[net-next-2.6.git] / mm / truncate.c

/*
 * mm/truncate.c - code for taking down pages from address_spaces
 *
 * Copyright (C) 2002, Linus Torvalds
 *
 * 10Sep2002	akpm@zip.com.au
 *		Initial version.
 */

#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/module.h>
#include <linux/pagemap.h>
#include <linux/pagevec.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/buffer_head.h>	/* grr. try_to_release_page,
				   do_invalidatepage */


/**
 * do_invalidatepage - invalidate part of all of a page
 * @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;
#ifdef CONFIG_BLOCK
	if (!invalidatepage)
		invalidatepage = block_invalidatepage;
#endif
	if (invalidatepage)
		(*invalidatepage)(page, offset);
}

static inline void truncate_partial_page(struct page *page, unsigned partial)
{
	memclear_highpage_flush(page, partial, PAGE_CACHE_SIZE-partial);
	if (PagePrivate(page))
		do_invalidatepage(page, partial);
}

void cancel_dirty_page(struct page *page, unsigned int account_size)
{
	/* If we're cancelling the page, it had better not be mapped any more */
	if (page_mapped(page)) {
		static unsigned int warncount;

		WARN_ON(++warncount < 5);
	}
		
	if (TestClearPageDirty(page)) {
		struct address_space *mapping = page->mapping;
		if (mapping && mapping_cap_account_dirty(mapping)) {
			dec_zone_page_state(page, NR_FILE_DIRTY);
			if (account_size)
				task_io_account_cancelled_write(account_size);
		}
	}
}
EXPORT_SYMBOL(cancel_dirty_page);

/*
 * If truncate cannot remove the fs-private metadata from the page, the page
 * becomes anonymous.  It will be left on the LRU and may even be mapped into
 * user pagetables if we're racing with filemap_nopage().
 *
 * 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
 * its lock, b) when a concurrent invalidate_inode_pages got there first and
 * c) when tmpfs swizzles a page between a tmpfs inode and swapper_space.
 */
static void
truncate_complete_page(struct address_space *mapping, struct page *page)
{
	if (page->mapping != mapping)
		return;

	cancel_dirty_page(page, PAGE_CACHE_SIZE);

	if (PagePrivate(page))
		do_invalidatepage(page, 0);

	ClearPageUptodate(page);
	ClearPageMappedToDisk(page);
	remove_from_page_cache(page);
	page_cache_release(page);	/* pagecache ref */
}

/*
 * This is for invalidate_inode_pages().  That function can be called at
 * any time, and is not supposed to throw away dirty pages.  But pages can
 * be marked dirty at any time too, so use remove_mapping which safely
 * discards clean, unused pages.
 *
 * Returns non-zero if the page was successfully invalidated.
 */
static int
invalidate_complete_page(struct address_space *mapping, struct page *page)
{
	int ret;

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

	if (PagePrivate(page) && !try_to_release_page(page, 0))
		return 0;

	ret = remove_mapping(mapping, page);

	return ret;
}

/**
 * truncate_inode_pages - truncate range of pages specified by start and
 * end byte offsets
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
 * @lend: offset to which to truncate
 *
 * Truncate the page cache, removing the pages that are between
 * specified offsets (and zeroing out partial page
 * (if lstart is not page aligned)).
 *
 * 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.
 */
void truncate_inode_pages_range(struct address_space *mapping,
				loff_t lstart, loff_t lend)
{
	const pgoff_t start = (lstart + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
	pgoff_t end;
	const unsigned partial = lstart & (PAGE_CACHE_SIZE - 1);
	struct pagevec pvec;
	pgoff_t next;
	int i;

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

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

	pagevec_init(&pvec, 0);
	next = start;
	while (next <= end &&
	       pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
			pgoff_t page_index = page->index;

			if (page_index > end) {
				next = page_index;
				break;
			}

			if (page_index > next)
				next = page_index;
			next++;
			if (TestSetPageLocked(page))
				continue;
			if (PageWriteback(page)) {
				unlock_page(page);
				continue;
			}
			truncate_complete_page(mapping, page);
			unlock_page(page);
		}
		pagevec_release(&pvec);
		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;
		}
		if (pvec.pages[0]->index > end) {
			pagevec_release(&pvec);
			break;
		}
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];

			if (page->index > end)
				break;
			lock_page(page);
			wait_on_page_writeback(page);
			if (page->index > next)
				next = page->index;
			next++;
			truncate_complete_page(mapping, page);
			unlock_page(page);
		}
		pagevec_release(&pvec);
	}
}
EXPORT_SYMBOL(truncate_inode_pages_range);

