[PATCH v2 00/27] Finish the create_empty_buffers() transition

[PATCH v2 00/27] Finish the create_empty_buffers() transition

[Matthew Wilcox (Oracle)]

Pankaj recently added folio_create_empty_buffers() as the folio
equivalent to create_empty_buffers().  This patch set finishes
the conversion by first converting all remaining filesystems
to call folio_create_empty_buffers(), then renaming it back
to create_empty_buffers().  I took the opportunity to make a few
simplifications like making folio_create_empty_buffers() return the head
buffer and extracting get_nth_bh() from nilfs2.

A few of the patches in this series aren't directly related to
create_empty_buffers(), but I saw them while I was working on this and
thought they'd be easy enough to add to this series.  Compile-tested only,
other than ext4.

v2:
 - Added the patch to fix the return type from grow_dev_page()
 - Fixed typo in subject line noticed by Andreas
 - Fixed missed assignment to 'bh' spotted by Andreas
 - Exported folio_copy() spotted by Ryusuke
 - Added various Reviewed-by tags from Pankaj, Ryusuke & Andreas

Matthew Wilcox (Oracle) (27):
  buffer: Return bool from grow_dev_folio()
  buffer: Make folio_create_empty_buffers() return a buffer_head
  mpage: Convert map_buffer_to_folio() to folio_create_empty_buffers()
  ext4: Convert to folio_create_empty_buffers
  buffer: Add get_nth_bh()
  gfs2: Convert inode unstuffing to use a folio
  gfs2: Convert gfs2_getbuf() to folios
  gfs2: Convert gfs2_getjdatabuf to use a folio
  gfs2: Convert gfs2_write_buf_to_page() to use a folio
  nilfs2: Convert nilfs_mdt_freeze_buffer to use a folio
  nilfs2: Convert nilfs_grab_buffer() to use a folio
  nilfs2: Convert nilfs_copy_page() to nilfs_copy_folio()
  nilfs2: Convert nilfs_mdt_forget_block() to use a folio
  nilfs2: Convert nilfs_mdt_get_frozen_buffer to use a folio
  nilfs2: Remove nilfs_page_get_nth_block
  nilfs2: Convert nilfs_lookup_dirty_data_buffers to use
    folio_create_empty_buffers
  ntfs: Convert ntfs_read_block() to use a folio
  ntfs: Convert ntfs_writepage to use a folio
  ntfs: Convert ntfs_prepare_pages_for_non_resident_write() to folios
  ntfs3: Convert ntfs_zero_range() to use a folio
  ocfs2: Convert ocfs2_map_page_blocks to use a folio
  reiserfs: Convert writepage to use a folio
  ufs: Add ufs_get_locked_folio and ufs_put_locked_folio
  ufs: Use ufs_get_locked_folio() in ufs_alloc_lastblock()
  ufs; Convert ufs_change_blocknr() to use folios
  ufs: Remove ufs_get_locked_page()
  buffer: Remove folio_create_empty_buffers()

 fs/buffer.c                 |  68 +++++-----
 fs/ext4/inode.c             |  14 +-
 fs/ext4/move_extent.c       |  11 +-
 fs/gfs2/aops.c              |   2 +-
 fs/gfs2/bmap.c              |  48 ++++---
 fs/gfs2/meta_io.c           |  61 ++++-----
 fs/gfs2/quota.c             |  37 +++---
 fs/mpage.c                  |   3 +-
 fs/nilfs2/mdt.c             |  66 +++++-----
 fs/nilfs2/page.c            |  76 +++++------
 fs/nilfs2/page.h            |  11 --
 fs/nilfs2/segment.c         |   7 +-
 fs/ntfs/aops.c              | 255 +++++++++++++++++-------------------
 fs/ntfs/file.c              |  89 ++++++-------
 fs/ntfs3/file.c             |  31 ++---
 fs/ocfs2/aops.c             |  19 +--
 fs/reiserfs/inode.c         |  80 +++++------
 fs/ufs/balloc.c             |  20 ++-
 fs/ufs/inode.c              |  25 ++--
 fs/ufs/util.c               |  34 +++--
 fs/ufs/util.h               |  10 +-
 include/linux/buffer_head.h |  28 +++-
 mm/util.c                   |   1 +
 23 files changed, 481 insertions(+), 515 deletions(-)

-- 
2.40.1

[PATCH v2 01/27] buffer: Return bool from grow_dev_folio()

[Matthew Wilcox (Oracle)]

Rename grow_dev_page() to grow_dev_folio() and make it return a bool.
Document what that bool means; it's more subtle than it first appears.
Also rename the 'failed' label to 'unlock' beacuse it's not exactly
'failed'.  It just hasn't succeeded.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/buffer.c | 39 ++++++++++++++++++++++-----------------
 1 file changed, 22 insertions(+), 17 deletions(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index b33cc74e9649..dec41d84044b 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1024,24 +1024,26 @@ static sector_t folio_init_buffers(struct folio *folio,
 }
 
 /*
- * Create the page-cache page that contains the requested block.
+ * Create the page-cache folio that contains the requested block.
  *
  * This is used purely for blockdev mappings.
+ *
+ * Returns false if we have a 'permanent' failure.  Returns true if
+ * we succeeded, or the caller should retry.
  */
-static int
-grow_dev_page(struct block_device *bdev, sector_t block,
+static bool grow_dev_folio(struct block_device *bdev, sector_t block,
 	      pgoff_t index, int size, int sizebits, gfp_t gfp)
 {
 	struct inode *inode = bdev->bd_inode;
 	struct folio *folio;
 	struct buffer_head *bh;
 	sector_t end_block;
-	int ret = 0;
+	bool ret;
 
 	folio = __filemap_get_folio(inode->i_mapping, index,
 			FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
 	if (IS_ERR(folio))
-		return PTR_ERR(folio);
+		return false;
 
 	bh = folio_buffers(folio);
 	if (bh) {
@@ -1050,14 +1052,17 @@ grow_dev_page(struct block_device *bdev, sector_t block,
 					(sector_t)index << sizebits, size);
 			goto done;
 		}
+
+		/* Caller should retry if this call fails */
+		ret = true;
 		if (!try_to_free_buffers(folio))
-			goto failed;
+			goto unlock;
 	}
 
-	ret = -ENOMEM;
+	ret = false;
 	bh = folio_alloc_buffers(folio, size, gfp | __GFP_ACCOUNT);
 	if (!bh)
-		goto failed;
+		goto unlock;
 
 	/*
 	 * Link the folio to the buffers and initialise them.  Take the
@@ -1070,19 +1075,19 @@ grow_dev_page(struct block_device *bdev, sector_t block,
 			(sector_t)index << sizebits, size);
 	spin_unlock(&inode->i_mapping->private_lock);
 done:
-	ret = (block < end_block) ? 1 : -ENXIO;
-failed:
+	ret = block < end_block;
+unlock:
 	folio_unlock(folio);
 	folio_put(folio);
 	return ret;
 }
 
 /*
- * Create buffers for the specified block device block's page.  If
- * that page was dirty, the buffers are set dirty also.
+ * Create buffers for the specified block device block's folio.  If
+ * that folio was dirty, the buffers are set dirty also.
  */
-static int
-grow_buffers(struct block_device *bdev, sector_t block, int size, gfp_t gfp)
+static bool grow_buffers(struct block_device *bdev, sector_t block,
+		int size, gfp_t gfp)
 {
 	pgoff_t index;
 	int sizebits;
@@ -1099,11 +1104,11 @@ grow_buffers(struct block_device *bdev, sector_t block, int size, gfp_t gfp)
 			"device %pg\n",
 			__func__, (unsigned long long)block,
 			bdev);
-		return -EIO;
+		return false;
 	}
 
-	/* Create a page with the proper size buffers.. */
-	return grow_dev_page(bdev, block, index, size, sizebits, gfp);
+	/* Create a folio with the proper size buffers.. */
+	return grow_dev_folio(bdev, block, index, size, sizebits, gfp);
 }
 
 static struct buffer_head *
-- 
2.40.1

[PATCH v2 02/27] buffer: Make folio_create_empty_buffers() return a buffer_head

[Matthew Wilcox (Oracle)]

Almost all callers want to know the first BH that was allocated
for this folio.  We already have that handy, so return it.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/buffer.c                 | 24 +++++++++++++-----------
 include/linux/buffer_head.h |  4 ++--
 2 files changed, 15 insertions(+), 13 deletions(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index dec41d84044b..81cdf36e5196 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1646,8 +1646,8 @@ EXPORT_SYMBOL(block_invalidate_folio);
  * block_dirty_folio() via private_lock.  try_to_free_buffers
  * is already excluded via the folio lock.
  */
-void folio_create_empty_buffers(struct folio *folio, unsigned long blocksize,
-				unsigned long b_state)
+struct buffer_head *folio_create_empty_buffers(struct folio *folio,
+		unsigned long blocksize, unsigned long b_state)
 {
 	struct buffer_head *bh, *head, *tail;
 	gfp_t gfp = GFP_NOFS | __GFP_ACCOUNT | __GFP_NOFAIL;
@@ -1674,6 +1674,8 @@ void folio_create_empty_buffers(struct folio *folio, unsigned long blocksize,
 	}
 	folio_attach_private(folio, head);
 	spin_unlock(&folio->mapping->private_lock);
+
+	return head;
 }
 EXPORT_SYMBOL(folio_create_empty_buffers);
 
@@ -1775,13 +1777,15 @@ static struct buffer_head *folio_create_buffers(struct folio *folio,
 						struct inode *inode,
 						unsigned int b_state)
 {
+	struct buffer_head *bh;
+
 	BUG_ON(!folio_test_locked(folio));
 
-	if (!folio_buffers(folio))
-		folio_create_empty_buffers(folio,
-					   1 << READ_ONCE(inode->i_blkbits),
-					   b_state);
-	return folio_buffers(folio);
+	bh = folio_buffers(folio);
+	if (!bh)
+		bh = folio_create_empty_buffers(folio,
+				1 << READ_ONCE(inode->i_blkbits), b_state);
+	return bh;
 }
 
 /*
@@ -2681,10 +2685,8 @@ int block_truncate_page(struct address_space *mapping,
 		return PTR_ERR(folio);
 
 	bh = folio_buffers(folio);
-	if (!bh) {
-		folio_create_empty_buffers(folio, blocksize, 0);
-		bh = folio_buffers(folio);
-	}
+	if (!bh)
+		bh = folio_create_empty_buffers(folio, blocksize, 0);
 
 	/* Find the buffer that contains "offset" */
 	offset = offset_in_folio(folio, from);
diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index 3dc4720e4773..1001244a8941 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -203,8 +203,8 @@ struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size,
 		bool retry);
 void create_empty_buffers(struct page *, unsigned long,
 			unsigned long b_state);
-void folio_create_empty_buffers(struct folio *folio, unsigned long blocksize,
-				unsigned long b_state);
+struct buffer_head *folio_create_empty_buffers(struct folio *folio,
+		unsigned long blocksize, unsigned long b_state);
 void end_buffer_read_sync(struct buffer_head *bh, int uptodate);
 void end_buffer_write_sync(struct buffer_head *bh, int uptodate);
 void end_buffer_async_write(struct buffer_head *bh, int uptodate);
-- 
2.40.1

[PATCH v2 03/27] mpage: Convert map_buffer_to_folio() to folio_create_empty_buffers()

[Matthew Wilcox (Oracle)]

Saves a folio->page->folio conversion.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/mpage.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/fs/mpage.c b/fs/mpage.c
index 242e213ee064..964a6efe594d 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -119,8 +119,7 @@ static void map_buffer_to_folio(struct folio *folio, struct buffer_head *bh,
 			folio_mark_uptodate(folio);
 			return;
 		}
-		create_empty_buffers(&folio->page, i_blocksize(inode), 0);
-		head = folio_buffers(folio);
+		head = folio_create_empty_buffers(folio, i_blocksize(inode), 0);
 	}
 
 	page_bh = head;
-- 
2.40.1

[PATCH v2 04/27] ext4: Convert to folio_create_empty_buffers

[Matthew Wilcox (Oracle)]

Remove an unnecessary folio->page->folio conversion and take advantage
of the new return value from folio_create_empty_buffers().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Pankaj Raghav <p.raghav@samsung.com>
---
 fs/ext4/inode.c       | 14 +++++---------
 fs/ext4/move_extent.c | 11 +++++------
 2 files changed, 10 insertions(+), 15 deletions(-)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index a6838f54ae91..31c8f2641870 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1032,10 +1032,8 @@ static int ext4_block_write_begin(struct folio *folio, loff_t pos, unsigned len,
 	BUG_ON(from > to);
 
 	head = folio_buffers(folio);
-	if (!head) {
-		create_empty_buffers(&folio->page, blocksize, 0);
-		head = folio_buffers(folio);
-	}
+	if (!head)
+		head = folio_create_empty_buffers(folio, blocksize, 0);
 	bbits = ilog2(blocksize);
 	block = (sector_t)folio->index << (PAGE_SHIFT - bbits);
 
@@ -1165,7 +1163,7 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping,
 	 * starting the handle.
 	 */
 	if (!folio_buffers(folio))
-		create_empty_buffers(&folio->page, inode->i_sb->s_blocksize, 0);
+		folio_create_empty_buffers(folio, inode->i_sb->s_blocksize, 0);
 
 	folio_unlock(folio);
 
@@ -3655,10 +3653,8 @@ static int __ext4_block_zero_page_range(handle_t *handle,
 	iblock = index << (PAGE_SHIFT - inode->i_sb->s_blocksize_bits);
 
 	bh = folio_buffers(folio);
-	if (!bh) {
-		create_empty_buffers(&folio->page, blocksize, 0);
-		bh = folio_buffers(folio);
-	}
+	if (!bh)
+		bh = folio_create_empty_buffers(folio, blocksize, 0);
 
 	/* Find the buffer that contains "offset" */
 	pos = blocksize;
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 18a9e7c47975..7fe448fb948b 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -183,10 +183,8 @@ mext_page_mkuptodate(struct folio *folio, unsigned from, unsigned to)
 
 	blocksize = i_blocksize(inode);
 	head = folio_buffers(folio);
-	if (!head) {
-		create_empty_buffers(&folio->page, blocksize, 0);
-		head = folio_buffers(folio);
-	}
+	if (!head)
+		head = folio_create_empty_buffers(folio, blocksize, 0);
 
 	block = (sector_t)folio->index << (PAGE_SHIFT - inode->i_blkbits);
 	for (bh = head, block_start = 0; bh != head || !block_start;
@@ -380,9 +378,10 @@ move_extent_per_page(struct file *o_filp, struct inode *donor_inode,
 	}
 	/* Perform all necessary steps similar write_begin()/write_end()
 	 * but keeping in mind that i_size will not change */
-	if (!folio_buffers(folio[0]))
-		create_empty_buffers(&folio[0]->page, 1 << orig_inode->i_blkbits, 0);
 	bh = folio_buffers(folio[0]);
+	if (!bh)
+		bh = folio_create_empty_buffers(folio[0],
+				1 << orig_inode->i_blkbits, 0);
 	for (i = 0; i < data_offset_in_page; i++)
 		bh = bh->b_this_page;
 	for (i = 0; i < block_len_in_page; i++) {
-- 
2.40.1

[PATCH v2 05/27] buffer: Add get_nth_bh()

[Matthew Wilcox (Oracle)]

Extract this useful helper from nilfs_page_get_nth_block()

