[linux.git] / fs / nilfs2 / page.c

// SPDX-License-Identifier: GPL-2.0+
/*
 * Buffer/page management specific to NILFS
 *
 * Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation.
 *
 * Written by Ryusuke Konishi and Seiji Kihara.
 */

#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/swap.h>
#include <linux/bitops.h>
#include <linux/page-flags.h>
#include <linux/list.h>
#include <linux/highmem.h>
#include <linux/pagevec.h>
#include <linux/gfp.h>
#include "nilfs.h"
#include "page.h"
#include "mdt.h"


#define NILFS_BUFFER_INHERENT_BITS					\
	(BIT(BH_Uptodate) | BIT(BH_Mapped) | BIT(BH_NILFS_Node) |	\
	 BIT(BH_NILFS_Volatile) | BIT(BH_NILFS_Checked))

static struct buffer_head *__nilfs_get_folio_block(struct folio *folio,
		unsigned long block, pgoff_t index, int blkbits,
		unsigned long b_state)

{
	unsigned long first_block;
	struct buffer_head *bh = folio_buffers(folio);

	if (!bh)
		bh = create_empty_buffers(folio, 1 << blkbits, b_state);

	first_block = (unsigned long)index << (PAGE_SHIFT - blkbits);
	bh = get_nth_bh(bh, block - first_block);

	touch_buffer(bh);
	wait_on_buffer(bh);
	return bh;
}

struct buffer_head *nilfs_grab_buffer(struct inode *inode,
				      struct address_space *mapping,
				      unsigned long blkoff,
				      unsigned long b_state)
{
	int blkbits = inode->i_blkbits;
	pgoff_t index = blkoff >> (PAGE_SHIFT - blkbits);
	struct folio *folio;
	struct buffer_head *bh;

	folio = filemap_grab_folio(mapping, index);
	if (IS_ERR(folio))
		return NULL;

	bh = __nilfs_get_folio_block(folio, blkoff, index, blkbits, b_state);
	if (unlikely(!bh)) {
		folio_unlock(folio);
		folio_put(folio);
		return NULL;
	}
	return bh;
}

/**
 * nilfs_forget_buffer - discard dirty state
 * @bh: buffer head of the buffer to be discarded
 */
void nilfs_forget_buffer(struct buffer_head *bh)
{
	struct page *page = bh->b_page;
	const unsigned long clear_bits =
		(BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) |
		 BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) |
		 BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected));

	lock_buffer(bh);
	set_mask_bits(&bh->b_state, clear_bits, 0);
	if (nilfs_page_buffers_clean(page))
		__nilfs_clear_page_dirty(page);

	bh->b_blocknr = -1;
	ClearPageUptodate(page);
	ClearPageMappedToDisk(page);
	unlock_buffer(bh);
	brelse(bh);
}

/**
 * nilfs_copy_buffer -- copy buffer data and flags
 * @dbh: destination buffer
 * @sbh: source buffer
 */
void nilfs_copy_buffer(struct buffer_head *dbh, struct buffer_head *sbh)
{
	void *kaddr0, *kaddr1;
	unsigned long bits;
	struct page *spage = sbh->b_page, *dpage = dbh->b_page;
	struct buffer_head *bh;

	kaddr0 = kmap_atomic(spage);
	kaddr1 = kmap_atomic(dpage);
	memcpy(kaddr1 + bh_offset(dbh), kaddr0 + bh_offset(sbh), sbh->b_size);
	kunmap_atomic(kaddr1);
	kunmap_atomic(kaddr0);

	dbh->b_state = sbh->b_state & NILFS_BUFFER_INHERENT_BITS;
	dbh->b_blocknr = sbh->b_blocknr;
	dbh->b_bdev = sbh->b_bdev;

	bh = dbh;
	bits = sbh->b_state & (BIT(BH_Uptodate) | BIT(BH_Mapped));
	while ((bh = bh->b_this_page) != dbh) {
		lock_buffer(bh);
		bits &= bh->b_state;
		unlock_buffer(bh);
	}
	if (bits & BIT(BH_Uptodate))
		SetPageUptodate(dpage);
	else
		ClearPageUptodate(dpage);
	if (bits & BIT(BH_Mapped))
		SetPageMappedToDisk(dpage);
	else
		ClearPageMappedToDisk(dpage);
}

/**
 * nilfs_page_buffers_clean - check if a page has dirty buffers or not.
 * @page: page to be checked
 *
 * nilfs_page_buffers_clean() returns zero if the page has dirty buffers.
 * Otherwise, it returns non-zero value.
 */
int nilfs_page_buffers_clean(struct page *page)
{
	struct buffer_head *bh, *head;

	bh = head = page_buffers(page);
	do {
		if (buffer_dirty(bh))
			return 0;
		bh = bh->b_this_page;
	} while (bh != head);
	return 1;
}

void nilfs_page_bug(struct page *page)
{
	struct address_space *m;
	unsigned long ino;

	if (unlikely(!page)) {
		printk(KERN_CRIT "NILFS_PAGE_BUG(NULL)\n");
		return;
	}

	m = page->mapping;
	ino = m ? m->host->i_ino : 0;

	printk(KERN_CRIT "NILFS_PAGE_BUG(%p): cnt=%d index#=%llu flags=0x%lx "
	       "mapping=%p ino=%lu\n",
	       page, page_ref_count(page),
	       (unsigned long long)page->index, page->flags, m, ino);

	if (page_has_buffers(page)) {
		struct buffer_head *bh, *head;
		int i = 0;

		bh = head = page_buffers(page);
		do {
			printk(KERN_CRIT
			       " BH[%d] %p: cnt=%d block#=%llu state=0x%lx\n",
			       i++, bh, atomic_read(&bh->b_count),
			       (unsigned long long)bh->b_blocknr, bh->b_state);
			bh = bh->b_this_page;
		} while (bh != head);
	}
}

