ext4: add delalloc support for inline data

For delayed allocation mode, we write to inline data if the file
is small enough. And in case of we write to some offset larger
than the inline size, the 1st page is dirtied, so that
ext4_da_writepages can handle the conversion. When the 1st page
is initialized with blocks, the inline part is removed.

Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

4 files changed, 262 insertions, 9 deletions

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 9f4efc6..268636a 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -2022,6 +2022,8 @@ int ext4_get_block_write(struct inode *inode, sector_t iblock,
 			 struct buffer_head *bh_result, int create);
 int ext4_get_block(struct inode *inode, sector_t iblock,
 				struct buffer_head *bh_result, int create);
+int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
+			   struct buffer_head *bh, int create);
 int ext4_walk_page_buffers(handle_t *handle,
 			   struct buffer_head *head,
 			   unsigned from,
@@ -2031,6 +2033,8 @@ int ext4_walk_page_buffers(handle_t *handle,
 				     struct buffer_head *bh));
 int do_journal_get_write_access(handle_t *handle,
 				struct buffer_head *bh);
+#define FALL_BACK_TO_NONDELALLOC 1
+#define CONVERT_INLINE_DATA	 2
 
 extern struct inode *ext4_iget(struct super_block *, unsigned long);
 extern int  ext4_write_inode(struct inode *, struct writeback_control *);
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 01274b1..65f7ffb 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -771,6 +771,183 @@ ext4_journalled_write_inline_data(struct inode *inode,
 	return iloc.bh;
 }
 
+/*
+ * Try to make the page cache and handle ready for the inline data case.
+ * We can call this function in 2 cases:
+ * 1. The inode is created and the first write exceeds inline size. We can
+ *    clear the inode state safely.
+ * 2. The inode has inline data, then we need to read the data, make it
+ *    update and dirty so that ext4_da_writepages can handle it. We don't
+ *    need to start the journal since the file's metatdata isn't changed now.
+ */
+static int ext4_da_convert_inline_data_to_extent(struct address_space *mapping,
+						 struct inode *inode,
+						 unsigned flags,
+						 void **fsdata)
+{
+	int ret = 0, inline_size;
+	struct page *page;
+
+	page = grab_cache_page_write_begin(mapping, 0, flags);
+	if (!page)
+		return -ENOMEM;
+
+	down_read(&EXT4_I(inode)->xattr_sem);
+	if (!ext4_has_inline_data(inode)) {
+		ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+		goto out;
+	}
+
+	inline_size = ext4_get_inline_size(inode);
+
+	if (!PageUptodate(page)) {
+		ret = ext4_read_inline_page(inode, page);
+		if (ret < 0)
+			goto out;
+	}
+
+	ret = __block_write_begin(page, 0, inline_size,
+				  ext4_da_get_block_prep);
+	if (ret) {
+		ext4_truncate_failed_write(inode);
+		goto out;
+	}
+
+	SetPageDirty(page);
+	SetPageUptodate(page);
+	ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+	*fsdata = (void *)CONVERT_INLINE_DATA;
+
+out:
+	up_read(&EXT4_I(inode)->xattr_sem);
+	if (page) {
+		unlock_page(page);
+		page_cache_release(page);
+	}
+	return ret;
+}
+
+/*
+ * Prepare the write for the inline data.
+ * If the the data can be written into the inode, we just read
+ * the page and make it uptodate, and start the journal.
+ * Otherwise read the page, makes it dirty so that it can be
+ * handle in writepages(the i_disksize update is left to the
+ * normal ext4_da_write_end).
+ */
+int ext4_da_write_inline_data_begin(struct address_space *mapping,
+				    struct inode *inode,
+				    loff_t pos, unsigned len,
+				    unsigned flags,
+				    struct page **pagep,
+				    void **fsdata)
+{
+	int ret, inline_size;
+	handle_t *handle;
+	struct page *page;
+	struct ext4_iloc iloc;
+
+	ret = ext4_get_inode_loc(inode, &iloc);
+	if (ret)
+		return ret;
+
+	handle = ext4_journal_start(inode, 1);
+	if (IS_ERR(handle)) {
+		ret = PTR_ERR(handle);
+		handle = NULL;
+		goto out;
+	}
+
+	inline_size = ext4_get_max_inline_size(inode);
+
+	ret = -ENOSPC;
+	if (inline_size >= pos + len) {
+		ret = ext4_prepare_inline_data(handle, inode, pos + len);
+		if (ret && ret != -ENOSPC)
+			goto out;
+	}
+
+	if (ret == -ENOSPC) {
+		ret = ext4_da_convert_inline_data_to_extent(mapping,
+							    inode,
+							    flags,
+							    fsdata);
+		goto out;
+	}
+
+	/*
+	 * We cannot recurse into the filesystem as the transaction
+	 * is already started.
+	 */
+	flags |= AOP_FLAG_NOFS;
+
+	page = grab_cache_page_write_begin(mapping, 0, flags);
+	if (!page) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	down_read(&EXT4_I(inode)->xattr_sem);
+	if (!ext4_has_inline_data(inode)) {
+		ret = 0;
+		goto out_release_page;
+	}
+
+	if (!PageUptodate(page)) {
+		ret = ext4_read_inline_page(inode, page);
+		if (ret < 0)
+			goto out_release_page;
+	}
+
+	up_read(&EXT4_I(inode)->xattr_sem);
+	*pagep = page;
+	handle = NULL;
+	brelse(iloc.bh);
+	return 1;
+out_release_page:
+	up_read(&EXT4_I(inode)->xattr_sem);
+	unlock_page(page);
+	page_cache_release(page);
+out:
+	if (handle)
+		ext4_journal_stop(handle);
+	brelse(iloc.bh);
+	return ret;
+}
+
+int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+				  unsigned len, unsigned copied,
+				  struct page *page)
+{
+	int i_size_changed = 0;
+
+	copied = ext4_write_inline_data_end(inode, pos, len, copied, page);
+
+	/*
+	 * No need to use i_size_read() here, the i_size
+	 * cannot change under us because we hold i_mutex.
+	 *
+	 * But it's important to update i_size while still holding page lock:
+	 * page writeout could otherwise come in and zero beyond i_size.
+	 */
+	if (pos+copied > inode->i_size) {
+		i_size_write(inode, pos+copied);
+		i_size_changed = 1;
+	}
+	unlock_page(page);
+	page_cache_release(page);
+
+	/*
+	 * Don't mark the inode dirty under page lock. First, it unnecessarily
+	 * makes the holding time of page lock longer. Second, it forces lock
+	 * ordering of page lock and transaction start for journaling
+	 * filesystems.
+	 */
+	if (i_size_changed)
+		mark_inode_dirty(inode);
+
+	return copied;
+}
 