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/page.h            |  7 +------
 include/linux/buffer_head.h | 22 ++++++++++++++++++++++
 2 files changed, 23 insertions(+), 6 deletions(-)

diff --git a/fs/nilfs2/page.h b/fs/nilfs2/page.h
index 21ddcdd4d63e..344d71942d36 100644
--- a/fs/nilfs2/page.h
+++ b/fs/nilfs2/page.h
@@ -55,12 +55,7 @@ unsigned long nilfs_find_uncommitted_extent(struct inode *inode,
 static inline struct buffer_head *
 nilfs_page_get_nth_block(struct page *page, unsigned int count)
 {
-	struct buffer_head *bh = page_buffers(page);
-
-	while (count-- > 0)
-		bh = bh->b_this_page;
-	get_bh(bh);
-	return bh;
+	return get_nth_bh(page_buffers(page), count);
 }
 
 #endif /* _NILFS_PAGE_H */
diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index 1001244a8941..3d85a0cf0ca5 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -457,6 +457,28 @@ __bread(struct block_device *bdev, sector_t block, unsigned size)
 	return __bread_gfp(bdev, block, size, __GFP_MOVABLE);
 }
 
+/**
+ * get_nth_bh - Get a reference on the n'th buffer after this one.
+ * @bh: The buffer to start counting from.
+ * @count: How many buffers to skip.
+ *
+ * This is primarily useful for finding the nth buffer in a folio; in
+ * that case you pass the head buffer and the byte offset in the folio
+ * divided by the block size.  It can be used for other purposes, but
+ * it will wrap at the end of the folio rather than returning NULL or
+ * proceeding to the next folio for you.
+ *
+ * Return: The requested buffer with an elevated refcount.
+ */
+static inline __must_check
+struct buffer_head *get_nth_bh(struct buffer_head *bh, unsigned int count)
+{
+	while (count--)
+		bh = bh->b_this_page;
+	get_bh(bh);
+	return bh;
+}
+
 bool block_dirty_folio(struct address_space *mapping, struct folio *folio);
 
 #ifdef CONFIG_BUFFER_HEAD
-- 
2.40.1

[PATCH v2 06/27] gfs2: Convert inode unstuffing to use a folio

[Matthew Wilcox (Oracle)]

Use the folio APIs, removing numerous hidden calls to compound_head().
Also remove the stale comment about the page being looked up if it's NULL.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/bmap.c | 48 +++++++++++++++++++++++-------------------------
 1 file changed, 23 insertions(+), 25 deletions(-)

diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index f1d92e4d1538..f7ea1af9d842 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -43,53 +43,51 @@ struct metapath {
 static int punch_hole(struct gfs2_inode *ip, u64 offset, u64 length);
 
 /**
- * gfs2_unstuffer_page - unstuff a stuffed inode into a block cached by a page
+ * gfs2_unstuffer_folio - unstuff a stuffed inode into a block cached by a folio
  * @ip: the inode
  * @dibh: the dinode buffer
  * @block: the block number that was allocated
- * @page: The (optional) page. This is looked up if @page is NULL
+ * @folio: The folio.
  *
  * Returns: errno
  */
-
-static int gfs2_unstuffer_page(struct gfs2_inode *ip, struct buffer_head *dibh,
-			       u64 block, struct page *page)
+static int gfs2_unstuffer_folio(struct gfs2_inode *ip, struct buffer_head *dibh,
+			       u64 block, struct folio *folio)
 {
 	struct inode *inode = &ip->i_inode;
 
-	if (!PageUptodate(page)) {
-		void *kaddr = kmap(page);
+	if (!folio_test_uptodate(folio)) {
+		void *kaddr = kmap_local_folio(folio, 0);
 		u64 dsize = i_size_read(inode);
  
 		memcpy(kaddr, dibh->b_data + sizeof(struct gfs2_dinode), dsize);
-		memset(kaddr + dsize, 0, PAGE_SIZE - dsize);
-		kunmap(page);
+		memset(kaddr + dsize, 0, folio_size(folio) - dsize);
+		kunmap_local(kaddr);
 
-		SetPageUptodate(page);
+		folio_mark_uptodate(folio);
 	}
 
 	if (gfs2_is_jdata(ip)) {
-		struct buffer_head *bh;
+		struct buffer_head *bh = folio_buffers(folio);
 
-		if (!page_has_buffers(page))
-			create_empty_buffers(page, BIT(inode->i_blkbits),
-					     BIT(BH_Uptodate));
+		if (!bh)
+			bh = folio_create_empty_buffers(folio,
+				BIT(inode->i_blkbits), BIT(BH_Uptodate));
 
-		bh = page_buffers(page);
 		if (!buffer_mapped(bh))
 			map_bh(bh, inode->i_sb, block);
 
 		set_buffer_uptodate(bh);
 		gfs2_trans_add_data(ip->i_gl, bh);
 	} else {
-		set_page_dirty(page);
+		folio_mark_dirty(folio);
 		gfs2_ordered_add_inode(ip);
 	}
 
 	return 0;
 }
 
-static int __gfs2_unstuff_inode(struct gfs2_inode *ip, struct page *page)
+static int __gfs2_unstuff_inode(struct gfs2_inode *ip, struct folio *folio)
 {
 	struct buffer_head *bh, *dibh;
 	struct gfs2_dinode *di;
@@ -118,7 +116,7 @@ static int __gfs2_unstuff_inode(struct gfs2_inode *ip, struct page *page)
 					      dibh, sizeof(struct gfs2_dinode));
 			brelse(bh);
 		} else {
-			error = gfs2_unstuffer_page(ip, dibh, block, page);
+			error = gfs2_unstuffer_folio(ip, dibh, block, folio);
 			if (error)
 				goto out_brelse;
 		}
@@ -157,17 +155,17 @@ static int __gfs2_unstuff_inode(struct gfs2_inode *ip, struct page *page)
 int gfs2_unstuff_dinode(struct gfs2_inode *ip)
 {
 	struct inode *inode = &ip->i_inode;
-	struct page *page;
+	struct folio *folio;
 	int error;
 
 	down_write(&ip->i_rw_mutex);
-	page = grab_cache_page(inode->i_mapping, 0);
-	error = -ENOMEM;
-	if (!page)
+	folio = filemap_grab_folio(inode->i_mapping, 0);
+	error = PTR_ERR(folio);
+	if (IS_ERR(folio))
 		goto out;
-	error = __gfs2_unstuff_inode(ip, page);
-	unlock_page(page);
-	put_page(page);
+	error = __gfs2_unstuff_inode(ip, folio);
+	folio_unlock(folio);
+	folio_put(folio);
 out:
 	up_write(&ip->i_rw_mutex);
 	return error;
-- 
2.40.1

[PATCH v2 07/27] gfs2: Convert gfs2_getbuf() to folios

[Matthew Wilcox (Oracle)]

Remove several folio->page->folio conversions.  Also use __GFP_NOFAIL
instead of calling yield() and the new get_nth_bh().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/meta_io.c | 39 +++++++++++++++++----------------------
 1 file changed, 17 insertions(+), 22 deletions(-)

diff --git a/fs/gfs2/meta_io.c b/fs/gfs2/meta_io.c
index 924361fa510b..f1fac1b45059 100644
--- a/fs/gfs2/meta_io.c
+++ b/fs/gfs2/meta_io.c
@@ -115,7 +115,7 @@ struct buffer_head *gfs2_getbuf(struct gfs2_glock *gl, u64 blkno, int create)
 {
 	struct address_space *mapping = gfs2_glock2aspace(gl);
 	struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
-	struct page *page;
+	struct folio *folio;
 	struct buffer_head *bh;
 	unsigned int shift;
 	unsigned long index;
@@ -129,36 +129,31 @@ struct buffer_head *gfs2_getbuf(struct gfs2_glock *gl, u64 blkno, int create)
 	bufnum = blkno - (index << shift);  /* block buf index within page */
 
 	if (create) {
-		for (;;) {
-			page = grab_cache_page(mapping, index);
-			if (page)
-				break;
-			yield();
-		}
-		if (!page_has_buffers(page))
-			create_empty_buffers(page, sdp->sd_sb.sb_bsize, 0);
+		folio = __filemap_get_folio(mapping, index,
+				FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+				mapping_gfp_mask(mapping) | __GFP_NOFAIL);
+		bh = folio_buffers(folio);
+		if (!bh)
+			bh = folio_create_empty_buffers(folio,
+				sdp->sd_sb.sb_bsize, 0);
 	} else {
-		page = find_get_page_flags(mapping, index,
-						FGP_LOCK|FGP_ACCESSED);
-		if (!page)
+		folio = __filemap_get_folio(mapping, index,
+				FGP_LOCK | FGP_ACCESSED, 0);
+		if (IS_ERR(folio))
 			return NULL;
-		if (!page_has_buffers(page)) {
-			bh = NULL;
-			goto out_unlock;
-		}
+		bh = folio_buffers(folio);
 	}
 
-	/* Locate header for our buffer within our page */
-	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
-		/* Do nothing */;
-	get_bh(bh);
+	if (!bh)
+		goto out_unlock;
 
+	bh = get_nth_bh(bh, bufnum);
 	if (!buffer_mapped(bh))
 		map_bh(bh, sdp->sd_vfs, blkno);
 
 out_unlock:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	return bh;
 }
-- 
2.40.1

[PATCH v2 08/27] gfs2: Convert gfs2_getjdatabuf to use a folio

[Matthew Wilcox (Oracle)]

Use the folio APIs, saving four hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/meta_io.c | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/fs/gfs2/meta_io.c b/fs/gfs2/meta_io.c
index f1fac1b45059..f6d40d51f5ed 100644
--- a/fs/gfs2/meta_io.c
+++ b/fs/gfs2/meta_io.c
@@ -400,26 +400,20 @@ static struct buffer_head *gfs2_getjdatabuf(struct gfs2_inode *ip, u64 blkno)
 {
 	struct address_space *mapping = ip->i_inode.i_mapping;
 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
-	struct page *page;
+	struct folio *folio;
 	struct buffer_head *bh;
 	unsigned int shift = PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift;
 	unsigned long index = blkno >> shift; /* convert block to page */
 	unsigned int bufnum = blkno - (index << shift);
 
-	page = find_get_page_flags(mapping, index, FGP_LOCK|FGP_ACCESSED);
-	if (!page)
-		return NULL;
-	if (!page_has_buffers(page)) {
-		unlock_page(page);
-		put_page(page);
+	folio = __filemap_get_folio(mapping, index, FGP_LOCK | FGP_ACCESSED, 0);
+	if (IS_ERR(folio))
 		return NULL;
-	}
-	/* Locate header for our buffer within our page */
-	for (bh = page_buffers(page); bufnum--; bh = bh->b_this_page)
-		/* Do nothing */;
-	get_bh(bh);
-	unlock_page(page);
-	put_page(page);
+	bh = folio_buffers(folio);
+	if (bh)
+		bh = get_nth_bh(bh, bufnum);
+	folio_unlock(folio);
+	folio_put(folio);
 	return bh;
 }
 
-- 
2.40.1

[PATCH v2 09/27] gfs2: Convert gfs2_write_buf_to_page() to use a folio

[Matthew Wilcox (Oracle)]

Remove several folio->page->folio conversions.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/quota.c | 37 +++++++++++++++++++------------------
 1 file changed, 19 insertions(+), 18 deletions(-)

diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index 6affd261a754..e45654ee4f4d 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -760,7 +760,7 @@ static int gfs2_write_buf_to_page(struct gfs2_sbd *sdp, unsigned long index,
 	struct gfs2_inode *ip = GFS2_I(sdp->sd_quota_inode);
 	struct inode *inode = &ip->i_inode;
 	struct address_space *mapping = inode->i_mapping;
-	struct page *page;
+	struct folio *folio;
 	struct buffer_head *bh;
 	u64 blk;
 	unsigned bsize = sdp->sd_sb.sb_bsize, bnum = 0, boff = 0;
@@ -769,15 +769,15 @@ static int gfs2_write_buf_to_page(struct gfs2_sbd *sdp, unsigned long index,
 	blk = index << (PAGE_SHIFT - sdp->sd_sb.sb_bsize_shift);
 	boff = off % bsize;
 
-	page = grab_cache_page(mapping, index);
-	if (!page)
-		return -ENOMEM;
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, bsize, 0);
+	folio = filemap_grab_folio(mapping, index);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
+	bh = folio_buffers(folio);
+	if (!bh)
+		bh = folio_create_empty_buffers(folio, bsize, 0);
 
-	bh = page_buffers(page);
-	for(;;) {
-		/* Find the beginning block within the page */
+	for (;;) {
+		/* Find the beginning block within the folio */
 		if (pg_off >= ((bnum * bsize) + bsize)) {
 			bh = bh->b_this_page;
 			bnum++;
@@ -790,9 +790,10 @@ static int gfs2_write_buf_to_page(struct gfs2_sbd *sdp, unsigned long index,
 				goto unlock_out;
 			/* If it's a newly allocated disk block, zero it */
 			if (buffer_new(bh))
-				zero_user(page, bnum * bsize, bh->b_size);
+				folio_zero_range(folio, bnum * bsize,
+						bh->b_size);
 		}
-		if (PageUptodate(page))
+		if (folio_test_uptodate(folio))
 			set_buffer_uptodate(bh);
 		if (bh_read(bh, REQ_META | REQ_PRIO) < 0)
 			goto unlock_out;
@@ -808,17 +809,17 @@ static int gfs2_write_buf_to_page(struct gfs2_sbd *sdp, unsigned long index,
 		break;
 	}
 
-	/* Write to the page, now that we have setup the buffer(s) */
-	memcpy_to_page(page, off, buf, bytes);
-	flush_dcache_page(page);
-	unlock_page(page);
-	put_page(page);
+	/* Write to the folio, now that we have setup the buffer(s) */
+	memcpy_to_folio(folio, off, buf, bytes);
+	flush_dcache_folio(folio);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	return 0;
 
 unlock_out:
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 	return -EIO;
 }
 
-- 
2.40.1

[PATCH v2 10/27] nilfs2: Convert nilfs_mdt_freeze_buffer to use a folio

[Matthew Wilcox (Oracle)]

Remove a number of folio->page->folio conversions.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/mdt.c | 20 +++++++++++---------
 1 file changed, 11 insertions(+), 9 deletions(-)

diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 19c8158605ed..db2260d6e44d 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -560,17 +560,19 @@ int nilfs_mdt_freeze_buffer(struct inode *inode, struct buffer_head *bh)
 {
 	struct nilfs_shadow_map *shadow = NILFS_MDT(inode)->mi_shadow;
 	struct buffer_head *bh_frozen;
-	struct page *page;
+	struct folio *folio;
 	int blkbits = inode->i_blkbits;
 
-	page = grab_cache_page(shadow->inode->i_mapping, bh->b_folio->index);
-	if (!page)
-		return -ENOMEM;
+	folio = filemap_grab_folio(shadow->inode->i_mapping,
+			bh->b_folio->index);
+	if (IS_ERR(folio))
+		return PTR_ERR(folio);
 
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << blkbits, 0);
+	bh_frozen = folio_buffers(folio);
+	if (!bh_frozen)
+		bh_frozen = folio_create_empty_buffers(folio, 1 << blkbits, 0);
 