/**
 * nilfs_copy_folio -- copy the folio with buffers
 * @dst: destination folio
 * @src: source folio
 * @copy_dirty: flag whether to copy dirty states on the folio's buffer heads.
 *
 * This function is for both data folios and btnode folios.  The dirty flag
 * should be treated by caller.  The folio must not be under i/o.
 * Both src and dst folio must be locked
 */
static void nilfs_copy_folio(struct folio *dst, struct folio *src,
		bool copy_dirty)
{
	struct buffer_head *dbh, *dbufs, *sbh;
	unsigned long mask = NILFS_BUFFER_INHERENT_BITS;

	BUG_ON(folio_test_writeback(dst));

	sbh = folio_buffers(src);
	dbh = folio_buffers(dst);
	if (!dbh)
		dbh = create_empty_buffers(dst, sbh->b_size, 0);

	if (copy_dirty)
		mask |= BIT(BH_Dirty);

	dbufs = dbh;
	do {
		lock_buffer(sbh);
		lock_buffer(dbh);
		dbh->b_state = sbh->b_state & mask;
		dbh->b_blocknr = sbh->b_blocknr;
		dbh->b_bdev = sbh->b_bdev;
		sbh = sbh->b_this_page;
		dbh = dbh->b_this_page;
	} while (dbh != dbufs);

	folio_copy(dst, src);

	if (folio_test_uptodate(src) && !folio_test_uptodate(dst))
		folio_mark_uptodate(dst);
	else if (!folio_test_uptodate(src) && folio_test_uptodate(dst))
		folio_clear_uptodate(dst);
	if (folio_test_mappedtodisk(src) && !folio_test_mappedtodisk(dst))
		folio_set_mappedtodisk(dst);
	else if (!folio_test_mappedtodisk(src) && folio_test_mappedtodisk(dst))
		folio_clear_mappedtodisk(dst);

	do {
		unlock_buffer(sbh);
		unlock_buffer(dbh);
		sbh = sbh->b_this_page;
		dbh = dbh->b_this_page;
	} while (dbh != dbufs);
}

int nilfs_copy_dirty_pages(struct address_space *dmap,
			   struct address_space *smap)
{
	struct folio_batch fbatch;
	unsigned int i;
	pgoff_t index = 0;
	int err = 0;

	folio_batch_init(&fbatch);
repeat:
	if (!filemap_get_folios_tag(smap, &index, (pgoff_t)-1,
				PAGECACHE_TAG_DIRTY, &fbatch))
		return 0;

	for (i = 0; i < folio_batch_count(&fbatch); i++) {
		struct folio *folio = fbatch.folios[i], *dfolio;

		folio_lock(folio);
		if (unlikely(!folio_test_dirty(folio)))
			NILFS_PAGE_BUG(&folio->page, "inconsistent dirty state");

		dfolio = filemap_grab_folio(dmap, folio->index);
		if (unlikely(IS_ERR(dfolio))) {
			/* No empty page is added to the page cache */
			folio_unlock(folio);
			err = PTR_ERR(dfolio);
			break;
		}
		if (unlikely(!folio_buffers(folio)))
			NILFS_PAGE_BUG(&folio->page,
				       "found empty page in dat page cache");

		nilfs_copy_folio(dfolio, folio, true);
		filemap_dirty_folio(folio_mapping(dfolio), dfolio);

		folio_unlock(dfolio);
		folio_put(dfolio);
		folio_unlock(folio);
	}
	folio_batch_release(&fbatch);
	cond_resched();

	if (likely(!err))
		goto repeat;
	return err;
}

/**
 * nilfs_copy_back_pages -- copy back pages to original cache from shadow cache
 * @dmap: destination page cache
 * @smap: source page cache
 *
 * No pages must be added to the cache during this process.
 * This must be ensured by the caller.
 */
void nilfs_copy_back_pages(struct address_space *dmap,
			   struct address_space *smap)
{
	struct folio_batch fbatch;
	unsigned int i, n;
	pgoff_t start = 0;

	folio_batch_init(&fbatch);
repeat:
	n = filemap_get_folios(smap, &start, ~0UL, &fbatch);
	if (!n)
		return;

	for (i = 0; i < folio_batch_count(&fbatch); i++) {
		struct folio *folio = fbatch.folios[i], *dfolio;
		pgoff_t index = folio->index;

		folio_lock(folio);
		dfolio = filemap_lock_folio(dmap, index);
		if (!IS_ERR(dfolio)) {
			/* overwrite existing folio in the destination cache */
			WARN_ON(folio_test_dirty(dfolio));
			nilfs_copy_folio(dfolio, folio, false);
			folio_unlock(dfolio);
			folio_put(dfolio);
			/* Do we not need to remove folio from smap here? */
		} else {
			struct folio *f;

