (inside __delete_from_page_cache()) in the specified range.  Thus
 * mapping->nrpages can be non-zero when this function returns even after
 * truncation of the whole mapping.
394 395 396 397 398
 */
void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
{
	truncate_inode_pages_range(mapping, lstart, (loff_t)-1);
}
Linus Torvalds's avatar
Linus Torvalds committed
399 400
EXPORT_SYMBOL(truncate_inode_pages);

401 402 403 404 405 406 407 408 409 410 411
/**
 * truncate_inode_pages_final - truncate *all* pages before inode dies
 * @mapping: mapping to truncate
 *
 * Called under (and serialized by) inode->i_mutex.
 *
 * Filesystems have to use this in the .evict_inode path to inform the
 * VM that this is the final truncate and the inode is going away.
 */
void truncate_inode_pages_final(struct address_space *mapping)
{
412
	unsigned long nrexceptional;
413 414 415 416 417 418 419 420 421 422 423 424 425
	unsigned long nrpages;

	/*
	 * Page reclaim can not participate in regular inode lifetime
	 * management (can't call iput()) and thus can race with the
	 * inode teardown.  Tell it when the address space is exiting,
	 * so that it does not install eviction information after the
	 * final truncate has begun.
	 */
	mapping_set_exiting(mapping);

	/*
	 * When reclaim installs eviction entries, it increases
426
	 * nrexceptional first, then decreases nrpages.  Make sure we see
427 428 429 430
	 * this in the right order or we might miss an entry.
	 */
	nrpages = mapping->nrpages;
	smp_rmb();
431
	nrexceptional = mapping->nrexceptional;
432

433
	if (nrpages || nrexceptional) {
434 435 436 437 438 439 440 441 442 443 444 445 446 447
		/*
		 * As truncation uses a lockless tree lookup, cycle
		 * the tree lock to make sure any ongoing tree
		 * modification that does not see AS_EXITING is
		 * completed before starting the final truncate.
		 */
		spin_lock_irq(&mapping->tree_lock);
		spin_unlock_irq(&mapping->tree_lock);

		truncate_inode_pages(mapping, 0);
	}
}
EXPORT_SYMBOL(truncate_inode_pages_final);

448 449 450 451 452 453 454 455 456 457 458 459 460 461
/**
 * 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,
462
		pgoff_t start, pgoff_t end)
Linus Torvalds's avatar
Linus Torvalds committed
463
{
464
	pgoff_t indices[PAGEVEC_SIZE];
Linus Torvalds's avatar
Linus Torvalds committed
465
	struct pagevec pvec;
466
	pgoff_t index = start;
467 468
	unsigned long ret;
	unsigned long count = 0;
Linus Torvalds's avatar
Linus Torvalds committed
469 470 471
	int i;

	pagevec_init(&pvec, 0);
472 473 474
	while (index <= end && pagevec_lookup_entries(&pvec, mapping, index,
			min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1,
			indices)) {
Linus Torvalds's avatar
Linus Torvalds committed
475 476
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
477

478
			/* We rely upon deletion not changing page->index */
479
			index = indices[i];
480 481
			if (index > end)
				break;
482

483 484 485 486 487
			if (radix_tree_exceptional_entry(page)) {
				clear_exceptional_entry(mapping, index, page);
				continue;
			}

488 489 490
			if (!trylock_page(page))
				continue;
			WARN_ON(page->index != index);
491
			ret = invalidate_inode_page(page);
Linus Torvalds's avatar
Linus Torvalds committed
492
			unlock_page(page);
493 494 495 496 497
			/*
			 * Invalidation is a hint that the page is no longer
			 * of interest and try to speed up its reclaim.
			 */
			if (!ret)
498
				deactivate_file_page(page);
499
			count += ret;
Linus Torvalds's avatar
Linus Torvalds committed
500
		}
501
		pagevec_remove_exceptionals(&pvec);
Linus Torvalds's avatar
Linus Torvalds committed
502
		pagevec_release(&pvec);
503
		cond_resched();
504
		index++;
Linus Torvalds's avatar
Linus Torvalds committed
505
	}
506
	return count;
Linus Torvalds's avatar
Linus Torvalds committed
507
}
508
EXPORT_SYMBOL(invalidate_mapping_pages);
Linus Torvalds's avatar
Linus Torvalds committed
509

510 511 512 513
/*
 * 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
514 515
 * shrink_page_list() has a temp ref on them, or because they're transiently
 * sitting in the lru_cache_add() pagevecs.
516 517 518 519
 */
static int
invalidate_complete_page2(struct address_space *mapping, struct page *page)
{
520 521
	unsigned long flags;

522 523 524
	if (page->mapping != mapping)
		return 0;

525
	if (page_has_private(page) && !try_to_release_page(page, GFP_KERNEL))
526 527
		return 0;

528
	spin_lock_irqsave(&mapping->tree_lock, flags);
529 530 531
	if (PageDirty(page))
		goto failed;