-	bh_frozen = nilfs_page_get_nth_block(page, bh_offset(bh) >> blkbits);
+	bh_frozen = get_nth_bh(bh_frozen, bh_offset(bh) >> blkbits);
 
 	if (!buffer_uptodate(bh_frozen))
 		nilfs_copy_buffer(bh_frozen, bh);
@@ -582,8 +584,8 @@ int nilfs_mdt_freeze_buffer(struct inode *inode, struct buffer_head *bh)
 		brelse(bh_frozen); /* already frozen */
 	}
 
-	unlock_page(page);
-	put_page(page);
+	folio_unlock(folio);
+	folio_put(folio);
 	return 0;
 }
 
-- 
2.40.1

[PATCH v2 11/27] nilfs2: Convert nilfs_grab_buffer() to use a folio

[Matthew Wilcox (Oracle)]

Remove a number of folio->page->folio conversions.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/page.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c
index b4e54d079b7d..1c075bd906c9 100644
--- a/fs/nilfs2/page.c
+++ b/fs/nilfs2/page.c
@@ -25,19 +25,19 @@
 	(BIT(BH_Uptodate) | BIT(BH_Mapped) | BIT(BH_NILFS_Node) |	\
 	 BIT(BH_NILFS_Volatile) | BIT(BH_NILFS_Checked))
 
-static struct buffer_head *
-__nilfs_get_page_block(struct page *page, unsigned long block, pgoff_t index,
-		       int blkbits, unsigned long b_state)
+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;
+	struct buffer_head *bh = folio_buffers(folio);
 
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << blkbits, b_state);
+	if (!bh)
+		bh = folio_create_empty_buffers(folio, 1 << blkbits, b_state);
 
 	first_block = (unsigned long)index << (PAGE_SHIFT - blkbits);
-	bh = nilfs_page_get_nth_block(page, block - first_block);
+	bh = get_nth_bh(bh, block - first_block);
 
 	touch_buffer(bh);
 	wait_on_buffer(bh);
@@ -51,17 +51,17 @@ struct buffer_head *nilfs_grab_buffer(struct inode *inode,
 {
 	int blkbits = inode->i_blkbits;
 	pgoff_t index = blkoff >> (PAGE_SHIFT - blkbits);
-	struct page *page;
+	struct folio *folio;
 	struct buffer_head *bh;
 
-	page = grab_cache_page(mapping, index);
-	if (unlikely(!page))
+	folio = filemap_grab_folio(mapping, index);
+	if (IS_ERR(folio))
 		return NULL;
 
-	bh = __nilfs_get_page_block(page, blkoff, index, blkbits, b_state);
+	bh = __nilfs_get_folio_block(folio, blkoff, index, blkbits, b_state);
 	if (unlikely(!bh)) {
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 		return NULL;
 	}
 	return bh;
-- 
2.40.1

[PATCH v2 12/27] nilfs2: Convert nilfs_copy_page() to nilfs_copy_folio()

[Matthew Wilcox (Oracle)]

Both callers already have a folio, so pass it in and use it directly.
Removes a lot of hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/page.c | 50 +++++++++++++++++++++++++-----------------------
 mm/util.c        |  1 +
 2 files changed, 27 insertions(+), 24 deletions(-)

diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c
index 1c075bd906c9..696215d899bf 100644
--- a/fs/nilfs2/page.c
+++ b/fs/nilfs2/page.c
@@ -184,30 +184,32 @@ void nilfs_page_bug(struct page *page)
 }
 
 /**
- * nilfs_copy_page -- copy the page with buffers
- * @dst: destination page
- * @src: source page
- * @copy_dirty: flag whether to copy dirty states on the page's buffer heads.
+ * 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 pages and btnode pages.  The dirty flag
- * should be treated by caller.  The page must not be under i/o.
- * Both src and dst page must be locked
+ * 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_page(struct page *dst, struct page *src, int copy_dirty)
+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(PageWriteback(dst));
+	BUG_ON(folio_test_writeback(dst));
 
-	sbh = page_buffers(src);
-	if (!page_has_buffers(dst))
-		create_empty_buffers(dst, sbh->b_size, 0);
+	sbh = folio_buffers(src);
+	dbh = folio_buffers(dst);
+	if (!dbh)
+		dbh = folio_create_empty_buffers(dst, sbh->b_size, 0);
 
 	if (copy_dirty)
 		mask |= BIT(BH_Dirty);
 
-	dbh = dbufs = page_buffers(dst);
+	dbufs = dbh;
 	do {
 		lock_buffer(sbh);
 		lock_buffer(dbh);
@@ -218,16 +220,16 @@ static void nilfs_copy_page(struct page *dst, struct page *src, int copy_dirty)
 		dbh = dbh->b_this_page;
 	} while (dbh != dbufs);
 
-	copy_highpage(dst, src);
+	folio_copy(dst, src);
 
-	if (PageUptodate(src) && !PageUptodate(dst))
-		SetPageUptodate(dst);
-	else if (!PageUptodate(src) && PageUptodate(dst))
-		ClearPageUptodate(dst);
-	if (PageMappedToDisk(src) && !PageMappedToDisk(dst))
-		SetPageMappedToDisk(dst);
-	else if (!PageMappedToDisk(src) && PageMappedToDisk(dst))
-		ClearPageMappedToDisk(dst);
+	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);
@@ -269,7 +271,7 @@ int nilfs_copy_dirty_pages(struct address_space *dmap,
 			NILFS_PAGE_BUG(&folio->page,
 				       "found empty page in dat page cache");
 
-		nilfs_copy_page(&dfolio->page, &folio->page, 1);
+		nilfs_copy_folio(dfolio, folio, true);
 		filemap_dirty_folio(folio_mapping(dfolio), dfolio);
 
 		folio_unlock(dfolio);
@@ -314,7 +316,7 @@ void nilfs_copy_back_pages(struct address_space *dmap,
 		if (!IS_ERR(dfolio)) {
 			/* overwrite existing folio in the destination cache */
 			WARN_ON(folio_test_dirty(dfolio));
-			nilfs_copy_page(&dfolio->page, &folio->page, 0);
+			nilfs_copy_folio(dfolio, folio, false);
 			folio_unlock(dfolio);
 			folio_put(dfolio);
 			/* Do we not need to remove folio from smap here? */
diff --git a/mm/util.c b/mm/util.c
index 6eddd891198e..aa01f6ea5a75 100644
--- a/mm/util.c
+++ b/mm/util.c
@@ -799,6 +799,7 @@ void folio_copy(struct folio *dst, struct folio *src)
 		cond_resched();
 	}
 }
+EXPORT_SYMBOL(folio_copy);
 
 int sysctl_overcommit_memory __read_mostly = OVERCOMMIT_GUESS;
 int sysctl_overcommit_ratio __read_mostly = 50;
-- 
2.40.1

[PATCH v2 13/27] nilfs2: Convert nilfs_mdt_forget_block() to use a folio

[Matthew Wilcox (Oracle)]

Remove a number of folio->page->folio conversions.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/mdt.c | 30 ++++++++++++++----------------
 1 file changed, 14 insertions(+), 16 deletions(-)

diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index db2260d6e44d..11b7cf4acc92 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -356,30 +356,28 @@ int nilfs_mdt_delete_block(struct inode *inode, unsigned long block)
  */
 int nilfs_mdt_forget_block(struct inode *inode, unsigned long block)
 {
-	pgoff_t index = (pgoff_t)block >>
-		(PAGE_SHIFT - inode->i_blkbits);
-	struct page *page;
-	unsigned long first_block;
+	pgoff_t index = block >> (PAGE_SHIFT - inode->i_blkbits);
+	struct folio *folio;
+	struct buffer_head *bh;
 	int ret = 0;
 	int still_dirty;
 
-	page = find_lock_page(inode->i_mapping, index);
-	if (!page)
+	folio = filemap_lock_folio(inode->i_mapping, index);
+	if (IS_ERR(folio))
 		return -ENOENT;
 
-	wait_on_page_writeback(page);
+	folio_wait_writeback(folio);
 
-	first_block = (unsigned long)index <<
-		(PAGE_SHIFT - inode->i_blkbits);
-	if (page_has_buffers(page)) {
-		struct buffer_head *bh;
-
-		bh = nilfs_page_get_nth_block(page, block - first_block);
+	bh = folio_buffers(folio);
+	if (bh) {
+		unsigned long first_block = index <<
+				(PAGE_SHIFT - inode->i_blkbits);
+		bh = get_nth_bh(bh, block - first_block);
 		nilfs_forget_buffer(bh);
 	}
-	still_dirty = PageDirty(page);
-	unlock_page(page);
-	put_page(page);
+	still_dirty = folio_test_dirty(folio);
+	folio_unlock(folio);
+	folio_put(folio);
 
 	if (still_dirty ||
 	    invalidate_inode_pages2_range(inode->i_mapping, index, index) != 0)
-- 
2.40.1

[PATCH v2 14/27] nilfs2: Convert nilfs_mdt_get_frozen_buffer to use a folio

[Matthew Wilcox (Oracle)]

Remove a number of folio->page->folio conversions.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/mdt.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 11b7cf4acc92..7b754e6494d7 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -592,17 +592,19 @@ nilfs_mdt_get_frozen_buffer(struct inode *inode, struct buffer_head *bh)
 {
 	struct nilfs_shadow_map *shadow = NILFS_MDT(inode)->mi_shadow;
 	struct buffer_head *bh_frozen = NULL;
-	struct page *page;
+	struct folio *folio;
 	int n;
 
-	page = find_lock_page(shadow->inode->i_mapping, bh->b_folio->index);
-	if (page) {
-		if (page_has_buffers(page)) {
+	folio = filemap_lock_folio(shadow->inode->i_mapping,
+			bh->b_folio->index);
+	if (!IS_ERR(folio)) {
+		bh_frozen = folio_buffers(folio);
+		if (bh_frozen) {
 			n = bh_offset(bh) >> inode->i_blkbits;
-			bh_frozen = nilfs_page_get_nth_block(page, n);
+			bh_frozen = get_nth_bh(bh_frozen, n);
 		}
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 	}
 	return bh_frozen;
 }
-- 
2.40.1

[PATCH v2 15/27] nilfs2: Remove nilfs_page_get_nth_block

[Matthew Wilcox (Oracle)]

All users have now been converted to get_nth_block().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/page.h | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/fs/nilfs2/page.h b/fs/nilfs2/page.h
index 344d71942d36..d249ea1cefff 100644
--- a/fs/nilfs2/page.h
+++ b/fs/nilfs2/page.h
@@ -52,10 +52,4 @@ unsigned long nilfs_find_uncommitted_extent(struct inode *inode,
 #define NILFS_PAGE_BUG(page, m, a...) \
 	do { nilfs_page_bug(page); BUG(); } while (0)
 
-static inline struct buffer_head *
-nilfs_page_get_nth_block(struct page *page, unsigned int count)
-{
-	return get_nth_bh(page_buffers(page), count);
-}
-
 #endif /* _NILFS_PAGE_H */
-- 
2.40.1

[PATCH v2 16/27] nilfs2: Convert nilfs_lookup_dirty_data_buffers to use folio_create_empty_buffers

[Matthew Wilcox (Oracle)]

This function was already using a folio, so this update to the new API
removes a single folio->page->folio conversion.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
---
 fs/nilfs2/segment.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 7ec16879756e..94388fe83cf8 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -731,10 +731,9 @@ static size_t nilfs_lookup_dirty_data_buffers(struct inode *inode,
 			continue;
 		}
 		head = folio_buffers(folio);
-		if (!head) {
-			create_empty_buffers(&folio->page, i_blocksize(inode), 0);
-			head = folio_buffers(folio);
-		}
+		if (!head)
+			head = folio_create_empty_buffers(folio,
+					i_blocksize(inode), 0);
 		folio_unlock(folio);
 
 		bh = head;
-- 
2.40.1

[PATCH v2 17/27] ntfs: Convert ntfs_read_block() to use a folio

[Matthew Wilcox (Oracle)]

The caller already has the folio, so pass it in and use the folio API
throughout saving five hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ntfs/aops.c | 44 +++++++++++++++++++-------------------------
 1 file changed, 19 insertions(+), 25 deletions(-)

diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index 4e158bce4192..d66a9f5ffde9 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -145,13 +145,12 @@ static void ntfs_end_buffer_async_read(struct buffer_head *bh, int uptodate)
 }
 
 /**
- * ntfs_read_block - fill a @page of an address space with data
- * @page:	page cache page to fill with data
+ * ntfs_read_block - fill a @folio of an address space with data
+ * @folio:	page cache folio to fill with data
  *
- * Fill the page @page of the address space belonging to the @page->host inode.
  * We read each buffer asynchronously and when all buffers are read in, our io
  * completion handler ntfs_end_buffer_read_async(), if required, automatically
- * applies the mst fixups to the page before finally marking it uptodate and
+ * applies the mst fixups to the folio before finally marking it uptodate and
  * unlocking it.
  *
  * We only enforce allocated_size limit because i_size is checked for in
@@ -161,7 +160,7 @@ static void ntfs_end_buffer_async_read(struct buffer_head *bh, int uptodate)
  *
  * Contains an adapted version of fs/buffer.c::block_read_full_folio().
  */
-static int ntfs_read_block(struct page *page)
+static int ntfs_read_block(struct folio *folio)
 {
 	loff_t i_size;
 	VCN vcn;
@@ -178,7 +177,7 @@ static int ntfs_read_block(struct page *page)
 	int i, nr;
 	unsigned char blocksize_bits;
 
-	vi = page->mapping->host;
+	vi = folio->mapping->host;
 	ni = NTFS_I(vi);
 	vol = ni->vol;
 
@@ -188,15 +187,10 @@ static int ntfs_read_block(struct page *page)
 	blocksize = vol->sb->s_blocksize;
 	blocksize_bits = vol->sb->s_blocksize_bits;
 
-	if (!page_has_buffers(page)) {
-		create_empty_buffers(page, blocksize, 0);
-		if (unlikely(!page_has_buffers(page))) {
-			unlock_page(page);
-			return -ENOMEM;
-		}
-	}
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
+	head = folio_buffers(folio);
+	if (!head)
+		head = folio_create_empty_buffers(folio, blocksize, 0);
+	bh = head;
 
 	/*
 	 * We may be racing with truncate.  To avoid some of the problems we
@@ -205,11 +199,11 @@ static int ntfs_read_block(struct page *page)
 	 * may leave some buffers unmapped which are now allocated.  This is
 	 * not a problem since these buffers will just get mapped when a write
 	 * occurs.  In case of a shrinking truncate, we will detect this later
-	 * on due to the runlist being incomplete and if the page is being
+	 * on due to the runlist being incomplete and if the folio is being
 	 * fully truncated, truncate will throw it away as soon as we unlock
 	 * it so no need to worry what we do with it.
 	 */
-	iblock = (s64)page->index << (PAGE_SHIFT - blocksize_bits);
+	iblock = (s64)folio->index << (PAGE_SHIFT - blocksize_bits);
 	read_lock_irqsave(&ni->size_lock, flags);
 	lblock = (ni->allocated_size + blocksize - 1) >> blocksize_bits;
 	init_size = ni->initialized_size;
@@ -221,7 +215,7 @@ static int ntfs_read_block(struct page *page)
 	}
 	zblock = (init_size + blocksize - 1) >> blocksize_bits;
 