/**
 * truncate_inode_pages - truncate *all* the pages from an offset
 * @mapping: mapping to truncate
 * @lstart: offset from which to truncate
 *
 * Called under (and serialised by) inode->i_mutex.
 */
void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
{
	truncate_inode_pages_range(mapping, lstart, (loff_t)-1);
}
EXPORT_SYMBOL(truncate_inode_pages);

/**
 * 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)
{
	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)) {
		for (i = 0; i < pagevec_count(&pvec); i++) {
			struct page *page = pvec.pages[i];
			pgoff_t index;
			int lock_failed;

			lock_failed = TestSetPageLocked(page);

			/*
			 * 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;
			next++;
			if (lock_failed)
				continue;

			if (PageDirty(page) || PageWriteback(page))
				goto unlock;
			if (page_mapped(page))
				goto unlock;
			ret += invalidate_complete_page(mapping, page);
unlock:
			unlock_page(page);
			if (next > end)
				break;
		}
		pagevec_release(&pvec);
	}
	return ret;
}

unsigned long invalidate_inode_pages(struct address_space *mapping)
{
	return invalidate_mapping_pages(mapping, 0, ~0UL);
}
EXPORT_SYMBOL(invalidate_inode_pages);

/*
 * 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
 * shrink_list() has a temp ref on them, or because they're transiently sitting
 * in the lru_cache_add() pagevecs.
 */
static int
invalidate_complete_page2(struct address_space *mapping, struct page *page)
{
	if (page->mapping != mapping)
		return 0;

	if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL))
		return 0;

	write_lock_irq(&mapping->tree_lock);
	if (PageDirty(page))
		goto failed;

	BUG_ON(PagePrivate(page));
	__remove_from_page_cache(page);
	write_unlock_irq(&mapping->tree_lock);
	ClearPageUptodate(page);
	page_cache_release(page);	/* pagecache ref */
	return 1;
failed:
	write_unlock_irq(&mapping->tree_lock);
	return 0;
}

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);
}

/**
 * invalidate_inode_pages2_range - remove range of pages from an address_space
 * @mapping: the address_space
 * @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.
 *
 * Returns -EIO if any pages could not be invalidated.
 */
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;
	int did_range_unmap = 0;
	int wrapped = 0;

	pagevec_init(&pvec, 0);
	next = start;
	while (next <= end && !ret && !wrapped &&
		pagevec_lookup(&pvec, mapping, next,
			min(end - next, (pgoff_t)PAGEVEC_SIZE - 1) + 1)) {
		for (i = 0; !ret && i < pagevec_count(&pvec); i++) {
			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);
			while (page_mapped(page)) {
				if (!did_range_unmap) {
					/*
					 * Zap the rest of the file in one hit.
					 */
					unmap_mapping_range(mapping,
					   (loff_t)page_index<<PAGE_CACHE_SHIFT,
					   (loff_t)(end - page_index + 1)
							<< PAGE_CACHE_SHIFT,
					    0);
					did_range_unmap = 1;
				} else {
					/*
					 * Just zap this page
					 */
					unmap_mapping_range(mapping,
					  (loff_t)page_index<<PAGE_CACHE_SHIFT,
					  PAGE_CACHE_SIZE, 0);
				}
			}
			ret = do_launder_page(mapping, page);
			if (ret == 0 && !invalidate_complete_page2(mapping, page))
				ret = -EIO;
			unlock_page(page);
		}
		pagevec_release(&pvec);
		cond_resched();
	}
	WARN_ON_ONCE(ret);
	return ret;
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2_range);