532
	BUG_ON(page_has_private(page));
533
	__delete_from_page_cache(page, NULL);
534
	spin_unlock_irqrestore(&mapping->tree_lock, flags);
535 536 537 538

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

539
	put_page(page);	/* pagecache ref */
540 541
	return 1;
failed:
542
	spin_unlock_irqrestore(&mapping->tree_lock, flags);
543 544 545
	return 0;
}

546 547 548 549 550 551 552 553 554
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);
}

Linus Torvalds's avatar
Linus Torvalds committed
555 556
/**
 * invalidate_inode_pages2_range - remove range of pages from an address_space
557
 * @mapping: the address_space
Linus Torvalds's avatar
Linus Torvalds committed
558 559 560 561 562 563
 * @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.
 *
564
 * Returns -EBUSY if any pages could not be invalidated.
Linus Torvalds's avatar
Linus Torvalds committed
565 566 567 568
 */
int invalidate_inode_pages2_range(struct address_space *mapping,
				  pgoff_t start, pgoff_t end)
{
569
	pgoff_t indices[PAGEVEC_SIZE];
Linus Torvalds's avatar
Linus Torvalds committed
570
	struct pagevec pvec;
571
	pgoff_t index;
Linus Torvalds's avatar
Linus Torvalds committed
572 573
	int i;
	int ret = 0;
574
	int ret2 = 0;
Linus Torvalds's avatar
Linus Torvalds committed
575 576
	int did_range_unmap = 0;

577
	cleancache_invalidate_inode(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
578
	pagevec_init(&pvec, 0);
579
	index = start;
580 581 582
	while (index <= end && pagevec_lookup_entries(&pvec, mapping, index,
			min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1,
			indices)) {
583
		for (i = 0; i < pagevec_count(&pvec); i++) {
Linus Torvalds's avatar
Linus Torvalds committed
584
			struct page *page = pvec.pages[i];
585 586

			/* We rely upon deletion not changing page->index */
587
			index = indices[i];
588 589
			if (index > end)
				break;
Linus Torvalds's avatar
Linus Torvalds committed
590

591 592 593 594 595
			if (radix_tree_exceptional_entry(page)) {
				clear_exceptional_entry(mapping, index, page);
				continue;
			}

Linus Torvalds's avatar
Linus Torvalds committed
596
			lock_page(page);
597
			WARN_ON(page->index != index);
Linus Torvalds's avatar
Linus Torvalds committed
598 599 600 601 602
			if (page->mapping != mapping) {
				unlock_page(page);
				continue;
			}
			wait_on_page_writeback(page);
603
			if (page_mapped(page)) {
Linus Torvalds's avatar
Linus Torvalds committed
604 605 606 607 608
				if (!did_range_unmap) {
					/*
					 * Zap the rest of the file in one hit.
					 */
					unmap_mapping_range(mapping,
609
					   (loff_t)index << PAGE_SHIFT,
610
					   (loff_t)(1 + end - index)
611 612
							 << PAGE_SHIFT,
							 0);
Linus Torvalds's avatar
Linus Torvalds committed
613 614 615 616 617 618
					did_range_unmap = 1;
				} else {
					/*
					 * Just zap this page
					 */
					unmap_mapping_range(mapping,
619 620
					   (loff_t)index << PAGE_SHIFT,
					   PAGE_SIZE, 0);
Linus Torvalds's avatar
Linus Torvalds committed
621 622
				}
			}
623
			BUG_ON(page_mapped(page));
624 625 626
			ret2 = do_launder_page(mapping, page);
			if (ret2 == 0) {
				if (!invalidate_complete_page2(mapping, page))
627
					ret2 = -EBUSY;
628 629 630
			}
			if (ret2 < 0)
				ret = ret2;
Linus Torvalds's avatar
Linus Torvalds committed
631 632
			unlock_page(page);
		}
633
		pagevec_remove_exceptionals(&pvec);
Linus Torvalds's avatar
Linus Torvalds committed
634 635
		pagevec_release(&pvec);
		cond_resched();
636
		index++;
Linus Torvalds's avatar
Linus Torvalds committed
637
	}
638
	cleancache_invalidate_inode(mapping);
Linus Torvalds's avatar
Linus Torvalds committed
639 640 641 642 643 644
	return ret;
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2_range);

/**
 * invalidate_inode_pages2 - remove all pages from an address_space
645
 * @mapping: the address_space
Linus Torvalds's avatar
Linus Torvalds committed
646 647 648 649
 *
 * Any pages which are found to be mapped into pagetables are unmapped prior to
 * invalidation.
 *
650
 * Returns -EBUSY if any pages could not be invalidated.
Linus Torvalds's avatar
Linus Torvalds committed
651 652 653 654 655 656
 */
int invalidate_inode_pages2(struct address_space *mapping)
{
	return invalidate_inode_pages2_range(mapping, 0, -1);
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2);
npiggin@suse.de's avatar
npiggin@suse.de committed
657 658 659 660