			/* move the folio to the destination cache */
			xa_lock_irq(&smap->i_pages);
			f = __xa_erase(&smap->i_pages, index);
			WARN_ON(folio != f);
			smap->nrpages--;
			xa_unlock_irq(&smap->i_pages);

			xa_lock_irq(&dmap->i_pages);
			f = __xa_store(&dmap->i_pages, index, folio, GFP_NOFS);
			if (unlikely(f)) {
				/* Probably -ENOMEM */
				folio->mapping = NULL;
				folio_put(folio);
			} else {
				folio->mapping = dmap;
				dmap->nrpages++;
				if (folio_test_dirty(folio))
					__xa_set_mark(&dmap->i_pages, index,
							PAGECACHE_TAG_DIRTY);
			}
			xa_unlock_irq(&dmap->i_pages);
		}
		folio_unlock(folio);
	}
	folio_batch_release(&fbatch);
	cond_resched();

	goto repeat;
}

/**
 * nilfs_clear_dirty_pages - discard dirty pages in address space
 * @mapping: address space with dirty pages for discarding
 * @silent: suppress [true] or print [false] warning messages
 */
void nilfs_clear_dirty_pages(struct address_space *mapping, bool silent)
{
	struct folio_batch fbatch;
	unsigned int i;
	pgoff_t index = 0;

	folio_batch_init(&fbatch);

	while (filemap_get_folios_tag(mapping, &index, (pgoff_t)-1,
				PAGECACHE_TAG_DIRTY, &fbatch)) {
		for (i = 0; i < folio_batch_count(&fbatch); i++) {
			struct folio *folio = fbatch.folios[i];

			folio_lock(folio);

			/*
			 * This folio may have been removed from the address
			 * space by truncation or invalidation when the lock
			 * was acquired.  Skip processing in that case.
			 */
			if (likely(folio->mapping == mapping))
				nilfs_clear_dirty_page(&folio->page, silent);

			folio_unlock(folio);
		}
		folio_batch_release(&fbatch);
		cond_resched();
	}
}

/**
 * nilfs_clear_dirty_page - discard dirty page
 * @page: dirty page that will be discarded
 * @silent: suppress [true] or print [false] warning messages
 */
void nilfs_clear_dirty_page(struct page *page, bool silent)
{
	struct inode *inode = page->mapping->host;
	struct super_block *sb = inode->i_sb;

	BUG_ON(!PageLocked(page));

	if (!silent)
		nilfs_warn(sb, "discard dirty page: offset=%lld, ino=%lu",
			   page_offset(page), inode->i_ino);

	ClearPageUptodate(page);
	ClearPageMappedToDisk(page);

	if (page_has_buffers(page)) {
		struct buffer_head *bh, *head;
		const unsigned long clear_bits =
			(BIT(BH_Uptodate) | BIT(BH_Dirty) | BIT(BH_Mapped) |
			 BIT(BH_Async_Write) | BIT(BH_NILFS_Volatile) |
			 BIT(BH_NILFS_Checked) | BIT(BH_NILFS_Redirected));

		bh = head = page_buffers(page);
		do {
			lock_buffer(bh);
			if (!silent)
				nilfs_warn(sb,
					   "discard dirty block: blocknr=%llu, size=%zu",
					   (u64)bh->b_blocknr, bh->b_size);

			set_mask_bits(&bh->b_state, clear_bits, 0);
			unlock_buffer(bh);
		} while (bh = bh->b_this_page, bh != head);
	}

	__nilfs_clear_page_dirty(page);
}

unsigned int nilfs_page_count_clean_buffers(struct page *page,
					    unsigned int from, unsigned int to)
{
	unsigned int block_start, block_end;
	struct buffer_head *bh, *head;
	unsigned int nc = 0;

	for (bh = head = page_buffers(page), block_start = 0;
	     bh != head || !block_start;
	     block_start = block_end, bh = bh->b_this_page) {
		block_end = block_start + bh->b_size;
		if (block_end > from && block_start < to && !buffer_dirty(bh))
			nc++;
	}
	return nc;
}

/*
 * NILFS2 needs clear_page_dirty() in the following two cases:
 *
 * 1) For B-tree node pages and data pages of DAT file, NILFS2 clears dirty
 *    flag of pages when it copies back pages from shadow cache to the
 *    original cache.
 *
 * 2) Some B-tree operations like insertion or deletion may dispose buffers
 *    in dirty state, and this needs to cancel the dirty state of their pages.
 */
int __nilfs_clear_page_dirty(struct page *page)
{
	struct address_space *mapping = page->mapping;

	if (mapping) {
		xa_lock_irq(&mapping->i_pages);
		if (test_bit(PG_dirty, &page->flags)) {
			__xa_clear_mark(&mapping->i_pages, page_index(page),
					     PAGECACHE_TAG_DIRTY);
			xa_unlock_irq(&mapping->i_pages);
			return clear_page_dirty_for_io(page);
		}
		xa_unlock_irq(&mapping->i_pages);
		return 0;
	}
	return TestClearPageDirty(page);
}