-	/* Loop through all the buffers in the page. */
+	/* Loop through all the buffers in the folio. */
 	rl = NULL;
 	nr = i = 0;
 	do {
@@ -299,7 +293,7 @@ static int ntfs_read_block(struct page *page)
 			if (!err)
 				err = -EIO;
 			bh->b_blocknr = -1;
-			SetPageError(page);
+			folio_set_error(folio);
 			ntfs_error(vol->sb, "Failed to read from inode 0x%lx, "
 					"attribute type 0x%x, vcn 0x%llx, "
 					"offset 0x%x because its location on "
@@ -312,13 +306,13 @@ static int ntfs_read_block(struct page *page)
 		/*
 		 * Either iblock was outside lblock limits or
 		 * ntfs_rl_vcn_to_lcn() returned error.  Just zero that portion
-		 * of the page and set the buffer uptodate.
+		 * of the folio and set the buffer uptodate.
 		 */
 handle_hole:
 		bh->b_blocknr = -1UL;
 		clear_buffer_mapped(bh);
 handle_zblock:
-		zero_user(page, i * blocksize, blocksize);
+		folio_zero_range(folio, i * blocksize, blocksize);
 		if (likely(!err))
 			set_buffer_uptodate(bh);
 	} while (i++, iblock++, (bh = bh->b_this_page) != head);
@@ -349,11 +343,11 @@ static int ntfs_read_block(struct page *page)
 		return 0;
 	}
 	/* No i/o was scheduled on any of the buffers. */
-	if (likely(!PageError(page)))
-		SetPageUptodate(page);
+	if (likely(!folio_test_error(folio)))
+		folio_mark_uptodate(folio);
 	else /* Signal synchronous i/o error. */
 		nr = -EIO;
-	unlock_page(page);
+	folio_unlock(folio);
 	return nr;
 }
 
@@ -433,7 +427,7 @@ static int ntfs_read_folio(struct file *file, struct folio *folio)
 	/* NInoNonResident() == NInoIndexAllocPresent() */
 	if (NInoNonResident(ni)) {
 		/* Normal, non-resident data stream. */
-		return ntfs_read_block(page);
+		return ntfs_read_block(folio);
 	}
 	/*
 	 * Attribute is resident, implying it is not compressed or encrypted.
-- 
2.40.1

[PATCH v2 18/27] ntfs: Convert ntfs_writepage to use a folio

[Matthew Wilcox (Oracle)]

use folio APIs throughout.  Saves many hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ntfs/aops.c | 211 +++++++++++++++++++++++--------------------------
 1 file changed, 100 insertions(+), 111 deletions(-)

diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index d66a9f5ffde9..c4426992a2ee 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -501,28 +501,29 @@ static int ntfs_read_folio(struct file *file, struct folio *folio)
 #ifdef NTFS_RW
 
 /**
- * ntfs_write_block - write a @page to the backing store
- * @page:	page cache page to write out
+ * ntfs_write_block - write a @folio to the backing store
+ * @folio:	page cache folio to write out
  * @wbc:	writeback control structure
  *
- * This function is for writing pages belonging to non-resident, non-mst
+ * This function is for writing folios belonging to non-resident, non-mst
  * protected attributes to their backing store.
  *
- * For a page with buffers, map and write the dirty buffers asynchronously
- * under page writeback. For a page without buffers, create buffers for the
- * page, then proceed as above.
+ * For a folio with buffers, map and write the dirty buffers asynchronously
+ * under folio writeback. For a folio without buffers, create buffers for the
+ * folio, then proceed as above.
  *
- * If a page doesn't have buffers the page dirty state is definitive. If a page
- * does have buffers, the page dirty state is just a hint, and the buffer dirty
- * state is definitive. (A hint which has rules: dirty buffers against a clean
- * page is illegal. Other combinations are legal and need to be handled. In
- * particular a dirty page containing clean buffers for example.)
+ * If a folio doesn't have buffers the folio dirty state is definitive. If
+ * a folio does have buffers, the folio dirty state is just a hint,
+ * and the buffer dirty state is definitive. (A hint which has rules:
+ * dirty buffers against a clean folio is illegal. Other combinations are
+ * legal and need to be handled. In particular a dirty folio containing
+ * clean buffers for example.)
  *
  * Return 0 on success and -errno on error.
  *
  * Based on ntfs_read_block() and __block_write_full_folio().
  */
-static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
+static int ntfs_write_block(struct folio *folio, struct writeback_control *wbc)
 {
 	VCN vcn;
 	LCN lcn;
@@ -540,41 +541,29 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 	bool need_end_writeback;
 	unsigned char blocksize_bits;
 
-	vi = page->mapping->host;
+	vi = folio->mapping->host;
 	ni = NTFS_I(vi);
 	vol = ni->vol;
 
 	ntfs_debug("Entering for inode 0x%lx, attribute type 0x%x, page index "
-			"0x%lx.", ni->mft_no, ni->type, page->index);
+			"0x%lx.", ni->mft_no, ni->type, folio->index);
 
 	BUG_ON(!NInoNonResident(ni));
 	BUG_ON(NInoMstProtected(ni));
 	blocksize = vol->sb->s_blocksize;
 	blocksize_bits = vol->sb->s_blocksize_bits;
-	if (!page_has_buffers(page)) {
-		BUG_ON(!PageUptodate(page));
-		create_empty_buffers(page, blocksize,
+	head = folio_buffers(folio);
+	if (!head) {
+		BUG_ON(!folio_test_uptodate(folio));
+		head = folio_create_empty_buffers(folio, blocksize,
 				(1 << BH_Uptodate) | (1 << BH_Dirty));
-		if (unlikely(!page_has_buffers(page))) {
-			ntfs_warning(vol->sb, "Error allocating page "
-					"buffers.  Redirtying page so we try "
-					"again later.");
-			/*
-			 * Put the page back on mapping->dirty_pages, but leave
-			 * its buffers' dirty state as-is.
-			 */
-			redirty_page_for_writepage(wbc, page);
-			unlock_page(page);
-			return 0;
-		}
 	}
-	bh = head = page_buffers(page);
-	BUG_ON(!bh);
+	bh = head;
 
 	/* NOTE: Different naming scheme to ntfs_read_block()! */
 
-	/* The first block in the page. */
-	block = (s64)page->index << (PAGE_SHIFT - blocksize_bits);
+	/* The first block in the folio. */
+	block = (s64)folio->index << (PAGE_SHIFT - blocksize_bits);
 
 	read_lock_irqsave(&ni->size_lock, flags);
 	i_size = i_size_read(vi);
@@ -591,14 +580,14 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 	 * Be very careful.  We have no exclusion from block_dirty_folio
 	 * here, and the (potentially unmapped) buffers may become dirty at
 	 * any time.  If a buffer becomes dirty here after we've inspected it
-	 * then we just miss that fact, and the page stays dirty.
+	 * then we just miss that fact, and the folio stays dirty.
 	 *
 	 * Buffers outside i_size may be dirtied by block_dirty_folio;
 	 * handle that here by just cleaning them.
 	 */
 
 	/*
-	 * Loop through all the buffers in the page, mapping all the dirty
+	 * Loop through all the buffers in the folio, mapping all the dirty
 	 * buffers to disk addresses and handling any aliases from the
 	 * underlying block device's mapping.
 	 */
@@ -610,13 +599,13 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 		if (unlikely(block >= dblock)) {
 			/*
 			 * Mapped buffers outside i_size will occur, because
-			 * this page can be outside i_size when there is a
+			 * this folio can be outside i_size when there is a
 			 * truncate in progress. The contents of such buffers
 			 * were zeroed by ntfs_writepage().
 			 *
 			 * FIXME: What about the small race window where
 			 * ntfs_writepage() has not done any clearing because
-			 * the page was within i_size but before we get here,
+			 * the folio was within i_size but before we get here,
 			 * vmtruncate() modifies i_size?
 			 */
 			clear_buffer_dirty(bh);
@@ -632,38 +621,38 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 		if (unlikely((block >= iblock) &&
 				(initialized_size < i_size))) {
 			/*
-			 * If this page is fully outside initialized
-			 * size, zero out all pages between the current
-			 * initialized size and the current page. Just
+			 * If this folio is fully outside initialized
+			 * size, zero out all folios between the current
+			 * initialized size and the current folio. Just
 			 * use ntfs_read_folio() to do the zeroing
 			 * transparently.
 			 */
 			if (block > iblock) {
 				// TODO:
-				// For each page do:
-				// - read_cache_page()
-				// Again for each page do:
-				// - wait_on_page_locked()
-				// - Check (PageUptodate(page) &&
-				//			!PageError(page))
+				// For each folio do:
+				// - read_cache_folio()
+				// Again for each folio do:
+				// - wait_on_folio_locked()
+				// - Check (folio_test_uptodate(folio) &&
+				//		!folio_test_error(folio))
 				// Update initialized size in the attribute and
 				// in the inode.
-				// Again, for each page do:
+				// Again, for each folio do:
 				//	block_dirty_folio();
-				// put_page()
+				// folio_put()
 				// We don't need to wait on the writes.
 				// Update iblock.
 			}
 			/*
-			 * The current page straddles initialized size. Zero
+			 * The current folio straddles initialized size. Zero
 			 * all non-uptodate buffers and set them uptodate (and
 			 * dirty?). Note, there aren't any non-uptodate buffers
-			 * if the page is uptodate.
-			 * FIXME: For an uptodate page, the buffers may need to
+			 * if the folio is uptodate.
+			 * FIXME: For an uptodate folio, the buffers may need to
 			 * be written out because they were not initialized on
 			 * disk before.
 			 */
-			if (!PageUptodate(page)) {
+			if (!folio_test_uptodate(folio)) {
 				// TODO:
 				// Zero any non-uptodate buffers up to i_size.
 				// Set them uptodate and dirty.
@@ -721,14 +710,14 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 			unsigned long *bpos, *bend;
 
 			/* Check if the buffer is zero. */
-			kaddr = kmap_atomic(page);
-			bpos = (unsigned long *)(kaddr + bh_offset(bh));
-			bend = (unsigned long *)((u8*)bpos + blocksize);
+			kaddr = kmap_local_folio(folio, bh_offset(bh));
+			bpos = (unsigned long *)kaddr;
+			bend = (unsigned long *)(kaddr + blocksize);
 			do {
 				if (unlikely(*bpos))
 					break;
 			} while (likely(++bpos < bend));
-			kunmap_atomic(kaddr);
+			kunmap_local(kaddr);
 			if (bpos == bend) {
 				/*
 				 * Buffer is zero and sparse, no need to write
@@ -768,7 +757,7 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 		if (err == -ENOENT || lcn == LCN_ENOENT) {
 			bh->b_blocknr = -1;
 			clear_buffer_dirty(bh);
-			zero_user(page, bh_offset(bh), blocksize);
+			folio_zero_range(folio, bh_offset(bh), blocksize);
 			set_buffer_uptodate(bh);
 			err = 0;
 			continue;
@@ -795,7 +784,7 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 	bh = head;
 
 	/* Just an optimization, so ->read_folio() is not called later. */
-	if (unlikely(!PageUptodate(page))) {
+	if (unlikely(!folio_test_uptodate(folio))) {
 		int uptodate = 1;
 		do {
 			if (!buffer_uptodate(bh)) {
@@ -805,7 +794,7 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 			}
 		} while ((bh = bh->b_this_page) != head);
 		if (uptodate)
-			SetPageUptodate(page);
+			folio_mark_uptodate(folio);
 	}
 
 	/* Setup all mapped, dirty buffers for async write i/o. */
@@ -820,7 +809,7 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 		} else if (unlikely(err)) {
 			/*
 			 * For the error case. The buffer may have been set
-			 * dirty during attachment to a dirty page.
+			 * dirty during attachment to a dirty folio.
 			 */
 			if (err != -ENOMEM)
 				clear_buffer_dirty(bh);
@@ -833,20 +822,20 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 			err = 0;
 		else if (err == -ENOMEM) {
 			ntfs_warning(vol->sb, "Error allocating memory. "
-					"Redirtying page so we try again "
+					"Redirtying folio so we try again "
 					"later.");
 			/*
-			 * Put the page back on mapping->dirty_pages, but
+			 * Put the folio back on mapping->dirty_pages, but
 			 * leave its buffer's dirty state as-is.
 			 */
-			redirty_page_for_writepage(wbc, page);
+			folio_redirty_for_writepage(wbc, folio);
 			err = 0;
 		} else
-			SetPageError(page);
+			folio_set_error(folio);
 	}
 
-	BUG_ON(PageWriteback(page));
-	set_page_writeback(page);	/* Keeps try_to_free_buffers() away. */
+	BUG_ON(folio_test_writeback(folio));
+	folio_start_writeback(folio);	/* Keeps try_to_free_buffers() away. */
 
 	/* Submit the prepared buffers for i/o. */
 	need_end_writeback = true;
@@ -858,11 +847,11 @@ static int ntfs_write_block(struct page *page, struct writeback_control *wbc)
 		}
 		bh = next;
 	} while (bh != head);
-	unlock_page(page);
+	folio_unlock(folio);
 
-	/* If no i/o was started, need to end_page_writeback(). */
+	/* If no i/o was started, need to end writeback here. */
 	if (unlikely(need_end_writeback))
-		end_page_writeback(page);
+		folio_end_writeback(folio);
 
 	ntfs_debug("Done.");
 	return err;
@@ -1331,8 +1320,9 @@ static int ntfs_write_mst_block(struct page *page,
  */
 static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 {
+	struct folio *folio = page_folio(page);
 	loff_t i_size;
-	struct inode *vi = page->mapping->host;
+	struct inode *vi = folio->mapping->host;
 	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);
 	char *addr;
 	ntfs_attr_search_ctx *ctx = NULL;
@@ -1341,14 +1331,13 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 	int err;
 