/**
 * invalidate_inode_pages2 - remove all pages from an address_space
 * @mapping: the address_space
 *
 * Any pages which are found to be mapped into pagetables are unmapped prior to
 * invalidation.
 *
 * Returns -EIO if any pages could not be invalidated.
 */
int invalidate_inode_pages2(struct address_space *mapping)
{
	return invalidate_inode_pages2_range(mapping, 0, -1);
}
EXPORT_SYMBOL_GPL(invalidate_inode_pages2);
Commit	Line	Data
1da177e4 LT	1	/*
	2	* mm/truncate.c - code for taking down pages from address_spaces
	3	*
	4	* Copyright (C) 2002, Linus Torvalds
	5	*
	6	* 10Sep2002 akpm@zip.com.au
	7	* Initial version.
	8	*/
	9
	10	#include <linux/kernel.h>
	11	#include <linux/mm.h>
0fd0e6b0	12	#include <linux/swap.h>
1da177e4 LT	13	#include <linux/module.h>
	14	#include <linux/pagemap.h>
	15	#include <linux/pagevec.h>
e08748ce	16	#include <linux/task_io_accounting_ops.h>
1da177e4	17	#include <linux/buffer_head.h> /* grr. try_to_release_page,
aaa4059b	18	do_invalidatepage */
1da177e4 LT	19
1da177e4 LT	20
cf9a2ae8 DH	21	/**
	22	* do_invalidatepage - invalidate part of all of a page
	23	* @page: the page which is affected
	24	* @offset: the index of the truncation point
	25	*
	26	* do_invalidatepage() is called when all or part of the page has become
	27	* invalidated by a truncate operation.
	28	*
	29	* do_invalidatepage() does not have to release all buffers, but it must
	30	* ensure that no dirty buffer is left outside @offset and that no I/O
	31	* is underway against any of the blocks which are outside the truncation
	32	* point. Because the caller is about to free (and possibly reuse) those
	33	* blocks on-disk.
	34	*/
	35	void do_invalidatepage(struct page *page, unsigned long offset)
	36	{
	37	void (invalidatepage)(struct page , unsigned long);
	38	invalidatepage = page->mapping->a_ops->invalidatepage;
9361401e	39	#ifdef CONFIG_BLOCK
cf9a2ae8 DH	40	if (!invalidatepage)
cf9a2ae8 DH	41	invalidatepage = block_invalidatepage;
9361401e	42	#endif
cf9a2ae8 DH	43	if (invalidatepage)
	44	(*invalidatepage)(page, offset);
	45	}
	46
1da177e4 LT	47	static inline void truncate_partial_page(struct page *page, unsigned partial)
	48	{
	49	memclear_highpage_flush(page, partial, PAGE_CACHE_SIZE-partial);
	50	if (PagePrivate(page))
	51	do_invalidatepage(page, partial);
	52	}
	53
fba2591b LT	54	void cancel_dirty_page(struct page *page, unsigned int account_size)
	55	{
	56	/* If we're cancelling the page, it had better not be mapped any more */
	57	if (page_mapped(page)) {
	58	static unsigned int warncount;
	59
	60	WARN_ON(++warncount < 5);
	61	}
	62
8368e328 LT	63	if (TestClearPageDirty(page)) {
	64	struct address_space *mapping = page->mapping;
	65	if (mapping && mapping_cap_account_dirty(mapping)) {
	66	dec_zone_page_state(page, NR_FILE_DIRTY);
	67	if (account_size)
	68	task_io_account_cancelled_write(account_size);
	69	}
3e67c098	70	}
fba2591b	71	}
8368e328	72	EXPORT_SYMBOL(cancel_dirty_page);
fba2591b	73
1da177e4 LT	74	/*
	75	* If truncate cannot remove the fs-private metadata from the page, the page
	76	* becomes anonymous. It will be left on the LRU and may even be mapped into
	77	* user pagetables if we're racing with filemap_nopage().
	78	*
	79	* We need to bale out if page->mapping is no longer equal to the original
	80	* mapping. This happens a) when the VM reclaimed the page while we waited on
	81	* its lock, b) when a concurrent invalidate_inode_pages got there first and
	82	* c) when tmpfs swizzles a page between a tmpfs inode and swapper_space.
	83	*/
	84	static void
	85	truncate_complete_page(struct address_space mapping, struct page page)
	86	{
	87	if (page->mapping != mapping)
	88	return;
	89
3e67c098 AM	90	cancel_dirty_page(page, PAGE_CACHE_SIZE);
3e67c098 AM	91
1da177e4 LT	92	if (PagePrivate(page))