/**
 * nilfs_find_uncommitted_extent - find extent of uncommitted data
 * @inode: inode
 * @start_blk: start block offset (in)
 * @blkoff: start offset of the found extent (out)
 *
 * This function searches an extent of buffers marked "delayed" which
 * starts from a block offset equal to or larger than @start_blk.  If
 * such an extent was found, this will store the start offset in
 * @blkoff and return its length in blocks.  Otherwise, zero is
 * returned.
 */
unsigned long nilfs_find_uncommitted_extent(struct inode *inode,
					    sector_t start_blk,
					    sector_t *blkoff)
{
	unsigned int i, nr_folios;
	pgoff_t index;
	unsigned long length = 0;
	struct folio_batch fbatch;
	struct folio *folio;

	if (inode->i_mapping->nrpages == 0)
		return 0;

	index = start_blk >> (PAGE_SHIFT - inode->i_blkbits);

	folio_batch_init(&fbatch);

repeat:
	nr_folios = filemap_get_folios_contig(inode->i_mapping, &index, ULONG_MAX,
			&fbatch);
	if (nr_folios == 0)
		return length;

	i = 0;
	do {
		folio = fbatch.folios[i];

		folio_lock(folio);
		if (folio_buffers(folio)) {
			struct buffer_head *bh, *head;
			sector_t b;

			b = folio->index << (PAGE_SHIFT - inode->i_blkbits);
			bh = head = folio_buffers(folio);
			do {
				if (b < start_blk)
					continue;
				if (buffer_delay(bh)) {
					if (length == 0)
						*blkoff = b;
					length++;
				} else if (length > 0) {
					goto out_locked;
				}
			} while (++b, bh = bh->b_this_page, bh != head);
		} else {
			if (length > 0)
				goto out_locked;
		}
		folio_unlock(folio);

	} while (++i < nr_folios);

	folio_batch_release(&fbatch);
	cond_resched();
	goto repeat;