/**
 * truncate_pagecache - unmap and remove pagecache that has been truncated
 * @inode: inode
661
 * @newsize: new file size
npiggin@suse.de's avatar
npiggin@suse.de committed
662 663 664 665 666 667 668 669 670 671 672
 *
 * 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.
 */
673
void truncate_pagecache(struct inode *inode, loff_t newsize)
npiggin@suse.de's avatar
npiggin@suse.de committed
674
{
675
	struct address_space *mapping = inode->i_mapping;
676
	loff_t holebegin = round_up(newsize, PAGE_SIZE);
677 678 679 680 681 682 683 684 685 686

	/*
	 * 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.
	 */
687 688 689
	unmap_mapping_range(mapping, holebegin, 0, 1);
	truncate_inode_pages(mapping, newsize);
	unmap_mapping_range(mapping, holebegin, 0, 1);
npiggin@suse.de's avatar
npiggin@suse.de committed
690 691 692
}
EXPORT_SYMBOL(truncate_pagecache);

693 694 695 696 697
/**
 * truncate_setsize - update inode and pagecache for a new file size
 * @inode: inode
 * @newsize: new file size
 *
698 699 700
 * 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.
701
 *
702 703 704
 * Must be called with a lock serializing truncates and writes (generally
 * i_mutex but e.g. xfs uses a different lock) and before all filesystem
 * specific block truncation has been performed.
705 706 707
 */
void truncate_setsize(struct inode *inode, loff_t newsize)
{
708 709
	loff_t oldsize = inode->i_size;

710
	i_size_write(inode, newsize);
711 712
	if (newsize > oldsize)
		pagecache_isize_extended(inode, oldsize, newsize);
713
	truncate_pagecache(inode, newsize);
714 715 716
}
EXPORT_SYMBOL(truncate_setsize);

717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744
/**
 * pagecache_isize_extended - update pagecache after extension of i_size
 * @inode:	inode for which i_size was extended
 * @from:	original inode size
 * @to:		new inode size
 *
 * Handle extension of inode size either caused by extending truncate or by
 * write starting after current i_size. We mark the page straddling current
 * i_size RO so that page_mkwrite() is called on the nearest write access to
 * the page.  This way filesystem can be sure that page_mkwrite() is called on
 * the page before user writes to the page via mmap after the i_size has been
 * changed.
 *
 * The function must be called after i_size is updated so that page fault
 * coming after we unlock the page will already see the new i_size.
 * The function must be called while we still hold i_mutex - this not only
 * makes sure i_size is stable but also that userspace cannot observe new
 * i_size value before we are prepared to store mmap writes at new inode size.
 */
void pagecache_isize_extended(struct inode *inode, loff_t from, loff_t to)
{
	int bsize = 1 << inode->i_blkbits;
	loff_t rounded_from;
	struct page *page;
	pgoff_t index;

	WARN_ON(to > inode->i_size);

745
	if (from >= to || bsize == PAGE_SIZE)
746 747 748
		return;
	/* Page straddling @from will not have any hole block created? */
	rounded_from = round_up(from, bsize);
749
	if (to <= rounded_from || !(rounded_from & (PAGE_SIZE - 1)))
750 751
		return;

752
	index = from >> PAGE_SHIFT;
753 754 755 756 757 758 759 760 761 762 763
	page = find_lock_page(inode->i_mapping, index);
	/* Page not cached? Nothing to do */
	if (!page)
		return;
	/*
	 * See clear_page_dirty_for_io() for details why set_page_dirty()
	 * is needed.
	 */
	if (page_mkclean(page))
		set_page_dirty(page);
	unlock_page(page);
764
	put_page(page);
765 766 767
}
EXPORT_SYMBOL(pagecache_isize_extended);

768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789
/**
 * truncate_pagecache_range - unmap and remove pagecache that is hole-punched
 * @inode: inode
 * @lstart: offset of beginning of hole
 * @lend: offset of last byte of hole
 *
 * 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_range(struct inode *inode, loff_t lstart, loff_t lend)
{
	struct address_space *mapping = inode->i_mapping;
	loff_t unmap_start = round_up(lstart, PAGE_SIZE);
	loff_t unmap_end = round_down(1 + lend, PAGE_SIZE) - 1;
	/*
	 * This rounding is currently just for example: unmap_mapping_range
	 * expands its hole outwards, whereas we want it to contract the hole
	 * inwards.  However, existing callers of truncate_pagecache_range are
790 791
	 * doing their own page rounding first.  Note that unmap_mapping_range
	 * allows holelen 0 for all, and we allow lend -1 for end of file.
792 793 794 795 796 797 798 799 800 801 802 803 804
	 */

	/*
	 * Unlike in truncate_pagecache, unmap_mapping_range is called only
	 * once (before truncating pagecache), and without "even_cows" flag:
	 * hole-punching should not remove private COWed pages from the hole.
	 */
	if ((u64)unmap_end > (u64)unmap_start)
		unmap_mapping_range(mapping, unmap_start,
				    1 + unmap_end - unmap_start, 0);
	truncate_inode_pages_range(mapping, lstart, lend);
}
EXPORT_SYMBOL(truncate_pagecache_range);