 retry_writepage:
-	BUG_ON(!PageLocked(page));
+	BUG_ON(!folio_test_locked(folio));
 	i_size = i_size_read(vi);
-	/* Is the page fully outside i_size? (truncate in progress) */
-	if (unlikely(page->index >= (i_size + PAGE_SIZE - 1) >>
+	/* Is the folio fully outside i_size? (truncate in progress) */
+	if (unlikely(folio->index >= (i_size + PAGE_SIZE - 1) >>
 			PAGE_SHIFT)) {
-		struct folio *folio = page_folio(page);
 		/*
-		 * The page may have dirty, unmapped buffers.  Make them
+		 * The folio may have dirty, unmapped buffers.  Make them
 		 * freeable here, so the page does not leak.
 		 */
 		block_invalidate_folio(folio, 0, folio_size(folio));
@@ -1367,7 +1356,7 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 	if (ni->type != AT_INDEX_ALLOCATION) {
 		/* If file is encrypted, deny access, just like NT4. */
 		if (NInoEncrypted(ni)) {
-			unlock_page(page);
+			folio_unlock(folio);
 			BUG_ON(ni->type != AT_DATA);
 			ntfs_debug("Denying write access to encrypted file.");
 			return -EACCES;
@@ -1378,14 +1367,14 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 			BUG_ON(ni->name_len);
 			// TODO: Implement and replace this with
 			// return ntfs_write_compressed_block(page);
-			unlock_page(page);
+			folio_unlock(folio);
 			ntfs_error(vi->i_sb, "Writing to compressed files is "
 					"not supported yet.  Sorry.");
 			return -EOPNOTSUPP;
 		}
 		// TODO: Implement and remove this check.
 		if (NInoNonResident(ni) && NInoSparse(ni)) {
-			unlock_page(page);
+			folio_unlock(folio);
 			ntfs_error(vi->i_sb, "Writing to sparse files is not "
 					"supported yet.  Sorry.");
 			return -EOPNOTSUPP;
@@ -1394,34 +1383,34 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 	/* NInoNonResident() == NInoIndexAllocPresent() */
 	if (NInoNonResident(ni)) {
 		/* We have to zero every time due to mmap-at-end-of-file. */
-		if (page->index >= (i_size >> PAGE_SHIFT)) {
-			/* The page straddles i_size. */
-			unsigned int ofs = i_size & ~PAGE_MASK;
-			zero_user_segment(page, ofs, PAGE_SIZE);
+		if (folio->index >= (i_size >> PAGE_SHIFT)) {
+			/* The folio straddles i_size. */
+			unsigned int ofs = i_size & (folio_size(folio) - 1);
+			folio_zero_segment(folio, ofs, folio_size(folio));
 		}
 		/* Handle mst protected attributes. */
 		if (NInoMstProtected(ni))
 			return ntfs_write_mst_block(page, wbc);
 		/* Normal, non-resident data stream. */
-		return ntfs_write_block(page, wbc);
+		return ntfs_write_block(folio, wbc);
 	}
 	/*
 	 * Attribute is resident, implying it is not compressed, encrypted, or
 	 * mst protected.  This also means the attribute is smaller than an mft
-	 * record and hence smaller than a page, so can simply return error on
-	 * any pages with index above 0.  Note the attribute can actually be
+	 * record and hence smaller than a folio, so can simply return error on
+	 * any folios with index above 0.  Note the attribute can actually be
 	 * marked compressed but if it is resident the actual data is not
 	 * compressed so we are ok to ignore the compressed flag here.
 	 */
-	BUG_ON(page_has_buffers(page));
-	BUG_ON(!PageUptodate(page));
-	if (unlikely(page->index > 0)) {
-		ntfs_error(vi->i_sb, "BUG()! page->index (0x%lx) > 0.  "
-				"Aborting write.", page->index);
-		BUG_ON(PageWriteback(page));
-		set_page_writeback(page);
-		unlock_page(page);
-		end_page_writeback(page);
+	BUG_ON(folio_buffers(folio));
+	BUG_ON(!folio_test_uptodate(folio));
+	if (unlikely(folio->index > 0)) {
+		ntfs_error(vi->i_sb, "BUG()! folio->index (0x%lx) > 0.  "
+				"Aborting write.", folio->index);
+		BUG_ON(folio_test_writeback(folio));
+		folio_start_writeback(folio);
+		folio_unlock(folio);
+		folio_end_writeback(folio);
 		return -EIO;
 	}
 	if (!NInoAttr(ni))
@@ -1454,12 +1443,12 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 	if (unlikely(err))
 		goto err_out;
 	/*
-	 * Keep the VM happy.  This must be done otherwise the radix-tree tag
-	 * PAGECACHE_TAG_DIRTY remains set even though the page is clean.
+	 * Keep the VM happy.  This must be done otherwise
+	 * PAGECACHE_TAG_DIRTY remains set even though the folio is clean.
 	 */
-	BUG_ON(PageWriteback(page));
-	set_page_writeback(page);
-	unlock_page(page);
+	BUG_ON(folio_test_writeback(folio));
+	folio_start_writeback(folio);
+	folio_unlock(folio);
 	attr_len = le32_to_cpu(ctx->attr->data.resident.value_length);
 	i_size = i_size_read(vi);
 	if (unlikely(attr_len > i_size)) {
@@ -1474,18 +1463,18 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 		/* Shrinking cannot fail. */
 		BUG_ON(err);
 	}
-	addr = kmap_atomic(page);
-	/* Copy the data from the page to the mft record. */
+	addr = kmap_local_folio(folio, 0);
+	/* Copy the data from the folio to the mft record. */
 	memcpy((u8*)ctx->attr +
 			le16_to_cpu(ctx->attr->data.resident.value_offset),
 			addr, attr_len);
-	/* Zero out of bounds area in the page cache page. */
-	memset(addr + attr_len, 0, PAGE_SIZE - attr_len);
-	kunmap_atomic(addr);
-	flush_dcache_page(page);
+	/* Zero out of bounds area in the page cache folio. */
+	memset(addr + attr_len, 0, folio_size(folio) - attr_len);
+	kunmap_local(addr);
+	flush_dcache_folio(folio);
 	flush_dcache_mft_record_page(ctx->ntfs_ino);
-	/* We are done with the page. */
-	end_page_writeback(page);
+	/* We are done with the folio. */
+	folio_end_writeback(folio);
 	/* Finally, mark the mft record dirty, so it gets written back. */
 	mark_mft_record_dirty(ctx->ntfs_ino);
 	ntfs_attr_put_search_ctx(ctx);
@@ -1496,18 +1485,18 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc)
 		ntfs_warning(vi->i_sb, "Error allocating memory. Redirtying "
 				"page so we try again later.");
 		/*
-		 * Put the page back on mapping->dirty_pages, but leave its
+		 * Put the folio back on mapping->dirty_pages, but leave its
 		 * buffers' dirty state as-is.
 		 */
-		redirty_page_for_writepage(wbc, page);
+		folio_redirty_for_writepage(wbc, folio);
 		err = 0;
 	} else {
 		ntfs_error(vi->i_sb, "Resident attribute write failed with "
 				"error %i.", err);
-		SetPageError(page);
+		folio_set_error(folio);
 		NVolSetErrors(ni->vol);
 	}
-	unlock_page(page);
+	folio_unlock(folio);
 	if (ctx)
 		ntfs_attr_put_search_ctx(ctx);
 	if (m)
-- 
2.40.1

[PATCH v2 19/27] ntfs: Convert ntfs_prepare_pages_for_non_resident_write() to folios

[Matthew Wilcox (Oracle)]

Convert each element of the pages array to a folio before using it.
This in no way renders the function large-folio safe, but it does remove
a lot of hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ntfs/file.c | 89 +++++++++++++++++++++++---------------------------
 1 file changed, 41 insertions(+), 48 deletions(-)

diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index cbc545999cfe..099141d20db6 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -567,7 +567,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 	LCN lcn;
 	s64 bh_pos, vcn_len, end, initialized_size;
 	sector_t lcn_block;
-	struct page *page;
+	struct folio *folio;
 	struct inode *vi;
 	ntfs_inode *ni, *base_ni = NULL;
 	ntfs_volume *vol;
@@ -601,20 +601,6 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 			(long long)pos, bytes);
 	blocksize = vol->sb->s_blocksize;
 	blocksize_bits = vol->sb->s_blocksize_bits;
-	u = 0;
-	do {
-		page = pages[u];
-		BUG_ON(!page);
-		/*
-		 * create_empty_buffers() will create uptodate/dirty buffers if
-		 * the page is uptodate/dirty.
-		 */
-		if (!page_has_buffers(page)) {
-			create_empty_buffers(page, blocksize, 0);
-			if (unlikely(!page_has_buffers(page)))
-				return -ENOMEM;
-		}
-	} while (++u < nr_pages);
 	rl_write_locked = false;
 	rl = NULL;
 	err = 0;
@@ -626,14 +612,21 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 	end = pos + bytes;
 	cend = (end + vol->cluster_size - 1) >> vol->cluster_size_bits;
 	/*
-	 * Loop over each page and for each page over each buffer.  Use goto to
+	 * Loop over each buffer in each folio.  Use goto to
 	 * reduce indentation.
 	 */
 	u = 0;
-do_next_page:
-	page = pages[u];
-	bh_pos = (s64)page->index << PAGE_SHIFT;
-	bh = head = page_buffers(page);
+do_next_folio:
+	folio = page_folio(pages[u]);
+	bh_pos = folio_pos(folio);
+	head = folio_buffers(folio);
+	if (!head)
+		/*
+		 * create_empty_buffers() will create uptodate/dirty
+		 * buffers if the folio is uptodate/dirty.
+		 */
+		head = folio_create_empty_buffers(folio, blocksize, 0);
+	bh = head;
 	do {
 		VCN cdelta;
 		s64 bh_end;
@@ -653,15 +646,15 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 			if (buffer_uptodate(bh))
 				continue;
 			/*
-			 * The buffer is not uptodate.  If the page is uptodate
+			 * The buffer is not uptodate.  If the folio is uptodate
 			 * set the buffer uptodate and otherwise ignore it.
 			 */
-			if (PageUptodate(page)) {
+			if (folio_test_uptodate(folio)) {
 				set_buffer_uptodate(bh);
 				continue;
 			}
 			/*
-			 * Neither the page nor the buffer are uptodate.  If
+			 * Neither the folio nor the buffer are uptodate.  If
 			 * the buffer is only partially being written to, we
 			 * need to read it in before the write, i.e. now.
 			 */
@@ -679,7 +672,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 					ntfs_submit_bh_for_read(bh);
 					*wait_bh++ = bh;
 				} else {
-					zero_user(page, bh_offset(bh),
+					folio_zero_range(folio, bh_offset(bh),
 							blocksize);
 					set_buffer_uptodate(bh);
 				}
@@ -706,7 +699,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 					(bh_cofs >> blocksize_bits);
 			set_buffer_mapped(bh);
 			/*
-			 * If the page is uptodate so is the buffer.  If the
+			 * If the folio is uptodate so is the buffer.  If the
 			 * buffer is fully outside the write, we ignore it if
 			 * it was already allocated and we mark it dirty so it
 			 * gets written out if we allocated it.  On the other
@@ -714,7 +707,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 			 * marking it dirty we set buffer_new so we can do
 			 * error recovery.
 			 */
-			if (PageUptodate(page)) {
+			if (folio_test_uptodate(folio)) {
 				if (!buffer_uptodate(bh))
 					set_buffer_uptodate(bh);
 				if (unlikely(was_hole)) {
@@ -754,7 +747,8 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 						ntfs_submit_bh_for_read(bh);
 						*wait_bh++ = bh;
 					} else {
-						zero_user(page, bh_offset(bh),
+						folio_zero_range(folio,
+								bh_offset(bh),
 								blocksize);
 						set_buffer_uptodate(bh);
 					}
@@ -773,7 +767,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 			 */
 			if (bh_end <= pos || bh_pos >= end) {
 				if (!buffer_uptodate(bh)) {
-					zero_user(page, bh_offset(bh),
+					folio_zero_range(folio, bh_offset(bh),
 							blocksize);
 					set_buffer_uptodate(bh);
 				}
@@ -786,7 +780,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 				u8 *kaddr;
 				unsigned pofs;
 					
-				kaddr = kmap_atomic(page);
+				kaddr = kmap_local_folio(folio, 0);
 				if (bh_pos < pos) {
 					pofs = bh_pos & ~PAGE_MASK;
 					memset(kaddr + pofs, 0, pos - bh_pos);
@@ -795,8 +789,8 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 					pofs = end & ~PAGE_MASK;
 					memset(kaddr + pofs, 0, bh_end - end);
 				}
-				kunmap_atomic(kaddr);
-				flush_dcache_page(page);
+				kunmap_local(kaddr);
+				flush_dcache_folio(folio);
 			}
 			continue;
 		}
@@ -809,11 +803,12 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 		initialized_size = ni->allocated_size;
 		read_unlock_irqrestore(&ni->size_lock, flags);
 		if (bh_pos > initialized_size) {
-			if (PageUptodate(page)) {
+			if (folio_test_uptodate(folio)) {
 				if (!buffer_uptodate(bh))
 					set_buffer_uptodate(bh);
 			} else if (!buffer_uptodate(bh)) {
-				zero_user(page, bh_offset(bh), blocksize);
+				folio_zero_range(folio, bh_offset(bh),
+						blocksize);
 				set_buffer_uptodate(bh);
 			}
 			continue;
@@ -927,17 +922,17 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 				bh->b_blocknr = -1;
 				/*
 				 * If the buffer is uptodate we skip it.  If it
-				 * is not but the page is uptodate, we can set
-				 * the buffer uptodate.  If the page is not
+				 * is not but the folio is uptodate, we can set
+				 * the buffer uptodate.  If the folio is not
 				 * uptodate, we can clear the buffer and set it
 				 * uptodate.  Whether this is worthwhile is
 				 * debatable and this could be removed.
 				 */
-				if (PageUptodate(page)) {
+				if (folio_test_uptodate(folio)) {
 					if (!buffer_uptodate(bh))
 						set_buffer_uptodate(bh);
 				} else if (!buffer_uptodate(bh)) {
-					zero_user(page, bh_offset(bh),
+					folio_zero_range(folio, bh_offset(bh),
 						blocksize);
 					set_buffer_uptodate(bh);
 				}
@@ -1167,7 +1162,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 	} while (bh_pos += blocksize, (bh = bh->b_this_page) != head);
 	/* If there are no errors, do the next page. */
 	if (likely(!err && ++u < nr_pages))
-		goto do_next_page;
+		goto do_next_folio;
 	/* If there are no errors, release the runlist lock if we took it. */
 	if (likely(!err)) {
 		if (unlikely(rl_write_locked)) {
@@ -1185,9 +1180,8 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 		bh = *--wait_bh;
 		wait_on_buffer(bh);
 		if (likely(buffer_uptodate(bh))) {
-			page = bh->b_page;
-			bh_pos = ((s64)page->index << PAGE_SHIFT) +
-					bh_offset(bh);
+			folio = bh->b_folio;
+			bh_pos = folio_pos(folio) + bh_offset(bh);
 			/*
 			 * If the buffer overflows the initialized size, need
 			 * to zero the overflowing region.
@@ -1197,7 +1191,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 
 				if (likely(bh_pos < initialized_size))
 					ofs = initialized_size - bh_pos;
-				zero_user_segment(page, bh_offset(bh) + ofs,
+				folio_zero_segment(folio, bh_offset(bh) + ofs,
 						blocksize);
 			}
 		} else /* if (unlikely(!buffer_uptodate(bh))) */
@@ -1324,21 +1318,20 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 	u = 0;
 	end = bh_cpos << vol->cluster_size_bits;
 	do {
-		page = pages[u];
-		bh = head = page_buffers(page);
+		folio = page_folio(pages[u]);
+		bh = head = folio_buffers(folio);
 		do {
 			if (u == nr_pages &&
-					((s64)page->index << PAGE_SHIFT) +
-					bh_offset(bh) >= end)
+			    folio_pos(folio) + bh_offset(bh) >= end)
 				break;
 			if (!buffer_new(bh))
 				continue;
 			clear_buffer_new(bh);
 			if (!buffer_uptodate(bh)) {
-				if (PageUptodate(page))
+				if (folio_test_uptodate(folio))
 					set_buffer_uptodate(bh);
 				else {
-					zero_user(page, bh_offset(bh),
+					folio_zero_range(folio, bh_offset(bh),
 							blocksize);
 					set_buffer_uptodate(bh);
 				}
-- 
2.40.1

[PATCH v2 20/27] ntfs3: Convert ntfs_zero_range() to use a folio

[Matthew Wilcox (Oracle)]

Use the folio API throughout, saving six hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ntfs3/file.c | 31 ++++++++++++++++---------------
 1 file changed, 16 insertions(+), 15 deletions(-)

diff --git a/fs/ntfs3/file.c b/fs/ntfs3/file.c
index ad4a70b5d432..135af3340a19 100644
--- a/fs/ntfs3/file.c
+++ b/fs/ntfs3/file.c
@@ -187,7 +187,7 @@ static int ntfs_zero_range(struct inode *inode, u64 vbo, u64 vbo_to)
 	struct buffer_head *head, *bh;
 	u32 bh_next, bh_off, to;
 	sector_t iblock;
-	struct page *page;
+	struct folio *folio;
 
 	for (; idx < idx_end; idx += 1, from = 0) {
 		page_off = (loff_t)idx << PAGE_SHIFT;
@@ -195,16 +195,17 @@ static int ntfs_zero_range(struct inode *inode, u64 vbo, u64 vbo_to)
 						       PAGE_SIZE;
 		iblock = page_off >> inode->i_blkbits;
 