out_locked:
	folio_unlock(folio);
	folio_batch_release(&fbatch);
	return length;
}
Commit	Line	Data
ae98043f	1	// SPDX-License-Identifier: GPL-2.0+
0bd49f94	2	/*
94ee1d91	3	* Buffer/page management specific to NILFS
0bd49f94 RK	4	*
	5	* Copyright (C) 2005-2008 Nippon Telegraph and Telephone Corporation.
	6	*
4b420ab4	7	* Written by Ryusuke Konishi and Seiji Kihara.
0bd49f94 RK	8	*/
	9
	10	#include <linux/pagemap.h>
	11	#include <linux/writeback.h>
	12	#include <linux/swap.h>
	13	#include <linux/bitops.h>
	14	#include <linux/page-flags.h>
	15	#include <linux/list.h>
	16	#include <linux/highmem.h>
	17	#include <linux/pagevec.h>
5a0e3ad6	18	#include <linux/gfp.h>
0bd49f94 RK	19	#include "nilfs.h"
	20	#include "page.h"
	21	#include "mdt.h"
	22
	23
4ce5c342 RK	24	#define NILFS_BUFFER_INHERENT_BITS \
	25	(BIT(BH_Uptodate) \| BIT(BH_Mapped) \| BIT(BH_NILFS_Node) \| \
	26	BIT(BH_NILFS_Volatile) \| BIT(BH_NILFS_Checked))
0bd49f94	27
c5521c76 MWO	28	static struct buffer_head __nilfs_get_folio_block(struct folio folio,
	29	unsigned long block, pgoff_t index, int blkbits,
	30	unsigned long b_state)
0bd49f94 RK	31
	32	{
	33	unsigned long first_block;
c5521c76	34	struct buffer_head *bh = folio_buffers(folio);
0bd49f94	35
c5521c76	36	if (!bh)
0a88810d	37	bh = create_empty_buffers(folio, 1 << blkbits, b_state);
0bd49f94	38
09cbfeaf	39	first_block = (unsigned long)index << (PAGE_SHIFT - blkbits);
c5521c76	40	bh = get_nth_bh(bh, block - first_block);
0bd49f94 RK	41
	42	touch_buffer(bh);
	43	wait_on_buffer(bh);
	44	return bh;
	45	}
	46
0bd49f94 RK	47	struct buffer_head nilfs_grab_buffer(struct inode inode,
	48	struct address_space *mapping,
	49	unsigned long blkoff,
	50	unsigned long b_state)
	51	{
	52	int blkbits = inode->i_blkbits;
09cbfeaf	53	pgoff_t index = blkoff >> (PAGE_SHIFT - blkbits);
c5521c76	54	struct folio *folio;
c1c1d709	55	struct buffer_head *bh;
0bd49f94	56
c5521c76 MWO	57	folio = filemap_grab_folio(mapping, index);
c5521c76 MWO	58	if (IS_ERR(folio))
0bd49f94 RK	59	return NULL;
0bd49f94 RK	60
c5521c76	61	bh = __nilfs_get_folio_block(folio, blkoff, index, blkbits, b_state);
0bd49f94	62	if (unlikely(!bh)) {
c5521c76 MWO	63	folio_unlock(folio);
c5521c76 MWO	64	folio_put(folio);
0bd49f94 RK	65	return NULL;
0bd49f94 RK	66	}
0bd49f94 RK	67	return bh;
	68	}
	69
	70	/**
	71	* nilfs_forget_buffer - discard dirty state
0bd49f94 RK	72	* @bh: buffer head of the buffer to be discarded
	73	*/
	74	void nilfs_forget_buffer(struct buffer_head *bh)
	75	{
	76	struct page *page = bh->b_page;
ead8ecff	77	const unsigned long clear_bits =
4ce5c342 RK	78	(BIT(BH_Uptodate) \| BIT(BH_Dirty) \| BIT(BH_Mapped) \|
	79	BIT(BH_Async_Write) \| BIT(BH_NILFS_Volatile) \|
	80	BIT(BH_NILFS_Checked) \| BIT(BH_NILFS_Redirected));
0bd49f94 RK	81
0bd49f94 RK	82	lock_buffer(bh);
ead8ecff	83	set_mask_bits(&bh->b_state, clear_bits, 0);
84338237	84	if (nilfs_page_buffers_clean(page))
0bd49f94 RK	85	__nilfs_clear_page_dirty(page);
0bd49f94 RK	86
0bd49f94 RK	87	bh->b_blocknr = -1;
	88	ClearPageUptodate(page);
	89	ClearPageMappedToDisk(page);
	90	unlock_buffer(bh);
	91	brelse(bh);
	92	}
	93
	94	/**
	95	* nilfs_copy_buffer -- copy buffer data and flags
	96	* @dbh: destination buffer
	97	* @sbh: source buffer
	98	*/
	99	void nilfs_copy_buffer(struct buffer_head dbh, struct buffer_head sbh)
	100	{
	101	void kaddr0, kaddr1;
	102	unsigned long bits;
	103	struct page spage = sbh->b_page, dpage = dbh->b_page;
	104	struct buffer_head *bh;
	105
7b9c0976 CW	106	kaddr0 = kmap_atomic(spage);
7b9c0976 CW	107	kaddr1 = kmap_atomic(dpage);
0bd49f94	108	memcpy(kaddr1 + bh_offset(dbh), kaddr0 + bh_offset(sbh), sbh->b_size);
7b9c0976 CW	109	kunmap_atomic(kaddr1);
7b9c0976 CW	110	kunmap_atomic(kaddr0);
0bd49f94 RK	111
	112	dbh->b_state = sbh->b_state & NILFS_BUFFER_INHERENT_BITS;
	113	dbh->b_blocknr = sbh->b_blocknr;
	114	dbh->b_bdev = sbh->b_bdev;
	115
	116	bh = dbh;
4ce5c342	117	bits = sbh->b_state & (BIT(BH_Uptodate) \| BIT(BH_Mapped));
0bd49f94 RK	118	while ((bh = bh->b_this_page) != dbh) {
	119	lock_buffer(bh);
	120	bits &= bh->b_state;
	121	unlock_buffer(bh);
	122	}
4ce5c342	123	if (bits & BIT(BH_Uptodate))
0bd49f94 RK	124	SetPageUptodate(dpage);
	125	else
	126	ClearPageUptodate(dpage);
4ce5c342	127	if (bits & BIT(BH_Mapped))
0bd49f94 RK	128	SetPageMappedToDisk(dpage);
	129	else
	130	ClearPageMappedToDisk(dpage);
	131	}
	132
	133	/**
	134	* nilfs_page_buffers_clean - check if a page has dirty buffers or not.
	135	* @page: page to be checked
	136	*
	137	* nilfs_page_buffers_clean() returns zero if the page has dirty buffers.