 int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
 {
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 5c91622..f16ae02 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1790,7 +1790,19 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
 	 * file system block.
 	 */
 	down_read((&EXT4_I(inode)->i_data_sem));
-	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+	if (ext4_has_inline_data(inode)) {
+		/*
+		 * We will soon create blocks for this page, and let
+		 * us pretend as if the blocks aren't allocated yet.
+		 * In case of clusters, we have to handle the work
+		 * of mapping from cluster so that the reserved space
+		 * is calculated properly.
+		 */
+		if ((EXT4_SB(inode->i_sb)->s_cluster_ratio > 1) &&
+		    ext4_find_delalloc_cluster(inode, map->m_lblk))
+			map->m_flags |= EXT4_MAP_FROM_CLUSTER;
+		retval = 0;
+	} else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
 		retval = ext4_ext_map_blocks(NULL, inode, map, 0);
 	else
 		retval = ext4_ind_map_blocks(NULL, inode, map, 0);
@@ -1841,8 +1853,8 @@ out_unlock:
  * We also have b_blocknr = physicalblock mapping unwritten extent and b_bdev
  * initialized properly.
  */
-static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
-				  struct buffer_head *bh, int create)
+int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
+			   struct buffer_head *bh, int create)
 {
 	struct ext4_map_blocks map;
 	int ret = 0;
@@ -2119,7 +2131,8 @@ static int ext4_da_writepages_trans_blocks(struct inode *inode)
  * mpage_da_map_and_submit to map a single contiguous memory region
  * and then write them.
  */
-static int write_cache_pages_da(struct address_space *mapping,
+static int write_cache_pages_da(handle_t *handle,
+				struct address_space *mapping,
 				struct writeback_control *wbc,
 				struct mpage_da_data *mpd,
 				pgoff_t *done_index)
@@ -2198,6 +2211,17 @@ static int write_cache_pages_da(struct address_space *mapping,
 			wait_on_page_writeback(page);
 			BUG_ON(PageWriteback(page));
 