	93	do_invalidatepage(page, 0);
	94
1da177e4 LT	95	ClearPageUptodate(page);
	96	ClearPageMappedToDisk(page);
	97	remove_from_page_cache(page);
	98	page_cache_release(page); /* pagecache ref */
	99	}
	100
	101	/*
	102	* This is for invalidate_inode_pages(). That function can be called at
	103	* any time, and is not supposed to throw away dirty pages. But pages can
0fd0e6b0 NP	104	* be marked dirty at any time too, so use remove_mapping which safely
0fd0e6b0 NP	105	* discards clean, unused pages.
1da177e4 LT	106	*
	107	* Returns non-zero if the page was successfully invalidated.
	108	*/
	109	static int
	110	invalidate_complete_page(struct address_space mapping, struct page page)
	111	{
0fd0e6b0 NP	112	int ret;
0fd0e6b0 NP	113
1da177e4 LT	114	if (page->mapping != mapping)
	115	return 0;
	116
	117	if (PagePrivate(page) && !try_to_release_page(page, 0))
	118	return 0;
	119
0fd0e6b0	120	ret = remove_mapping(mapping, page);
0fd0e6b0 NP	121
0fd0e6b0 NP	122	return ret;
1da177e4 LT	123	}
	124
	125	/**
d7339071 HR	126	* truncate_inode_pages - truncate range of pages specified by start and
d7339071 HR	127	* end byte offsets
1da177e4 LT	128	* @mapping: mapping to truncate
1da177e4 LT	129	* @lstart: offset from which to truncate
d7339071	130	* @lend: offset to which to truncate
1da177e4	131	*
d7339071 HR	132	* Truncate the page cache, removing the pages that are between
	133	* specified offsets (and zeroing out partial page
	134	* (if lstart is not page aligned)).
1da177e4 LT	135	*
	136	* Truncate takes two passes - the first pass is nonblocking. It will not
	137	* block on page locks and it will not block on writeback. The second pass
	138	* will wait. This is to prevent as much IO as possible in the affected region.
	139	* The first pass will remove most pages, so the search cost of the second pass
	140	* is low.
	141	*
	142	* When looking at page->index outside the page lock we need to be careful to
	143	* copy it into a local to avoid races (it could change at any time).
	144	*
	145	* We pass down the cache-hot hint to the page freeing code. Even if the
	146	* mapping is large, it is probably the case that the final pages are the most
	147	* recently touched, and freeing happens in ascending file offset order.
1da177e4	148	*/
d7339071 HR	149	void truncate_inode_pages_range(struct address_space *mapping,
d7339071 HR	150	loff_t lstart, loff_t lend)
1da177e4 LT	151	{
1da177e4 LT	152	const pgoff_t start = (lstart + PAGE_CACHE_SIZE-1) >> PAGE_CACHE_SHIFT;
d7339071	153	pgoff_t end;
1da177e4 LT	154	const unsigned partial = lstart & (PAGE_CACHE_SIZE - 1);
	155	struct pagevec pvec;
	156	pgoff_t next;
	157	int i;
	158
	159	if (mapping->nrpages == 0)
	160	return;
	161
d7339071 HR	162	BUG_ON((lend & (PAGE_CACHE_SIZE - 1)) != (PAGE_CACHE_SIZE - 1));
	163	end = (lend >> PAGE_CACHE_SHIFT);
	164
1da177e4 LT	165	pagevec_init(&pvec, 0);
1da177e4 LT	166	next = start;
d7339071 HR	167	while (next <= end &&
d7339071 HR	168	pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
1da177e4 LT	169	for (i = 0; i < pagevec_count(&pvec); i++) {
	170	struct page *page = pvec.pages[i];
	171	pgoff_t page_index = page->index;
	172
d7339071 HR	173	if (page_index > end) {
	174	next = page_index;
	175	break;
	176	}
	177
1da177e4 LT	178	if (page_index > next)
	179	next = page_index;
	180	next++;
	181	if (TestSetPageLocked(page))
	182	continue;
	183	if (PageWriteback(page)) {
	184	unlock_page(page);
	185	continue;
	186	}
	187	truncate_complete_page(mapping, page);
	188	unlock_page(page);
	189	}
	190	pagevec_release(&pvec);
	191	cond_resched();
	192	}
	193
	194	if (partial) {
	195	struct page *page = find_lock_page(mapping, start - 1);
	196	if (page) {
	197	wait_on_page_writeback(page);
	198	truncate_partial_page(page, partial);