-		page = find_or_create_page(mapping, idx,
-					   mapping_gfp_constraint(mapping,
-								  ~__GFP_FS));
-		if (!page)
-			return -ENOMEM;
+		folio = __filemap_get_folio(mapping, idx,
+				FGP_LOCK | FGP_ACCESSED | FGP_CREAT,
+				mapping_gfp_constraint(mapping, ~__GFP_FS));
+		if (IS_ERR(folio))
+			return PTR_ERR(folio);
 
-		if (!page_has_buffers(page))
-			create_empty_buffers(page, blocksize, 0);
+		head = folio_buffers(folio);
+		if (!head)
+			head = folio_create_empty_buffers(folio, blocksize, 0);
 
-		bh = head = page_buffers(page);
+		bh = head;
 		bh_off = 0;
 		do {
 			bh_next = bh_off + blocksize;
@@ -220,14 +221,14 @@ static int ntfs_zero_range(struct inode *inode, u64 vbo, u64 vbo_to)
 			}
 
 			/* Ok, it's mapped. Make sure it's up-to-date. */
-			if (PageUptodate(page))
+			if (folio_test_uptodate(folio))
 				set_buffer_uptodate(bh);
 
 			if (!buffer_uptodate(bh)) {
 				err = bh_read(bh, 0);
 				if (err < 0) {
-					unlock_page(page);
-					put_page(page);
+					folio_unlock(folio);
+					folio_put(folio);
 					goto out;
 				}
 			}
@@ -237,10 +238,10 @@ static int ntfs_zero_range(struct inode *inode, u64 vbo, u64 vbo_to)
 		} while (bh_off = bh_next, iblock += 1,
 			 head != (bh = bh->b_this_page));
 
-		zero_user_segment(page, from, to);
+		folio_zero_segment(folio, from, to);
 
-		unlock_page(page);
-		put_page(page);
+		folio_unlock(folio);
+		folio_put(folio);
 		cond_resched();
 	}
 out:
-- 
2.40.1

[PATCH v2 21/27] ocfs2: Convert ocfs2_map_page_blocks to use a folio

[Matthew Wilcox (Oracle)]

Convert the page argument to a folio and then use the folio APIs
throughout.  Replaces three hidden calls to compound_head() with one
explicit one.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ocfs2/aops.c | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 6ab03494fc6e..001ad1d288ec 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -568,10 +568,10 @@ static void ocfs2_clear_page_regions(struct page *page,
  * read-in the blocks at the tail of our file. Avoid reading them by
  * testing i_size against each block offset.
  */
-static int ocfs2_should_read_blk(struct inode *inode, struct page *page,
+static int ocfs2_should_read_blk(struct inode *inode, struct folio *folio,
 				 unsigned int block_start)
 {
-	u64 offset = page_offset(page) + block_start;
+	u64 offset = folio_pos(folio) + block_start;
 
 	if (ocfs2_sparse_alloc(OCFS2_SB(inode->i_sb)))
 		return 1;
@@ -593,15 +593,16 @@ int ocfs2_map_page_blocks(struct page *page, u64 *p_blkno,
 			  struct inode *inode, unsigned int from,
 			  unsigned int to, int new)
 {
+	struct folio *folio = page_folio(page);
 	int ret = 0;
 	struct buffer_head *head, *bh, *wait[2], **wait_bh = wait;
 	unsigned int block_end, block_start;
 	unsigned int bsize = i_blocksize(inode);
 
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, bsize, 0);
+	head = folio_buffers(folio);
+	if (!head)
+		head = folio_create_empty_buffers(folio, bsize, 0);
 
-	head = page_buffers(page);
 	for (bh = head, block_start = 0; bh != head || !block_start;
 	     bh = bh->b_this_page, block_start += bsize) {
 		block_end = block_start + bsize;
@@ -613,7 +614,7 @@ int ocfs2_map_page_blocks(struct page *page, u64 *p_blkno,
 		 * they may belong to unallocated clusters.
 		 */
 		if (block_start >= to || block_end <= from) {
-			if (PageUptodate(page))
+			if (folio_test_uptodate(folio))
 				set_buffer_uptodate(bh);
 			continue;
 		}
@@ -630,11 +631,11 @@ int ocfs2_map_page_blocks(struct page *page, u64 *p_blkno,
 			clean_bdev_bh_alias(bh);
 		}
 
-		if (PageUptodate(page)) {
+		if (folio_test_uptodate(folio)) {
 			set_buffer_uptodate(bh);
 		} else if (!buffer_uptodate(bh) && !buffer_delay(bh) &&
 			   !buffer_new(bh) &&
-			   ocfs2_should_read_blk(inode, page, block_start) &&
+			   ocfs2_should_read_blk(inode, folio, block_start) &&
 			   (block_start < from || block_end > to)) {
 			bh_read_nowait(bh, 0);
 			*wait_bh++=bh;
@@ -668,7 +669,7 @@ int ocfs2_map_page_blocks(struct page *page, u64 *p_blkno,
 		if (block_start >= to)
 			break;
 
-		zero_user(page, block_start, bh->b_size);
+		folio_zero_range(folio, block_start, bh->b_size);
 		set_buffer_uptodate(bh);
 		mark_buffer_dirty(bh);
 
-- 
2.40.1

[PATCH v2 22/27] reiserfs: Convert writepage to use a folio

[Matthew Wilcox (Oracle)]

Convert the incoming page to a folio and then use it throughout the
writeback path.  This definitely isn't enough to support large folios, but
I don't expect reiserfs to gain support for those before it is removed.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/reiserfs/inode.c | 80 ++++++++++++++++++++++-----------------------
 1 file changed, 40 insertions(+), 40 deletions(-)

diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index c8572346556f..d7df556220a4 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -2503,10 +2503,10 @@ static int map_block_for_writepage(struct inode *inode,
  * start/recovery path as __block_write_full_folio, along with special
  * code to handle reiserfs tails.
  */
-static int reiserfs_write_full_page(struct page *page,
+static int reiserfs_write_full_folio(struct folio *folio,
 				    struct writeback_control *wbc)
 {
-	struct inode *inode = page->mapping->host;
+	struct inode *inode = folio->mapping->host;
 	unsigned long end_index = inode->i_size >> PAGE_SHIFT;
 	int error = 0;
 	unsigned long block;
@@ -2514,7 +2514,7 @@ static int reiserfs_write_full_page(struct page *page,
 	struct buffer_head *head, *bh;
 	int partial = 0;
 	int nr = 0;
-	int checked = PageChecked(page);
+	int checked = folio_test_checked(folio);
 	struct reiserfs_transaction_handle th;
 	struct super_block *s = inode->i_sb;
 	int bh_per_page = PAGE_SIZE / s->s_blocksize;
@@ -2522,47 +2522,46 @@ static int reiserfs_write_full_page(struct page *page,
 
 	/* no logging allowed when nonblocking or from PF_MEMALLOC */
 	if (checked && (current->flags & PF_MEMALLOC)) {
-		redirty_page_for_writepage(wbc, page);
-		unlock_page(page);
+		folio_redirty_for_writepage(wbc, folio);
+		folio_unlock(folio);
 		return 0;
 	}
 
 	/*
-	 * The page dirty bit is cleared before writepage is called, which
+	 * The folio dirty bit is cleared before writepage is called, which
 	 * means we have to tell create_empty_buffers to make dirty buffers
-	 * The page really should be up to date at this point, so tossing
+	 * The folio really should be up to date at this point, so tossing
 	 * in the BH_Uptodate is just a sanity check.
 	 */
-	if (!page_has_buffers(page)) {
-		create_empty_buffers(page, s->s_blocksize,
+	head = folio_buffers(folio);
+	if (!head)
+		head = folio_create_empty_buffers(folio, s->s_blocksize,
 				     (1 << BH_Dirty) | (1 << BH_Uptodate));
-	}
-	head = page_buffers(page);
 
 	/*
-	 * last page in the file, zero out any contents past the
+	 * last folio in the file, zero out any contents past the
 	 * last byte in the file
 	 */
-	if (page->index >= end_index) {
+	if (folio->index >= end_index) {
 		unsigned last_offset;
 
 		last_offset = inode->i_size & (PAGE_SIZE - 1);
-		/* no file contents in this page */
-		if (page->index >= end_index + 1 || !last_offset) {
-			unlock_page(page);
+		/* no file contents in this folio */
+		if (folio->index >= end_index + 1 || !last_offset) {
+			folio_unlock(folio);
 			return 0;
 		}
-		zero_user_segment(page, last_offset, PAGE_SIZE);
+		folio_zero_segment(folio, last_offset, folio_size(folio));
 	}
 	bh = head;
-	block = page->index << (PAGE_SHIFT - s->s_blocksize_bits);
+	block = folio->index << (PAGE_SHIFT - s->s_blocksize_bits);
 	last_block = (i_size_read(inode) - 1) >> inode->i_blkbits;
 	/* first map all the buffers, logging any direct items we find */
 	do {
 		if (block > last_block) {
 			/*
 			 * This can happen when the block size is less than
-			 * the page size.  The corresponding bytes in the page
+			 * the folio size.  The corresponding bytes in the folio
 			 * were zero filled above
 			 */
 			clear_buffer_dirty(bh);
@@ -2589,7 +2588,7 @@ static int reiserfs_write_full_page(struct page *page,
 	 * blocks we're going to log
 	 */
 	if (checked) {
-		ClearPageChecked(page);
+		folio_clear_checked(folio);
 		reiserfs_write_lock(s);
 		error = journal_begin(&th, s, bh_per_page + 1);
 		if (error) {
@@ -2598,7 +2597,7 @@ static int reiserfs_write_full_page(struct page *page,
 		}
 		reiserfs_update_inode_transaction(inode);
 	}
-	/* now go through and lock any dirty buffers on the page */
+	/* now go through and lock any dirty buffers on the folio */
 	do {
 		get_bh(bh);
 		if (!buffer_mapped(bh))
@@ -2619,7 +2618,7 @@ static int reiserfs_write_full_page(struct page *page,
 			lock_buffer(bh);
 		} else {
 			if (!trylock_buffer(bh)) {
-				redirty_page_for_writepage(wbc, page);
+				folio_redirty_for_writepage(wbc, folio);
 				continue;
 			}
 		}
@@ -2636,13 +2635,13 @@ static int reiserfs_write_full_page(struct page *page,
 		if (error)
 			goto fail;
 	}
-	BUG_ON(PageWriteback(page));
-	set_page_writeback(page);
-	unlock_page(page);
+	BUG_ON(folio_test_writeback(folio));
+	folio_start_writeback(folio);
+	folio_unlock(folio);
 
 	/*
-	 * since any buffer might be the only dirty buffer on the page,
-	 * the first submit_bh can bring the page out of writeback.
+	 * since any buffer might be the only dirty buffer on the folio,
+	 * the first submit_bh can bring the folio out of writeback.
 	 * be careful with the buffers.
 	 */
 	do {
@@ -2659,10 +2658,10 @@ static int reiserfs_write_full_page(struct page *page,
 done:
 	if (nr == 0) {
 		/*
-		 * if this page only had a direct item, it is very possible for
+		 * if this folio only had a direct item, it is very possible for
 		 * no io to be required without there being an error.  Or,
 		 * someone else could have locked them and sent them down the
-		 * pipe without locking the page
+		 * pipe without locking the folio
 		 */
 		bh = head;
 		do {
@@ -2673,18 +2672,18 @@ static int reiserfs_write_full_page(struct page *page,
 			bh = bh->b_this_page;
 		} while (bh != head);
 		if (!partial)
-			SetPageUptodate(page);
-		end_page_writeback(page);
+			folio_mark_uptodate(folio);
+		folio_end_writeback(folio);
 	}
 	return error;
 
 fail:
 	/*
 	 * catches various errors, we need to make sure any valid dirty blocks
-	 * get to the media.  The page is currently locked and not marked for
+	 * get to the media.  The folio is currently locked and not marked for
 	 * writeback
 	 */
-	ClearPageUptodate(page);
+	folio_clear_uptodate(folio);
 	bh = head;
 	do {
 		get_bh(bh);
@@ -2694,16 +2693,16 @@ static int reiserfs_write_full_page(struct page *page,
 		} else {
 			/*
 			 * clear any dirty bits that might have come from
-			 * getting attached to a dirty page
+			 * getting attached to a dirty folio
 			 */
 			clear_buffer_dirty(bh);
 		}
 		bh = bh->b_this_page;
 	} while (bh != head);
-	SetPageError(page);
-	BUG_ON(PageWriteback(page));
-	set_page_writeback(page);
-	unlock_page(page);
+	folio_set_error(folio);
+	BUG_ON(folio_test_writeback(folio));
+	folio_start_writeback(folio);
+	folio_unlock(folio);
 	do {
 		struct buffer_head *next = bh->b_this_page;
 		if (buffer_async_write(bh)) {
@@ -2724,9 +2723,10 @@ static int reiserfs_read_folio(struct file *f, struct folio *folio)
 
 static int reiserfs_writepage(struct page *page, struct writeback_control *wbc)
 {
-	struct inode *inode = page->mapping->host;
+	struct folio *folio = page_folio(page);
+	struct inode *inode = folio->mapping->host;
 	reiserfs_wait_on_write_block(inode->i_sb);
-	return reiserfs_write_full_page(page, wbc);
+	return reiserfs_write_full_folio(folio, wbc);
 }
 
 static void reiserfs_truncate_failed_write(struct inode *inode)
-- 
2.40.1

[PATCH v2 23/27] ufs: Add ufs_get_locked_folio and ufs_put_locked_folio

[Matthew Wilcox (Oracle)]

Convert the _page variants to call them.  Saves a few hidden calls to
compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ufs/util.c | 43 +++++++++++++++++++++++++------------------
 fs/ufs/util.h | 13 +++++++++----
 2 files changed, 34 insertions(+), 22 deletions(-)

diff --git a/fs/ufs/util.c b/fs/ufs/util.c
index 08ddf41eaaad..151b400cb3b6 100644
--- a/fs/ufs/util.c
+++ b/fs/ufs/util.c
@@ -229,43 +229,50 @@ ufs_set_inode_dev(struct super_block *sb, struct ufs_inode_info *ufsi, dev_t dev
 		ufsi->i_u1.i_data[0] = cpu_to_fs32(sb, fs32);
 }
 
+struct page *ufs_get_locked_page(struct address_space *mapping, pgoff_t index)
+{
+	struct folio *folio = ufs_get_locked_folio(mapping, index);
+
+	if (folio)
+		return folio_file_page(folio, index);
+	return NULL;
+}
+
 /**
- * ufs_get_locked_page() - locate, pin and lock a pagecache page, if not exist
+ * ufs_get_locked_folio() - locate, pin and lock a pagecache folio, if not exist
  * read it from disk.
  * @mapping: the address_space to search
  * @index: the page index
  *
- * Locates the desired pagecache page, if not exist we'll read it,
+ * Locates the desired pagecache folio, if not exist we'll read it,
  * locks it, increments its reference
  * count and returns its address.
  *
  */
-
-struct page *ufs_get_locked_page(struct address_space *mapping,
+struct folio *ufs_get_locked_folio(struct address_space *mapping,
 				 pgoff_t index)
 {
 	struct inode *inode = mapping->host;
-	struct page *page = find_lock_page(mapping, index);
-	if (!page) {
-		page = read_mapping_page(mapping, index, NULL);
+	struct folio *folio = filemap_lock_folio(mapping, index);
+	if (!folio) {
+		folio = read_mapping_folio(mapping, index, NULL);
 