+			/*
+			 * If we have inline data and arrive here, it means that
+			 * we will soon create the block for the 1st page, so
+			 * we'd better clear the inline data here.
+			 */
+			if (ext4_has_inline_data(inode)) {
+				BUG_ON(ext4_test_inode_state(inode,
+						EXT4_STATE_MAY_INLINE_DATA));
+				ext4_destroy_inline_data(handle, inode);
+			}
+
 			if (mpd->next_page != page->index)
 				mpd->first_page = page->index;
 			mpd->next_page = page->index + 1;
@@ -2404,7 +2428,8 @@ retry:
 		 * contiguous region of logical blocks that need
 		 * blocks to be allocated by ext4 and submit them.
 		 */
-		ret = write_cache_pages_da(mapping, wbc, &mpd, &done_index);
+		ret = write_cache_pages_da(handle, mapping,
+					   wbc, &mpd, &done_index);
 		/*
 		 * If we have a contiguous extent of pages and we
 		 * haven't done the I/O yet, map the blocks and submit
@@ -2468,7 +2493,6 @@ out_writepages:
 	return ret;
 }
 
-#define FALL_BACK_TO_NONDELALLOC 1
 static int ext4_nonda_switch(struct super_block *sb)
 {
 	s64 free_blocks, dirty_blocks;
@@ -2525,6 +2549,19 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping,
 	}
 	*fsdata = (void *)0;
 	trace_ext4_da_write_begin(inode, pos, len, flags);
+
+	if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) {
+		ret = ext4_da_write_inline_data_begin(mapping, inode,
+						      pos, len, flags,
+						      pagep, fsdata);
+		if (ret < 0)
+			goto out;
+		if (ret == 1) {
+			ret = 0;
+			goto out;
+		}
+	}
+
 retry:
 	/*
 	 * With delayed allocation, we don't log the i_disksize update
@@ -2626,10 +2663,10 @@ static int ext4_da_write_end(struct file *file,
 	 * changes.  So let's piggyback the i_disksize mark_inode_dirty
 	 * into that.
 	 */
-
 	new_i_size = pos + copied;
 	if (copied && new_i_size > EXT4_I(inode)->i_disksize) {
-		if (ext4_da_should_update_i_disksize(page, end)) {
+		if (ext4_has_inline_data(inode) ||
+		    ext4_da_should_update_i_disksize(page, end)) {
 			down_write(&EXT4_I(inode)->i_data_sem);
 			if (new_i_size > EXT4_I(inode)->i_disksize)
 				EXT4_I(inode)->i_disksize = new_i_size;
@@ -2641,8 +2678,16 @@ static int ext4_da_write_end(struct file *file,
 			ext4_mark_inode_dirty(handle, inode);
 		}
 	}
-	ret2 = generic_write_end(file, mapping, pos, len, copied,
+
+	if (write_mode != CONVERT_INLINE_DATA &&
+	    ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA) &&
+	    ext4_has_inline_data(inode))
+		ret2 = ext4_da_write_inline_data_end(inode, pos, len, copied,
+						     page);
+	else
+		ret2 = generic_write_end(file, mapping, pos, len, copied,
 							page, fsdata);
+
 	copied = ret2;
 	if (ret2 < 0)
 		ret = ret2;
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 7095ac1..37e66f8 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -154,6 +154,15 @@ extern struct buffer_head *
 ext4_journalled_write_inline_data(struct inode *inode,
 				  unsigned len,
 				  struct page *page);
+extern int ext4_da_write_inline_data_begin(struct address_space *mapping,
+					   struct inode *inode,
+					   loff_t pos, unsigned len,
+					   unsigned flags,
+					   struct page **pagep,
+					   void **fsdata);
+extern int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+					 unsigned len, unsigned copied,
+					 struct page *page);
 # else  /* CONFIG_EXT4_FS_XATTR */
 
 static inline int
@@ -300,6 +309,24 @@ ext4_journalled_write_inline_data(struct inode *inode,
 {
 	return NULL;
 }
+
+static inline int
+ext4_da_write_inline_data_begin(struct address_space *mapping,
+				struct inode *inode,
+				loff_t pos, unsigned len,
+				unsigned flags,
+				struct page **pagep,
+				void **fsdata)
+{
+	return 0;
+}
+
+static inline int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+						unsigned len, unsigned copied,
+						struct page *page)
+{
+	return 0;
+}
 # endif  /* CONFIG_EXT4_FS_XATTR */
 
 #ifdef CONFIG_EXT4_FS_SECURITY