	138	* Otherwise, it returns non-zero value.
	139	*/
	140	int nilfs_page_buffers_clean(struct page *page)
	141	{
	142	struct buffer_head bh, head;
	143
	144	bh = head = page_buffers(page);
	145	do {
	146	if (buffer_dirty(bh))
	147	return 0;
	148	bh = bh->b_this_page;
	149	} while (bh != head);
	150	return 1;
	151	}
	152
	153	void nilfs_page_bug(struct page *page)
	154	{
	155	struct address_space *m;
aa405b1f	156	unsigned long ino;
0bd49f94 RK	157
	158	if (unlikely(!page)) {
	159	printk(KERN_CRIT "NILFS_PAGE_BUG(NULL)\n");
	160	return;
	161	}
	162
	163	m = page->mapping;
aa405b1f RK	164	ino = m ? m->host->i_ino : 0;
aa405b1f RK	165
0bd49f94 RK	166	printk(KERN_CRIT "NILFS_PAGE_BUG(%p): cnt=%d index#=%llu flags=0x%lx "
0bd49f94 RK	167	"mapping=%p ino=%lu\n",
fe896d18	168	page, page_ref_count(page),
0bd49f94 RK	169	(unsigned long long)page->index, page->flags, m, ino);
	170
	171	if (page_has_buffers(page)) {
	172	struct buffer_head bh, head;
	173	int i = 0;
	174
	175	bh = head = page_buffers(page);
	176	do {
	177	printk(KERN_CRIT
	178	" BH[%d] %p: cnt=%d block#=%llu state=0x%lx\n",
	179	i++, bh, atomic_read(&bh->b_count),
	180	(unsigned long long)bh->b_blocknr, bh->b_state);
	181	bh = bh->b_this_page;
	182	} while (bh != head);
	183	}
	184	}
	185
0bd49f94	186	/**
4093602d MWO	187	* nilfs_copy_folio -- copy the folio with buffers
	188	* @dst: destination folio
	189	* @src: source folio
	190	* @copy_dirty: flag whether to copy dirty states on the folio's buffer heads.
0bd49f94	191	*
4093602d MWO	192	* This function is for both data folios and btnode folios. The dirty flag
	193	* should be treated by caller. The folio must not be under i/o.
	194	* Both src and dst folio must be locked
0bd49f94	195	*/
4093602d MWO	196	static void nilfs_copy_folio(struct folio dst, struct folio src,
4093602d MWO	197	bool copy_dirty)
0bd49f94	198	{
e1ce8a97	199	struct buffer_head dbh, dbufs, *sbh;
0bd49f94 RK	200	unsigned long mask = NILFS_BUFFER_INHERENT_BITS;
0bd49f94 RK	201
4093602d	202	BUG_ON(folio_test_writeback(dst));
0bd49f94	203
4093602d MWO	204	sbh = folio_buffers(src);
	205	dbh = folio_buffers(dst);
	206	if (!dbh)
0a88810d	207	dbh = create_empty_buffers(dst, sbh->b_size, 0);
0bd49f94 RK	208
0bd49f94 RK	209	if (copy_dirty)
4ce5c342	210	mask \|= BIT(BH_Dirty);
0bd49f94	211
4093602d	212	dbufs = dbh;
0bd49f94 RK	213	do {
	214	lock_buffer(sbh);
	215	lock_buffer(dbh);
	216	dbh->b_state = sbh->b_state & mask;
	217	dbh->b_blocknr = sbh->b_blocknr;
	218	dbh->b_bdev = sbh->b_bdev;
	219	sbh = sbh->b_this_page;
	220	dbh = dbh->b_this_page;
	221	} while (dbh != dbufs);
	222
4093602d	223	folio_copy(dst, src);
0bd49f94	224
4093602d MWO	225	if (folio_test_uptodate(src) && !folio_test_uptodate(dst))
	226	folio_mark_uptodate(dst);
	227	else if (!folio_test_uptodate(src) && folio_test_uptodate(dst))
	228	folio_clear_uptodate(dst);
	229	if (folio_test_mappedtodisk(src) && !folio_test_mappedtodisk(dst))
	230	folio_set_mappedtodisk(dst);
	231	else if (!folio_test_mappedtodisk(src) && folio_test_mappedtodisk(dst))
	232	folio_clear_mappedtodisk(dst);
0bd49f94 RK	233
	234	do {
	235	unlock_buffer(sbh);
	236	unlock_buffer(dbh);
	237	sbh = sbh->b_this_page;
	238	dbh = dbh->b_this_page;
	239	} while (dbh != dbufs);
	240	}
	241
	242	int nilfs_copy_dirty_pages(struct address_space *dmap,
	243	struct address_space *smap)
	244	{
d4a16d31	245	struct folio_batch fbatch;
0bd49f94 RK	246	unsigned int i;
	247	pgoff_t index = 0;
	248	int err = 0;
	249
d4a16d31	250	folio_batch_init(&fbatch);
0bd49f94	251	repeat:
d4a16d31 VMO	252	if (!filemap_get_folios_tag(smap, &index, (pgoff_t)-1,
d4a16d31 VMO	253	PAGECACHE_TAG_DIRTY, &fbatch))
0bd49f94 RK	254	return 0;
0bd49f94 RK	255
d4a16d31 VMO	256	for (i = 0; i < folio_batch_count(&fbatch); i++) {
d4a16d31 VMO	257	struct folio folio = fbatch.folios[i], dfolio;
0bd49f94	258
d4a16d31 VMO	259	folio_lock(folio);
	260	if (unlikely(!folio_test_dirty(folio)))
	261	NILFS_PAGE_BUG(&folio->page, "inconsistent dirty state");
0bd49f94	262
d4a16d31	263	dfolio = filemap_grab_folio(dmap, folio->index);
66dabbb6	264	if (unlikely(IS_ERR(dfolio))) {
0bd49f94	265	/* No empty page is added to the page cache */
d4a16d31	266	folio_unlock(folio);
66dabbb6	267	err = PTR_ERR(dfolio);
0bd49f94 RK	268	break;
0bd49f94 RK	269	}
d4a16d31 VMO	270	if (unlikely(!folio_buffers(folio)))
d4a16d31 VMO	271	NILFS_PAGE_BUG(&folio->page,
0bd49f94 RK	272	"found empty page in dat page cache");
0bd49f94 RK	273
4093602d	274	nilfs_copy_folio(dfolio, folio, true);
d4a16d31	275	filemap_dirty_folio(folio_mapping(dfolio), dfolio);
0bd49f94	276