-		if (IS_ERR(page)) {
-			printk(KERN_ERR "ufs_change_blocknr: "
-			       "read_mapping_page error: ino %lu, index: %lu\n",
+		if (IS_ERR(folio)) {
+			printk(KERN_ERR "ufs_change_blocknr: read_mapping_folio error: ino %lu, index: %lu\n",
 			       mapping->host->i_ino, index);
-			return page;
+			return folio;
 		}
 
-		lock_page(page);
+		folio_lock(folio);
 
-		if (unlikely(page->mapping == NULL)) {
+		if (unlikely(folio->mapping == NULL)) {
 			/* Truncate got there first */
-			unlock_page(page);
-			put_page(page);
+			folio_unlock(folio);
+			folio_put(folio);
 			return NULL;
 		}
 	}
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-	return page;
+	if (!folio_buffers(folio))
+		folio_create_empty_buffers(folio, 1 << inode->i_blkbits, 0);
+	return folio;
 }
diff --git a/fs/ufs/util.h b/fs/ufs/util.h
index 89247193d96d..62542561d150 100644
--- a/fs/ufs/util.h
+++ b/fs/ufs/util.h
@@ -273,12 +273,17 @@ extern void _ubh_ubhcpymem_(struct ufs_sb_private_info *, unsigned char *, struc
 extern void _ubh_memcpyubh_(struct ufs_sb_private_info *, struct ufs_buffer_head *, unsigned char *, unsigned);
 
 /* This functions works with cache pages*/
-extern struct page *ufs_get_locked_page(struct address_space *mapping,
-					pgoff_t index);
+struct page *ufs_get_locked_page(struct address_space *mapping, pgoff_t index);
+struct folio *ufs_get_locked_folio(struct address_space *mapping, pgoff_t index);
+static inline void ufs_put_locked_folio(struct folio *folio)
+{
+       folio_unlock(folio);
+       folio_put(folio);
+}
+
 static inline void ufs_put_locked_page(struct page *page)
 {
-       unlock_page(page);
-       put_page(page);
+	ufs_put_locked_folio(page_folio(page));
 }
 
 
-- 
2.40.1

[PATCH v2 24/27] ufs: Use ufs_get_locked_folio() in ufs_alloc_lastblock()

[Matthew Wilcox (Oracle)]

Switch to the folio APIs, saving one folio->page->folio conversion.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ufs/inode.c | 25 ++++++++++++-------------
 1 file changed, 12 insertions(+), 13 deletions(-)

diff --git a/fs/ufs/inode.c b/fs/ufs/inode.c
index 338e4b97312f..ebce93b08281 100644
--- a/fs/ufs/inode.c
+++ b/fs/ufs/inode.c
@@ -1063,7 +1063,7 @@ static int ufs_alloc_lastblock(struct inode *inode, loff_t size)
 	struct ufs_sb_private_info *uspi = UFS_SB(sb)->s_uspi;
 	unsigned i, end;
 	sector_t lastfrag;
-	struct page *lastpage;
+	struct folio *folio;
 	struct buffer_head *bh;
 	u64 phys64;
 
@@ -1074,18 +1074,17 @@ static int ufs_alloc_lastblock(struct inode *inode, loff_t size)
 
 	lastfrag--;
 
-	lastpage = ufs_get_locked_page(mapping, lastfrag >>
+	folio = ufs_get_locked_folio(mapping, lastfrag >>
 				       (PAGE_SHIFT - inode->i_blkbits));
-       if (IS_ERR(lastpage)) {
-               err = -EIO;
-               goto out;
-       }
-
-       end = lastfrag & ((1 << (PAGE_SHIFT - inode->i_blkbits)) - 1);
-       bh = page_buffers(lastpage);
-       for (i = 0; i < end; ++i)
-               bh = bh->b_this_page;
+	if (IS_ERR(folio)) {
+		err = -EIO;
+		goto out;
+	}
 
+	end = lastfrag & ((1 << (PAGE_SHIFT - inode->i_blkbits)) - 1);
+	bh = folio_buffers(folio);
+	for (i = 0; i < end; ++i)
+		bh = bh->b_this_page;
 
        err = ufs_getfrag_block(inode, lastfrag, bh, 1);
 
@@ -1101,7 +1100,7 @@ static int ufs_alloc_lastblock(struct inode *inode, loff_t size)
 		*/
 	       set_buffer_uptodate(bh);
 	       mark_buffer_dirty(bh);
-	       set_page_dirty(lastpage);
+		folio_mark_dirty(folio);
        }
 
        if (lastfrag >= UFS_IND_FRAGMENT) {
@@ -1119,7 +1118,7 @@ static int ufs_alloc_lastblock(struct inode *inode, loff_t size)
 	       }
        }
 out_unlock:
-       ufs_put_locked_page(lastpage);
+       ufs_put_locked_folio(folio);
 out:
        return err;
 }
-- 
2.40.1

[PATCH v2 25/27] ufs; Convert ufs_change_blocknr() to use folios

[Matthew Wilcox (Oracle)]

Convert the locked_page argument to a folio, then use folios throughout.
Saves three hidden calls to compound_head().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ufs/balloc.c | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/fs/ufs/balloc.c b/fs/ufs/balloc.c
index 2436e3f82147..53c11be2b2c1 100644
--- a/fs/ufs/balloc.c
+++ b/fs/ufs/balloc.c
@@ -240,6 +240,7 @@ static void ufs_change_blocknr(struct inode *inode, sector_t beg,
 			       unsigned int count, sector_t oldb,
 			       sector_t newb, struct page *locked_page)
 {
+	struct folio *folio, *locked_folio = page_folio(locked_page);
 	const unsigned blks_per_page =
 		1 << (PAGE_SHIFT - inode->i_blkbits);
 	const unsigned mask = blks_per_page - 1;
@@ -247,42 +248,39 @@ static void ufs_change_blocknr(struct inode *inode, sector_t beg,
 	pgoff_t index, cur_index, last_index;
 	unsigned pos, j, lblock;
 	sector_t end, i;
-	struct page *page;
 	struct buffer_head *head, *bh;
 
 	UFSD("ENTER, ino %lu, count %u, oldb %llu, newb %llu\n",
 	      inode->i_ino, count,
 	     (unsigned long long)oldb, (unsigned long long)newb);
 
-	BUG_ON(!locked_page);
-	BUG_ON(!PageLocked(locked_page));
+	BUG_ON(!folio_test_locked(locked_folio));
 
-	cur_index = locked_page->index;
+	cur_index = locked_folio->index;
 	end = count + beg;
 	last_index = end >> (PAGE_SHIFT - inode->i_blkbits);
 	for (i = beg; i < end; i = (i | mask) + 1) {
 		index = i >> (PAGE_SHIFT - inode->i_blkbits);
 
 		if (likely(cur_index != index)) {
-			page = ufs_get_locked_page(mapping, index);
-			if (!page)/* it was truncated */
+			folio = ufs_get_locked_folio(mapping, index);
+			if (!folio) /* it was truncated */
 				continue;
-			if (IS_ERR(page)) {/* or EIO */
+			if (IS_ERR(folio)) {/* or EIO */
 				ufs_error(inode->i_sb, __func__,
 					  "read of page %llu failed\n",
 					  (unsigned long long)index);
 				continue;
 			}
 		} else
-			page = locked_page;
+			folio = locked_folio;
 
-		head = page_buffers(page);
+		head = folio_buffers(folio);
 		bh = head;
 		pos = i & mask;
 		for (j = 0; j < pos; ++j)
 			bh = bh->b_this_page;
 
-
 		if (unlikely(index == last_index))
 			lblock = end & mask;
 		else
@@ -313,7 +311,7 @@ static void ufs_change_blocknr(struct inode *inode, sector_t beg,
 		} while (bh != head);
 
 		if (likely(cur_index != index))
-			ufs_put_locked_page(page);
+			ufs_put_locked_folio(folio);
  	}
 	UFSD("EXIT\n");
 }
-- 
2.40.1

[PATCH v2 26/27] ufs: Remove ufs_get_locked_page()

[Matthew Wilcox (Oracle)]

Both callers are now converted to ufs_get_locked_folio().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/ufs/util.c | 9 ---------
 fs/ufs/util.h | 7 -------
 2 files changed, 16 deletions(-)

diff --git a/fs/ufs/util.c b/fs/ufs/util.c
index 151b400cb3b6..d32de30009a0 100644
--- a/fs/ufs/util.c
+++ b/fs/ufs/util.c
@@ -229,15 +229,6 @@ ufs_set_inode_dev(struct super_block *sb, struct ufs_inode_info *ufsi, dev_t dev
 		ufsi->i_u1.i_data[0] = cpu_to_fs32(sb, fs32);
 }
 
-struct page *ufs_get_locked_page(struct address_space *mapping, pgoff_t index)
-{
-	struct folio *folio = ufs_get_locked_folio(mapping, index);
-
-	if (folio)
-		return folio_file_page(folio, index);
-	return NULL;
-}
-
 /**
  * ufs_get_locked_folio() - locate, pin and lock a pagecache folio, if not exist
  * read it from disk.
diff --git a/fs/ufs/util.h b/fs/ufs/util.h
index 62542561d150..0ecd2ed792f5 100644
--- a/fs/ufs/util.h
+++ b/fs/ufs/util.h
@@ -273,7 +273,6 @@ extern void _ubh_ubhcpymem_(struct ufs_sb_private_info *, unsigned char *, struc
 extern void _ubh_memcpyubh_(struct ufs_sb_private_info *, struct ufs_buffer_head *, unsigned char *, unsigned);
 
 /* This functions works with cache pages*/
-struct page *ufs_get_locked_page(struct address_space *mapping, pgoff_t index);
 struct folio *ufs_get_locked_folio(struct address_space *mapping, pgoff_t index);
 static inline void ufs_put_locked_folio(struct folio *folio)
 {
@@ -281,12 +280,6 @@ static inline void ufs_put_locked_folio(struct folio *folio)
        folio_put(folio);
 }
 
-static inline void ufs_put_locked_page(struct page *page)
-{
-	ufs_put_locked_folio(page_folio(page));
-}
-
-
 /*
  * macros and inline function to get important structures from ufs_sb_private_info
  */
-- 
2.40.1

[PATCH v2 27/27] buffer: Remove folio_create_empty_buffers()

[Matthew Wilcox (Oracle)]

With all users converted, remove the old create_empty_buffers() and
rename folio_create_empty_buffers() to create_empty_buffers().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/buffer.c                 | 13 +++----------
 fs/ext4/inode.c             |  6 +++---
 fs/ext4/move_extent.c       |  4 ++--
 fs/gfs2/aops.c              |  2 +-
 fs/gfs2/bmap.c              |  2 +-
 fs/gfs2/meta_io.c           |  2 +-
 fs/gfs2/quota.c             |  2 +-
 fs/mpage.c                  |  2 +-
 fs/nilfs2/mdt.c             |  2 +-
 fs/nilfs2/page.c            |  4 ++--
 fs/nilfs2/segment.c         |  2 +-
 fs/ntfs/aops.c              |  4 ++--
 fs/ntfs/file.c              |  2 +-
 fs/ntfs3/file.c             |  2 +-
 fs/ocfs2/aops.c             |  2 +-
 fs/reiserfs/inode.c         |  2 +-
 fs/ufs/util.c               |  2 +-
 include/linux/buffer_head.h |  4 +---
 18 files changed, 25 insertions(+), 34 deletions(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index 81cdf36e5196..bf76b7a27bfa 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -1646,7 +1646,7 @@ EXPORT_SYMBOL(block_invalidate_folio);
  * block_dirty_folio() via private_lock.  try_to_free_buffers
  * is already excluded via the folio lock.
  */
-struct buffer_head *folio_create_empty_buffers(struct folio *folio,
+struct buffer_head *create_empty_buffers(struct folio *folio,
 		unsigned long blocksize, unsigned long b_state)
 {
 	struct buffer_head *bh, *head, *tail;
@@ -1677,13 +1677,6 @@ struct buffer_head *folio_create_empty_buffers(struct folio *folio,
 
 	return head;
 }
-EXPORT_SYMBOL(folio_create_empty_buffers);
-
-void create_empty_buffers(struct page *page,
-			unsigned long blocksize, unsigned long b_state)
-{
-	folio_create_empty_buffers(page_folio(page), blocksize, b_state);
-}
 EXPORT_SYMBOL(create_empty_buffers);
 
 /**
@@ -1783,7 +1776,7 @@ static struct buffer_head *folio_create_buffers(struct folio *folio,
 
 	bh = folio_buffers(folio);
 	if (!bh)
-		bh = folio_create_empty_buffers(folio,
+		bh = create_empty_buffers(folio,
 				1 << READ_ONCE(inode->i_blkbits), b_state);
 	return bh;
 }
@@ -2686,7 +2679,7 @@ int block_truncate_page(struct address_space *mapping,
 
 	bh = folio_buffers(folio);
 	if (!bh)
-		bh = folio_create_empty_buffers(folio, blocksize, 0);
+		bh = create_empty_buffers(folio, blocksize, 0);
 
 	/* Find the buffer that contains "offset" */
 	offset = offset_in_folio(folio, from);
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 31c8f2641870..61277f7f8722 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1033,7 +1033,7 @@ static int ext4_block_write_begin(struct folio *folio, loff_t pos, unsigned len,
 
 	head = folio_buffers(folio);
 	if (!head)
-		head = folio_create_empty_buffers(folio, blocksize, 0);
+		head = create_empty_buffers(folio, blocksize, 0);
 	bbits = ilog2(blocksize);
 	block = (sector_t)folio->index << (PAGE_SHIFT - bbits);
 
@@ -1163,7 +1163,7 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping,
 	 * starting the handle.
 	 */
 	if (!folio_buffers(folio))
-		folio_create_empty_buffers(folio, inode->i_sb->s_blocksize, 0);
+		create_empty_buffers(folio, inode->i_sb->s_blocksize, 0);
 
 	folio_unlock(folio);
 
@@ -3654,7 +3654,7 @@ static int __ext4_block_zero_page_range(handle_t *handle,
 
 	bh = folio_buffers(folio);
 	if (!bh)
-		bh = folio_create_empty_buffers(folio, blocksize, 0);
+		bh = create_empty_buffers(folio, blocksize, 0);
 