	199	unlock_page(page);
	200	page_cache_release(page);
	201	}
	202	}
	203
	204	next = start;
	205	for ( ; ; ) {
	206	cond_resched();
	207	if (!pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
	208	if (next == start)
	209	break;
	210	next = start;
	211	continue;
	212	}
d7339071 HR	213	if (pvec.pages[0]->index > end) {
	214	pagevec_release(&pvec);
	215	break;
	216	}
1da177e4 LT	217	for (i = 0; i < pagevec_count(&pvec); i++) {
	218	struct page *page = pvec.pages[i];
	219
d7339071 HR	220	if (page->index > end)
d7339071 HR	221	break;
1da177e4 LT	222	lock_page(page);
	223	wait_on_page_writeback(page);
	224	if (page->index > next)
	225	next = page->index;
	226	next++;
	227	truncate_complete_page(mapping, page);
	228	unlock_page(page);
	229	}
	230	pagevec_release(&pvec);
	231	}
	232	}
d7339071	233	EXPORT_SYMBOL(truncate_inode_pages_range);
1da177e4	234
d7339071 HR	235	/**
	236	* truncate_inode_pages - truncate all the pages from an offset
	237	* @mapping: mapping to truncate
	238	* @lstart: offset from which to truncate
	239	*
1b1dcc1b	240	* Called under (and serialised by) inode->i_mutex.
d7339071 HR	241	*/
	242	void truncate_inode_pages(struct address_space *mapping, loff_t lstart)
	243	{
	244	truncate_inode_pages_range(mapping, lstart, (loff_t)-1);
	245	}
1da177e4 LT	246	EXPORT_SYMBOL(truncate_inode_pages);
	247
	248	/**
	249	* invalidate_mapping_pages - Invalidate all the unlocked pages of one inode
	250	* @mapping: the address_space which holds the pages to invalidate
	251	* @start: the offset 'from' which to invalidate
	252	* @end: the offset 'to' which to invalidate (inclusive)
	253	*
	254	* This function only removes the unlocked pages, if you want to
	255	* remove all the pages of one inode, you must call truncate_inode_pages.
	256	*
	257	* invalidate_mapping_pages() will not block on IO activity. It will not
	258	* invalidate pages which are dirty, locked, under writeback or mapped into
	259	* pagetables.
	260	*/
	261	unsigned long invalidate_mapping_pages(struct address_space *mapping,
	262	pgoff_t start, pgoff_t end)
	263	{
	264	struct pagevec pvec;
	265	pgoff_t next = start;
	266	unsigned long ret = 0;
	267	int i;
	268
	269	pagevec_init(&pvec, 0);
	270	while (next <= end &&
	271	pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) {
	272	for (i = 0; i < pagevec_count(&pvec); i++) {
	273	struct page *page = pvec.pages[i];
e0f23603 N	274	pgoff_t index;
e0f23603 N	275	int lock_failed;
1da177e4	276
e0f23603 N	277	lock_failed = TestSetPageLocked(page);
	278
	279	/*
	280	* We really shouldn't be looking at the ->index of an
	281	* unlocked page. But we're not allowed to lock these
	282	* pages. So we rely upon nobody altering the ->index
	283	* of this (pinned-by-us) page.
	284	*/
	285	index = page->index;
	286	if (index > next)
	287	next = index;
1da177e4	288	next++;
e0f23603 N	289	if (lock_failed)
	290	continue;
	291
1da177e4 LT	292	if (PageDirty(page) \|\| PageWriteback(page))
	293	goto unlock;
	294	if (page_mapped(page))
	295	goto unlock;
	296	ret += invalidate_complete_page(mapping, page);
	297	unlock:
	298	unlock_page(page);
	299	if (next > end)
	300	break;
	301	}
	302	pagevec_release(&pvec);
1da177e4 LT	303	}
	304	return ret;
	305	}
	306
	307	unsigned long invalidate_inode_pages(struct address_space *mapping)
	308	{
	309	return invalidate_mapping_pages(mapping, 0, ~0UL);
	310	}
1da177e4 LT	311	EXPORT_SYMBOL(invalidate_inode_pages);
1da177e4 LT	312
bd4c8ce4 AM	313	/*
	314	* This is like invalidate_complete_page(), except it ignores the page's
	315	* refcount. We do this because invalidate_inode_pages2() needs stronger
	316	* invalidation guarantees, and cannot afford to leave pages behind because
	317	* shrink_list() has a temp ref on them, or because they're transiently sitting