d4a16d31 VMO	277	folio_unlock(dfolio);
	278	folio_put(dfolio);
	279	folio_unlock(folio);
0bd49f94	280	}
d4a16d31	281	folio_batch_release(&fbatch);
0bd49f94 RK	282	cond_resched();
	283
	284	if (likely(!err))
	285	goto repeat;
	286	return err;
	287	}
	288
	289	/**
7a65004b	290	* nilfs_copy_back_pages -- copy back pages to original cache from shadow cache
0bd49f94 RK	291	* @dmap: destination page cache
	292	* @smap: source page cache
	293	*
f611ff63	294	* No pages must be added to the cache during this process.
0bd49f94 RK	295	* This must be ensured by the caller.
	296	*/
	297	void nilfs_copy_back_pages(struct address_space *dmap,
	298	struct address_space *smap)
	299	{
f6e0e173	300	struct folio_batch fbatch;
0bd49f94	301	unsigned int i, n;
f6e0e173	302	pgoff_t start = 0;
0bd49f94	303
f6e0e173	304	folio_batch_init(&fbatch);
0bd49f94	305	repeat:
f6e0e173	306	n = filemap_get_folios(smap, &start, ~0UL, &fbatch);
0bd49f94 RK	307	if (!n)
0bd49f94 RK	308	return;
0bd49f94	309
f6e0e173 MWO	310	for (i = 0; i < folio_batch_count(&fbatch); i++) {
	311	struct folio folio = fbatch.folios[i], dfolio;
	312	pgoff_t index = folio->index;
	313
	314	folio_lock(folio);
	315	dfolio = filemap_lock_folio(dmap, index);
66dabbb6	316	if (!IS_ERR(dfolio)) {
f6e0e173 MWO	317	/* overwrite existing folio in the destination cache */
f6e0e173 MWO	318	WARN_ON(folio_test_dirty(dfolio));
4093602d	319	nilfs_copy_folio(dfolio, folio, false);
f6e0e173 MWO	320	folio_unlock(dfolio);
	321	folio_put(dfolio);
	322	/* Do we not need to remove folio from smap here? */
0bd49f94	323	} else {
f6e0e173	324	struct folio *f;
0bd49f94	325
f6e0e173	326	/* move the folio to the destination cache */
b93b0163	327	xa_lock_irq(&smap->i_pages);
f6e0e173 MWO	328	f = __xa_erase(&smap->i_pages, index);
f6e0e173 MWO	329	WARN_ON(folio != f);
0bd49f94	330	smap->nrpages--;
b93b0163	331	xa_unlock_irq(&smap->i_pages);
0bd49f94	332
b93b0163	333	xa_lock_irq(&dmap->i_pages);
f6e0e173 MWO	334	f = __xa_store(&dmap->i_pages, index, folio, GFP_NOFS);
f6e0e173 MWO	335	if (unlikely(f)) {
f611ff63	336	/* Probably -ENOMEM */
f6e0e173 MWO	337	folio->mapping = NULL;
f6e0e173 MWO	338	folio_put(folio);
0bd49f94	339	} else {
f6e0e173	340	folio->mapping = dmap;
0bd49f94	341	dmap->nrpages++;
f6e0e173 MWO	342	if (folio_test_dirty(folio))
f6e0e173 MWO	343	__xa_set_mark(&dmap->i_pages, index,
f611ff63	344	PAGECACHE_TAG_DIRTY);
0bd49f94	345	}
b93b0163	346	xa_unlock_irq(&dmap->i_pages);
0bd49f94	347	}
f6e0e173	348	folio_unlock(folio);
0bd49f94	349	}
f6e0e173	350	folio_batch_release(&fbatch);
0bd49f94 RK	351	cond_resched();
	352
	353	goto repeat;
	354	}
	355
8c26c4e2 VD	356	/**
	357	* nilfs_clear_dirty_pages - discard dirty pages in address space
	358	* @mapping: address space with dirty pages for discarding
	359	* @silent: suppress [true] or print [false] warning messages
	360	*/
	361	void nilfs_clear_dirty_pages(struct address_space *mapping, bool silent)
0bd49f94	362	{
243c5ea4	363	struct folio_batch fbatch;
0bd49f94 RK	364	unsigned int i;
	365	pgoff_t index = 0;
	366
243c5ea4	367	folio_batch_init(&fbatch);
0bd49f94	368
243c5ea4 VMO	369	while (filemap_get_folios_tag(mapping, &index, (pgoff_t)-1,
	370	PAGECACHE_TAG_DIRTY, &fbatch)) {
	371	for (i = 0; i < folio_batch_count(&fbatch); i++) {
	372	struct folio *folio = fbatch.folios[i];
0bd49f94	373
243c5ea4	374	folio_lock(folio);
782e53d0 RK	375
	376	/*
	377	* This folio may have been removed from the address
	378	* space by truncation or invalidation when the lock
	379	* was acquired. Skip processing in that case.
	380	*/
	381	if (likely(folio->mapping == mapping))
	382	nilfs_clear_dirty_page(&folio->page, silent);
	383
243c5ea4	384	folio_unlock(folio);
0bd49f94	385	}
243c5ea4	386	folio_batch_release(&fbatch);
0bd49f94 RK	387	cond_resched();
	388	}
	389	}
	390
8c26c4e2 VD	391	/**
	392	* nilfs_clear_dirty_page - discard dirty page
	393	* @page: dirty page that will be discarded
	394	* @silent: suppress [true] or print [false] warning messages
	395	*/
	396	void nilfs_clear_dirty_page(struct page *page, bool silent)
	397	{
	398	struct inode *inode = page->mapping->host;
	399	struct super_block *sb = inode->i_sb;
	400
dc33f5f3	401	BUG_ON(!PageLocked(page));
8c26c4e2	402
d6517deb	403	if (!silent)
a1d0747a JP	404	nilfs_warn(sb, "discard dirty page: offset=%lld, ino=%lu",
a1d0747a JP	405	page_offset(page), inode->i_ino);
8c26c4e2 VD	406
	407	ClearPageUptodate(page);
	408	ClearPageMappedToDisk(page);
	409
	410	if (page_has_buffers(page)) {
	411	struct buffer_head bh, head;
ead8ecff	412	const unsigned long clear_bits =