 	/* Find the buffer that contains "offset" */
 	pos = blocksize;
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 7fe448fb948b..3aa57376d9c2 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -184,7 +184,7 @@ mext_page_mkuptodate(struct folio *folio, unsigned from, unsigned to)
 	blocksize = i_blocksize(inode);
 	head = folio_buffers(folio);
 	if (!head)
-		head = folio_create_empty_buffers(folio, blocksize, 0);
+		head = create_empty_buffers(folio, blocksize, 0);
 
 	block = (sector_t)folio->index << (PAGE_SHIFT - inode->i_blkbits);
 	for (bh = head, block_start = 0; bh != head || !block_start;
@@ -380,7 +380,7 @@ move_extent_per_page(struct file *o_filp, struct inode *donor_inode,
 	 * but keeping in mind that i_size will not change */
 	bh = folio_buffers(folio[0]);
 	if (!bh)
-		bh = folio_create_empty_buffers(folio[0],
+		bh = create_empty_buffers(folio[0],
 				1 << orig_inode->i_blkbits, 0);
 	for (i = 0; i < data_offset_in_page; i++)
 		bh = bh->b_this_page;
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index 52fad5bcb047..74f3302348a5 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -130,7 +130,7 @@ static int __gfs2_jdata_write_folio(struct folio *folio,
 	if (folio_test_checked(folio)) {
 		folio_clear_checked(folio);
 		if (!folio_buffers(folio)) {
-			folio_create_empty_buffers(folio,
+			create_empty_buffers(folio,
 					inode->i_sb->s_blocksize,
 					BIT(BH_Dirty)|BIT(BH_Uptodate));
 		}
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index f7ea1af9d842..398db80cb9c6 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -71,7 +71,7 @@ static int gfs2_unstuffer_folio(struct gfs2_inode *ip, struct buffer_head *dibh,
 		struct buffer_head *bh = folio_buffers(folio);
 
 		if (!bh)
-			bh = folio_create_empty_buffers(folio,
+			bh = create_empty_buffers(folio,
 				BIT(inode->i_blkbits), BIT(BH_Uptodate));
 
 		if (!buffer_mapped(bh))
diff --git a/fs/gfs2/meta_io.c b/fs/gfs2/meta_io.c
index f6d40d51f5ed..25ceb0805df2 100644
--- a/fs/gfs2/meta_io.c
+++ b/fs/gfs2/meta_io.c
@@ -134,7 +134,7 @@ struct buffer_head *gfs2_getbuf(struct gfs2_glock *gl, u64 blkno, int create)
 				mapping_gfp_mask(mapping) | __GFP_NOFAIL);
 		bh = folio_buffers(folio);
 		if (!bh)
-			bh = folio_create_empty_buffers(folio,
+			bh = create_empty_buffers(folio,
 				sdp->sd_sb.sb_bsize, 0);
 	} else {
 		folio = __filemap_get_folio(mapping, index,
diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c
index e45654ee4f4d..570a60aa3480 100644
--- a/fs/gfs2/quota.c
+++ b/fs/gfs2/quota.c
@@ -774,7 +774,7 @@ static int gfs2_write_buf_to_page(struct gfs2_sbd *sdp, unsigned long index,
 		return PTR_ERR(folio);
 	bh = folio_buffers(folio);
 	if (!bh)
-		bh = folio_create_empty_buffers(folio, bsize, 0);
+		bh = create_empty_buffers(folio, bsize, 0);
 
 	for (;;) {
 		/* Find the beginning block within the folio */
diff --git a/fs/mpage.c b/fs/mpage.c
index 964a6efe594d..ffb064ed9d04 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -119,7 +119,7 @@ static void map_buffer_to_folio(struct folio *folio, struct buffer_head *bh,
 			folio_mark_uptodate(folio);
 			return;
 		}
-		head = folio_create_empty_buffers(folio, i_blocksize(inode), 0);
+		head = create_empty_buffers(folio, i_blocksize(inode), 0);
 	}
 
 	page_bh = head;
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 7b754e6494d7..c97c77a39668 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -568,7 +568,7 @@ int nilfs_mdt_freeze_buffer(struct inode *inode, struct buffer_head *bh)
 
 	bh_frozen = folio_buffers(folio);
 	if (!bh_frozen)
-		bh_frozen = folio_create_empty_buffers(folio, 1 << blkbits, 0);
+		bh_frozen = create_empty_buffers(folio, 1 << blkbits, 0);
 
 	bh_frozen = get_nth_bh(bh_frozen, bh_offset(bh) >> blkbits);
 
diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c
index 696215d899bf..06b04758f289 100644
--- a/fs/nilfs2/page.c
+++ b/fs/nilfs2/page.c
@@ -34,7 +34,7 @@ static struct buffer_head *__nilfs_get_folio_block(struct folio *folio,
 	struct buffer_head *bh = folio_buffers(folio);
 
 	if (!bh)
-		bh = folio_create_empty_buffers(folio, 1 << blkbits, b_state);
+		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);
@@ -204,7 +204,7 @@ static void nilfs_copy_folio(struct folio *dst, struct folio *src,
 	sbh = folio_buffers(src);
 	dbh = folio_buffers(dst);
 	if (!dbh)
-		dbh = folio_create_empty_buffers(dst, sbh->b_size, 0);
+		dbh = create_empty_buffers(dst, sbh->b_size, 0);
 
 	if (copy_dirty)
 		mask |= BIT(BH_Dirty);
diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c
index 94388fe83cf8..55e31cc903d1 100644
--- a/fs/nilfs2/segment.c
+++ b/fs/nilfs2/segment.c
@@ -732,7 +732,7 @@ static size_t nilfs_lookup_dirty_data_buffers(struct inode *inode,
 		}
 		head = folio_buffers(folio);
 		if (!head)
-			head = folio_create_empty_buffers(folio,
+			head = create_empty_buffers(folio,
 					i_blocksize(inode), 0);
 		folio_unlock(folio);
 
diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index c4426992a2ee..71e31e789b29 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -189,7 +189,7 @@ static int ntfs_read_block(struct folio *folio)
 
 	head = folio_buffers(folio);
 	if (!head)
-		head = folio_create_empty_buffers(folio, blocksize, 0);
+		head = create_empty_buffers(folio, blocksize, 0);
 	bh = head;
 
 	/*
@@ -555,7 +555,7 @@ static int ntfs_write_block(struct folio *folio, struct writeback_control *wbc)
 	head = folio_buffers(folio);
 	if (!head) {
 		BUG_ON(!folio_test_uptodate(folio));
-		head = folio_create_empty_buffers(folio, blocksize,
+		head = create_empty_buffers(folio, blocksize,
 				(1 << BH_Uptodate) | (1 << BH_Dirty));
 	}
 	bh = head;
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index 099141d20db6..297c0b9db621 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -625,7 +625,7 @@ static int ntfs_prepare_pages_for_non_resident_write(struct page **pages,
 		 * create_empty_buffers() will create uptodate/dirty
 		 * buffers if the folio is uptodate/dirty.
 		 */
-		head = folio_create_empty_buffers(folio, blocksize, 0);
+		head = create_empty_buffers(folio, blocksize, 0);
 	bh = head;
 	do {
 		VCN cdelta;
diff --git a/fs/ntfs3/file.c b/fs/ntfs3/file.c
index 135af3340a19..a5a30a24ce5d 100644
--- a/fs/ntfs3/file.c
+++ b/fs/ntfs3/file.c
@@ -203,7 +203,7 @@ static int ntfs_zero_range(struct inode *inode, u64 vbo, u64 vbo_to)
 
 		head = folio_buffers(folio);
 		if (!head)
-			head = folio_create_empty_buffers(folio, blocksize, 0);
+			head = create_empty_buffers(folio, blocksize, 0);
 
 		bh = head;
 		bh_off = 0;
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 001ad1d288ec..ba790219d528 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -601,7 +601,7 @@ int ocfs2_map_page_blocks(struct page *page, u64 *p_blkno,
 
 	head = folio_buffers(folio);
 	if (!head)
-		head = folio_create_empty_buffers(folio, bsize, 0);
+		head = create_empty_buffers(folio, bsize, 0);
 
 	for (bh = head, block_start = 0; bh != head || !block_start;
 	     bh = bh->b_this_page, block_start += bsize) {
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index d7df556220a4..1d825459ee6e 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -2535,7 +2535,7 @@ static int reiserfs_write_full_folio(struct folio *folio,
 	 */
 	head = folio_buffers(folio);
 	if (!head)
-		head = folio_create_empty_buffers(folio, s->s_blocksize,
+		head = create_empty_buffers(folio, s->s_blocksize,
 				     (1 << BH_Dirty) | (1 << BH_Uptodate));
 
 	/*
diff --git a/fs/ufs/util.c b/fs/ufs/util.c
index d32de30009a0..13ba34e6d64f 100644
--- a/fs/ufs/util.c
+++ b/fs/ufs/util.c
@@ -264,6 +264,6 @@ struct folio *ufs_get_locked_folio(struct address_space *mapping,
 		}
 	}
 	if (!folio_buffers(folio))
-		folio_create_empty_buffers(folio, 1 << inode->i_blkbits, 0);
+		create_empty_buffers(folio, 1 << inode->i_blkbits, 0);
 	return folio;
 }
diff --git a/include/linux/buffer_head.h b/include/linux/buffer_head.h
index 3d85a0cf0ca5..5f23ee599889 100644
--- a/include/linux/buffer_head.h
+++ b/include/linux/buffer_head.h
@@ -201,9 +201,7 @@ struct buffer_head *folio_alloc_buffers(struct folio *folio, unsigned long size,
 					gfp_t gfp);
 struct buffer_head *alloc_page_buffers(struct page *page, unsigned long size,
 		bool retry);
-void create_empty_buffers(struct page *, unsigned long,
-			unsigned long b_state);
-struct buffer_head *folio_create_empty_buffers(struct folio *folio,
+struct buffer_head *create_empty_buffers(struct folio *folio,
 		unsigned long blocksize, unsigned long b_state);
 void end_buffer_read_sync(struct buffer_head *bh, int uptodate);
 void end_buffer_write_sync(struct buffer_head *bh, int uptodate);
-- 
2.40.1

Re: [PATCH v2 01/27] buffer: Return bool from grow_dev_folio()

[Ryusuke Konishi]

On Tue, Oct 17, 2023 at 5:11 AM Matthew Wilcox (Oracle) wrote:
>
> Rename grow_dev_page() to grow_dev_folio() and make it return a bool.
> Document what that bool means; it's more subtle than it first appears.
> Also rename the 'failed' label to 'unlock' beacuse it's not exactly
> 'failed'.  It just hasn't succeeded.
>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> ---
>  fs/buffer.c | 39 ++++++++++++++++++++++-----------------
>  1 file changed, 22 insertions(+), 17 deletions(-)
>
> diff --git a/fs/buffer.c b/fs/buffer.c
> index b33cc74e9649..dec41d84044b 100644
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -1024,24 +1024,26 @@ static sector_t folio_init_buffers(struct folio *folio,
>  }
>
>  /*
> - * Create the page-cache page that contains the requested block.
> + * Create the page-cache folio that contains the requested block.
>   *
>   * This is used purely for blockdev mappings.
> + *
> + * Returns false if we have a 'permanent' failure.  Returns true if
> + * we succeeded, or the caller should retry.
>   */
> -static int
> -grow_dev_page(struct block_device *bdev, sector_t block,
> +static bool grow_dev_folio(struct block_device *bdev, sector_t block,
>               pgoff_t index, int size, int sizebits, gfp_t gfp)
>  {
>         struct inode *inode = bdev->bd_inode;
>         struct folio *folio;
>         struct buffer_head *bh;
>         sector_t end_block;
> -       int ret = 0;
> +       bool ret;
>
>         folio = __filemap_get_folio(inode->i_mapping, index,
>                         FGP_LOCK | FGP_ACCESSED | FGP_CREAT, gfp);
>         if (IS_ERR(folio))
> -               return PTR_ERR(folio);
> +               return false;
>
>         bh = folio_buffers(folio);
>         if (bh) {
> @@ -1050,14 +1052,17 @@ grow_dev_page(struct block_device *bdev, sector_t block,
>                                         (sector_t)index << sizebits, size);
>                         goto done;
>                 }
> +
> +               /* Caller should retry if this call fails */
> +               ret = true;
>                 if (!try_to_free_buffers(folio))
> -                       goto failed;
> +                       goto unlock;
>         }
>
> -       ret = -ENOMEM;
> +       ret = false;
>         bh = folio_alloc_buffers(folio, size, gfp | __GFP_ACCOUNT);
>         if (!bh)
> -               goto failed;
> +               goto unlock;
>
>         /*
>          * Link the folio to the buffers and initialise them.  Take the
> @@ -1070,19 +1075,19 @@ grow_dev_page(struct block_device *bdev, sector_t block,
>                         (sector_t)index << sizebits, size);
>         spin_unlock(&inode->i_mapping->private_lock);
>  done:
> -       ret = (block < end_block) ? 1 : -ENXIO;
> -failed:
> +       ret = block < end_block;
> +unlock:
>         folio_unlock(folio);
>         folio_put(folio);
>         return ret;
>  }
>
>  /*
> - * Create buffers for the specified block device block's page.  If
> - * that page was dirty, the buffers are set dirty also.
> + * Create buffers for the specified block device block's folio.  If
> + * that folio was dirty, the buffers are set dirty also.
>   */
> -static int
> -grow_buffers(struct block_device *bdev, sector_t block, int size, gfp_t gfp)
> +static bool grow_buffers(struct block_device *bdev, sector_t block,
> +               int size, gfp_t gfp)
>  {
>         pgoff_t index;
>         int sizebits;
> @@ -1099,11 +1104,11 @@ grow_buffers(struct block_device *bdev, sector_t block, int size, gfp_t gfp)
>                         "device %pg\n",
>                         __func__, (unsigned long long)block,
>                         bdev);
> -               return -EIO;
> +               return false;
>         }
>
> -       /* Create a page with the proper size buffers.. */
> -       return grow_dev_page(bdev, block, index, size, sizebits, gfp);
> +       /* Create a folio with the proper size buffers.. */
> +       return grow_dev_folio(bdev, block, index, size, sizebits, gfp);
>  }
>
>  static struct buffer_head *
> --
> 2.40.1

This changes the return type of grow_buffers() from "int"  to "bool".
But, it seems that the caller, __getblk_slow(), has not changed the
type of the variable "ret" that receives its return value:

        for (;;) {
                struct buffer_head *bh;
                int ret;

                bh = __find_get_block(bdev, block, size);
                if (bh)
                        return bh;

                ret = grow_buffers(bdev, block, size, gfp);
                if (ret < 0)
                        return NULL;
        }

So, it looks like the error check immediately after calling
grow_buffers() will not branch like before.
Is this okay ?   Or, am I missing some other changes?

Also, there is a typo in the changelog: "beacuse" -> "because".

Regards,
Ryusuke Konishi

Re: [PATCH v2 01/27] buffer: Return bool from grow_dev_folio()

[Matthew Wilcox]

On Wed, Oct 18, 2023 at 04:41:44AM +0900, Ryusuke Konishi wrote:
> On Tue, Oct 17, 2023 at 5:11 AM Matthew Wilcox (Oracle) wrote:
> >
> > Rename grow_dev_page() to grow_dev_folio() and make it return a bool.
> > Document what that bool means; it's more subtle than it first appears.
> > Also rename the 'failed' label to 'unlock' beacuse it's not exactly
> > 'failed'.  It just hasn't succeeded.
> 
> This changes the return type of grow_buffers() from "int"  to "bool".
> But, it seems that the caller, __getblk_slow(), has not changed the
> type of the variable "ret" that receives its return value:
[...]
> 
> So, it looks like the error check immediately after calling
> grow_buffers() will not branch like before.
> Is this okay ?   Or, am I missing some other changes?
> 
> Also, there is a typo in the changelog: "beacuse" -> "because".

Argh, yes.  Andrew, please drop this patch for now.  I'll submit
something better next cycle.