	318	* in the lru_cache_add() pagevecs.
	319	*/
	320	static int
	321	invalidate_complete_page2(struct address_space mapping, struct page page)
	322	{
	323	if (page->mapping != mapping)
	324	return 0;
	325
887ed2f3	326	if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL))
bd4c8ce4 AM	327	return 0;
	328
	329	write_lock_irq(&mapping->tree_lock);
	330	if (PageDirty(page))
	331	goto failed;
	332
	333	BUG_ON(PagePrivate(page));
	334	__remove_from_page_cache(page);
	335	write_unlock_irq(&mapping->tree_lock);
	336	ClearPageUptodate(page);
	337	page_cache_release(page); /* pagecache ref */
	338	return 1;
	339	failed:
	340	write_unlock_irq(&mapping->tree_lock);
	341	return 0;
	342	}
	343
e3db7691 TM	344	static int do_launder_page(struct address_space mapping, struct page page)
	345	{
	346	if (!PageDirty(page))
	347	return 0;
	348	if (page->mapping != mapping \|\| mapping->a_ops->launder_page == NULL)
	349	return 0;
	350	return mapping->a_ops->launder_page(page);
	351	}
	352
1da177e4 LT	353	/**
1da177e4 LT	354	* invalidate_inode_pages2_range - remove range of pages from an address_space
67be2dd1	355	* @mapping: the address_space
1da177e4 LT	356	* @start: the page offset 'from' which to invalidate
	357	* @end: the page offset 'to' which to invalidate (inclusive)
	358	*
	359	* Any pages which are found to be mapped into pagetables are unmapped prior to
	360	* invalidation.
	361	*
	362	* Returns -EIO if any pages could not be invalidated.
	363	*/
	364	int invalidate_inode_pages2_range(struct address_space *mapping,
	365	pgoff_t start, pgoff_t end)
	366	{
	367	struct pagevec pvec;
	368	pgoff_t next;
	369	int i;
	370	int ret = 0;
	371	int did_range_unmap = 0;
	372	int wrapped = 0;
	373
	374	pagevec_init(&pvec, 0);
	375	next = start;
	376	while (next <= end && !ret && !wrapped &&
	377	pagevec_lookup(&pvec, mapping, next,
	378	min(end - next, (pgoff_t)PAGEVEC_SIZE - 1) + 1)) {
	379	for (i = 0; !ret && i < pagevec_count(&pvec); i++) {
	380	struct page *page = pvec.pages[i];
	381	pgoff_t page_index;
1da177e4 LT	382
	383	lock_page(page);
	384	if (page->mapping != mapping) {
	385	unlock_page(page);
	386	continue;
	387	}
	388	page_index = page->index;
	389	next = page_index + 1;
	390	if (next == 0)
	391	wrapped = 1;
	392	if (page_index > end) {
	393	unlock_page(page);
	394	break;
	395	}
	396	wait_on_page_writeback(page);
	397	while (page_mapped(page)) {
	398	if (!did_range_unmap) {
	399	/*
	400	* Zap the rest of the file in one hit.
	401	*/
	402	unmap_mapping_range(mapping,
479ef592 OD	403	(loff_t)page_index<<PAGE_CACHE_SHIFT,
479ef592 OD	404	(loff_t)(end - page_index + 1)
1da177e4 LT	405	<< PAGE_CACHE_SHIFT,
	406	0);
	407	did_range_unmap = 1;
	408	} else {
	409	/*
	410	* Just zap this page
	411	*/
	412	unmap_mapping_range(mapping,
479ef592	413	(loff_t)page_index<<PAGE_CACHE_SHIFT,
1da177e4 LT	414	PAGE_CACHE_SIZE, 0);
	415	}
	416	}
e3db7691 TM	417	ret = do_launder_page(mapping, page);
e3db7691 TM	418	if (ret == 0 && !invalidate_complete_page2(mapping, page))
1da177e4	419	ret = -EIO;
1da177e4 LT	420	unlock_page(page);
	421	}
	422	pagevec_release(&pvec);
	423	cond_resched();
	424	}
8258d4a5	425	WARN_ON_ONCE(ret);
1da177e4 LT	426	return ret;
	427	}
	428	EXPORT_SYMBOL_GPL(invalidate_inode_pages2_range);
	429
	430	/**
	431	* invalidate_inode_pages2 - remove all pages from an address_space
67be2dd1	432	* @mapping: the address_space
1da177e4 LT	433	*
	434	* Any pages which are found to be mapped into pagetables are unmapped prior to
	435	* invalidation.
	436	*
	437	* Returns -EIO if any pages could not be invalidated.
	438	*/
	439	int invalidate_inode_pages2(struct address_space *mapping)
	440	{
	441	return invalidate_inode_pages2_range(mapping, 0, -1);
	442	}
	443	EXPORT_SYMBOL_GPL(invalidate_inode_pages2);