4ce5c342 RK	413	(BIT(BH_Uptodate) \| BIT(BH_Dirty) \| BIT(BH_Mapped) \|
	414	BIT(BH_Async_Write) \| BIT(BH_NILFS_Volatile) \|
	415	BIT(BH_NILFS_Checked) \| BIT(BH_NILFS_Redirected));
8c26c4e2 VD	416
	417	bh = head = page_buffers(page);
	418	do {
	419	lock_buffer(bh);
d6517deb	420	if (!silent)
a1d0747a JP	421	nilfs_warn(sb,
	422	"discard dirty block: blocknr=%llu, size=%zu",
	423	(u64)bh->b_blocknr, bh->b_size);
d6517deb	424
ead8ecff	425	set_mask_bits(&bh->b_state, clear_bits, 0);
8c26c4e2 VD	426	unlock_buffer(bh);
	427	} while (bh = bh->b_this_page, bh != head);
	428	}
	429
	430	__nilfs_clear_page_dirty(page);
	431	}
	432
0c6c44cb RK	433	unsigned int nilfs_page_count_clean_buffers(struct page *page,
0c6c44cb RK	434	unsigned int from, unsigned int to)
0bd49f94	435	{
0c6c44cb	436	unsigned int block_start, block_end;
0bd49f94	437	struct buffer_head bh, head;
0c6c44cb	438	unsigned int nc = 0;
0bd49f94 RK	439
	440	for (bh = head = page_buffers(page), block_start = 0;
	441	bh != head \|\| !block_start;
	442	block_start = block_end, bh = bh->b_this_page) {
	443	block_end = block_start + bh->b_size;
	444	if (block_end > from && block_start < to && !buffer_dirty(bh))
	445	nc++;
	446	}
	447	return nc;
	448	}
ae53a0a2	449
0bd49f94 RK	450	/*
	451	* NILFS2 needs clear_page_dirty() in the following two cases:
	452	*
e897be17 RK	453	* 1) For B-tree node pages and data pages of DAT file, NILFS2 clears dirty
	454	* flag of pages when it copies back pages from shadow cache to the
	455	* original cache.
0bd49f94 RK	456	*
	457	* 2) Some B-tree operations like insertion or deletion may dispose buffers
	458	* in dirty state, and this needs to cancel the dirty state of their pages.
	459	*/
	460	int __nilfs_clear_page_dirty(struct page *page)
	461	{
	462	struct address_space *mapping = page->mapping;
	463
	464	if (mapping) {
b93b0163	465	xa_lock_irq(&mapping->i_pages);
0bd49f94	466	if (test_bit(PG_dirty, &page->flags)) {
f611ff63	467	__xa_clear_mark(&mapping->i_pages, page_index(page),
0bd49f94	468	PAGECACHE_TAG_DIRTY);
b93b0163	469	xa_unlock_irq(&mapping->i_pages);
0bd49f94 RK	470	return clear_page_dirty_for_io(page);
0bd49f94 RK	471	}
b93b0163	472	xa_unlock_irq(&mapping->i_pages);
0bd49f94 RK	473	return 0;
	474	}
	475	return TestClearPageDirty(page);
	476	}
622daaff RK	477
	478	/**
	479	* nilfs_find_uncommitted_extent - find extent of uncommitted data
	480	* @inode: inode
	481	* @start_blk: start block offset (in)
	482	* @blkoff: start offset of the found extent (out)
	483	*
	484	* This function searches an extent of buffers marked "delayed" which
	485	* starts from a block offset equal to or larger than @start_blk. If
	486	* such an extent was found, this will store the start offset in
	487	* @blkoff and return its length in blocks. Otherwise, zero is
	488	* returned.
	489	*/
	490	unsigned long nilfs_find_uncommitted_extent(struct inode *inode,
	491	sector_t start_blk,
	492	sector_t *blkoff)
	493	{
24a1efb4	494	unsigned int i, nr_folios;
622daaff	495	pgoff_t index;
622daaff	496	unsigned long length = 0;
24a1efb4 VMO	497	struct folio_batch fbatch;
24a1efb4 VMO	498	struct folio *folio;
622daaff RK	499
	500	if (inode->i_mapping->nrpages == 0)
	501	return 0;
	502
09cbfeaf	503	index = start_blk >> (PAGE_SHIFT - inode->i_blkbits);
622daaff	504
24a1efb4	505	folio_batch_init(&fbatch);
622daaff RK	506
622daaff RK	507	repeat:
24a1efb4 VMO	508	nr_folios = filemap_get_folios_contig(inode->i_mapping, &index, ULONG_MAX,
	509	&fbatch);
	510	if (nr_folios == 0)
622daaff RK	511	return length;
622daaff RK	512
622daaff RK	513	i = 0;
622daaff RK	514	do {
24a1efb4	515	folio = fbatch.folios[i];
622daaff	516
24a1efb4 VMO	517	folio_lock(folio);
24a1efb4 VMO	518	if (folio_buffers(folio)) {
622daaff	519	struct buffer_head bh, head;
24a1efb4	520	sector_t b;
622daaff	521
24a1efb4 VMO	522	b = folio->index << (PAGE_SHIFT - inode->i_blkbits);
24a1efb4 VMO	523	bh = head = folio_buffers(folio);
622daaff RK	524	do {
	525	if (b < start_blk)
	526	continue;
	527	if (buffer_delay(bh)) {
	528	if (length == 0)
	529	*blkoff = b;
	530	length++;
	531	} else if (length > 0) {
	532	goto out_locked;
	533	}
	534	} while (++b, bh = bh->b_this_page, bh != head);
	535	} else {
	536	if (length > 0)
	537	goto out_locked;
622daaff	538	}
24a1efb4	539	folio_unlock(folio);
622daaff	540
24a1efb4	541	} while (++i < nr_folios);
622daaff	542
24a1efb4	543	folio_batch_release(&fbatch);
622daaff RK	544	cond_resched();
	545	goto repeat;
	546
	547	out_locked:
24a1efb4 VMO	548	folio_unlock(folio);
24a1efb4 VMO	549	folio_batch_release(&fbatch);
622daaff RK	550	return length;
622daaff RK	551	}