Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:
 "Improve performance for ext4 by allowing multiple process to perform
  direct I/O writes to preallocated blocks by using a shared inode lock
  instead of taking an exclusive lock.

  In addition, multiple bug fixes and cleanups"

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix incorrect options show of original mount_opt and extend mount_opt2
  ext4: Fix possible corruption when moving a directory
  ext4: init error handle resource before init group descriptors
  ext4: fix task hung in ext4_xattr_delete_inode
  jbd2: fix data missing when reusing bh which is ready to be checkpointed
  ext4: update s_journal_inum if it changes after journal replay
  ext4: fail ext4_iget if special inode unallocated
  ext4: fix function prototype mismatch for ext4_feat_ktype
  ext4: remove unnecessary variable initialization
  ext4: fix inode tree inconsistency caused by ENOMEM
  ext4: refuse to create ea block when umounted
  ext4: optimize ea_inode block expansion
  ext4: remove dead code in updating backup sb
  ext4: dio take shared inode lock when overwriting preallocated blocks
  ext4: don't show commit interval if it is zero
  ext4: use ext4_fc_tl_mem in fast-commit replay path
  ext4: improve xattr consistency checking and error reporting

1 files changed, 22 insertions, 12 deletions

diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 6bdf61a62c7962..0b8b4499e5ca18 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -202,8 +202,9 @@ ext4_extending_io(struct inode *inode, loff_t offset, size_t len)
 	return false;
 }
 
-/* Is IO overwriting allocated and initialized blocks? */
-static bool ext4_overwrite_io(struct inode *inode, loff_t pos, loff_t len)
+/* Is IO overwriting allocated or initialized blocks? */
+static bool ext4_overwrite_io(struct inode *inode,
+			      loff_t pos, loff_t len, bool *unwritten)
 {
 	struct ext4_map_blocks map;
 	unsigned int blkbits = inode->i_blkbits;
@@ -217,12 +218,15 @@ static bool ext4_overwrite_io(struct inode *inode, loff_t pos, loff_t len)
 	blklen = map.m_len;
 
 	err = ext4_map_blocks(NULL, inode, &map, 0);
+	if (err != blklen)
+		return false;
 	/*
 	 * 'err==len' means that all of the blocks have been preallocated,
-	 * regardless of whether they have been initialized or not. To exclude
-	 * unwritten extents, we need to check m_flags.
+	 * regardless of whether they have been initialized or not. We need to
+	 * check m_flags to distinguish the unwritten extents.
 	 */
-	return err == blklen && (map.m_flags & EXT4_MAP_MAPPED);
+	*unwritten = !(map.m_flags & EXT4_MAP_MAPPED);
+	return true;
 }
 
 static ssize_t ext4_generic_write_checks(struct kiocb *iocb,
@@ -431,11 +435,16 @@ static const struct iomap_dio_ops ext4_dio_write_ops = {
  * - For extending writes case we don't take the shared lock, since it requires
  *   updating inode i_disksize and/or orphan handling with exclusive lock.
  *
- * - shared locking will only be true mostly with overwrites. Otherwise we will
- *   switch to exclusive i_rwsem lock.
+ * - shared locking will only be true mostly with overwrites, including
+ *   initialized blocks and unwritten blocks. For overwrite unwritten blocks
+ *   we protect splitting extents by i_data_sem in ext4_inode_info, so we can
+ *   also release exclusive i_rwsem lock.
+ *
+ * - Otherwise we will switch to exclusive i_rwsem lock.
  */
 static ssize_t ext4_dio_write_checks(struct kiocb *iocb, struct iov_iter *from,
-				     bool *ilock_shared, bool *extend)
+				     bool *ilock_shared, bool *extend,
+				     bool *unwritten)
 {
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file_inode(file);
@@ -459,7 +468,7 @@ restart:
 	 * in file_modified().
 	 */
 	if (*ilock_shared && (!IS_NOSEC(inode) || *extend ||
-	     !ext4_overwrite_io(inode, offset, count))) {
+	     !ext4_overwrite_io(inode, offset, count, unwritten))) {
 		if (iocb->ki_flags & IOCB_NOWAIT) {
 			ret = -EAGAIN;
 			goto out;
@@ -491,7 +500,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from)
 	loff_t offset = iocb->ki_pos;
 	size_t count = iov_iter_count(from);
 	const struct iomap_ops *iomap_ops = &ext4_iomap_ops;
-	bool extend = false, unaligned_io = false;
+	bool extend = false, unaligned_io = false, unwritten = false;
 	bool ilock_shared = true;
 
 	/*
@@ -534,7 +543,8 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		return ext4_buffered_write_iter(iocb, from);
 	}
 
-	ret = ext4_dio_write_checks(iocb, from, &ilock_shared, &extend);
+	ret = ext4_dio_write_checks(iocb, from,
+				    &ilock_shared, &extend, &unwritten);
 	if (ret <= 0)
 		return ret;
 
@@ -582,7 +592,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from)
 		ext4_journal_stop(handle);
 	}
 
-	if (ilock_shared)
+	if (ilock_shared && !unwritten)
 		iomap_ops = &ext4_iomap_overwrite_ops;
 	ret = iomap_dio_rw(iocb, from, iomap_ops, &ext4_dio_write_ops,
 			   (unaligned_io || extend) ? IOMAP_DIO_FORCE_WAIT : 0,