Merge tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 fixes from Ted Ts'o:
 "A number of bug fixes for ext4:

   - Fix for the new fast_commit feature

   - Fix some error handling codepaths in whiteout handling and
     mountpoint sampling

   - Fix how we write ext4_error information so it goes through the
     journal when journalling is active, to avoid races that can lead to
     lost error information, superblock checksum failures, or DIF/DIX
     features"

* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: remove expensive flush on fast commit
  ext4: fix bug for rename with RENAME_WHITEOUT
  ext4: fix wrong list_splice in ext4_fc_cleanup
  ext4: use IS_ERR instead of IS_ERR_OR_NULL and set inode null when IS_ERR
  ext4: don't leak old mountpoint samples
  ext4: drop ext4_handle_dirty_super()
  ext4: fix superblock checksum failure when setting password salt
  ext4: use sbi instead of EXT4_SB(sb) in ext4_update_super()
  ext4: save error info to sb through journal if available
  ext4: protect superblock modifications with a buffer lock
  ext4: drop sync argument of ext4_commit_super()
  ext4: combine ext4_handle_error() and save_error_info()

fs/ext4/ext4_jbd2.c

@@ -372,20 +372,3 @@ int __ext4_handle_dirty_metadata(const char *where, unsigned int line,
 	}
 	return err;
 }
-
-int __ext4_handle_dirty_super(const char *where, unsigned int line,
-			      handle_t *handle, struct super_block *sb)
-{
-	struct buffer_head *bh = EXT4_SB(sb)->s_sbh;
-	int err = 0;
-
-	ext4_superblock_csum_set(sb);
-	if (ext4_handle_valid(handle)) {
-		err = jbd2_journal_dirty_metadata(handle, bh);
-		if (err)
-			ext4_journal_abort_handle(where, line, __func__,
-						  bh, handle, err);
-	} else
-		mark_buffer_dirty(bh);
-	return err;
-}

fs/ext4/ext4_jbd2.h

@@ -244,9 +244,6 @@ int __ext4_handle_dirty_metadata(const char *where, unsigned int line,
 				 handle_t *handle, struct inode *inode,
 				 struct buffer_head *bh);
 
-int __ext4_handle_dirty_super(const char *where, unsigned int line,
-			      handle_t *handle, struct super_block *sb);
-
 #define ext4_journal_get_write_access(handle, bh) \
 	__ext4_journal_get_write_access(__func__, __LINE__, (handle), (bh))
 #define ext4_forget(handle, is_metadata, inode, bh, block_nr) \
@@ -257,8 +254,6 @@ int __ext4_handle_dirty_super(const char *where, unsigned int line,
 #define ext4_handle_dirty_metadata(handle, inode, bh) \
 	__ext4_handle_dirty_metadata(__func__, __LINE__, (handle), (inode), \
 				     (bh))
-#define ext4_handle_dirty_super(handle, sb) \
-	__ext4_handle_dirty_super(__func__, __LINE__, (handle), (sb))
 
 handle_t *__ext4_journal_start_sb(struct super_block *sb, unsigned int line,
 				  int type, int blocks, int rsv_blocks,

fs/ext4/fast_commit.c

@@ -604,13 +604,13 @@ void ext4_fc_track_range(handle_t *handle, struct inode *inode, ext4_lblk_t star
 	trace_ext4_fc_track_range(inode, start, end, ret);
 }
 
-static void ext4_fc_submit_bh(struct super_block *sb)
+static void ext4_fc_submit_bh(struct super_block *sb, bool is_tail)
 {
 	int write_flags = REQ_SYNC;
 	struct buffer_head *bh = EXT4_SB(sb)->s_fc_bh;
 
-	/* TODO: REQ_FUA | REQ_PREFLUSH is unnecessarily expensive. */
-	if (test_opt(sb, BARRIER))
+	/* Add REQ_FUA | REQ_PREFLUSH only its tail */
+	if (test_opt(sb, BARRIER) && is_tail)
 		write_flags |= REQ_FUA | REQ_PREFLUSH;
 	lock_buffer(bh);
 	set_buffer_dirty(bh);
@@ -684,7 +684,7 @@ static u8 *ext4_fc_reserve_space(struct super_block *sb, int len, u32 *crc)
 		*crc = ext4_chksum(sbi, *crc, tl, sizeof(*tl));
 	if (pad_len > 0)
 		ext4_fc_memzero(sb, tl + 1, pad_len, crc);
-	ext4_fc_submit_bh(sb);
+	ext4_fc_submit_bh(sb, false);
 
 	ret = jbd2_fc_get_buf(EXT4_SB(sb)->s_journal, &bh);
 	if (ret)
@@ -741,7 +741,7 @@ static int ext4_fc_write_tail(struct super_block *sb, u32 crc)
 	tail.fc_crc = cpu_to_le32(crc);
 	ext4_fc_memcpy(sb, dst, &tail.fc_crc, sizeof(tail.fc_crc), NULL);
 
-	ext4_fc_submit_bh(sb);
+	ext4_fc_submit_bh(sb, true);
 
 	return 0;
 }
@@ -1268,7 +1268,7 @@ static void ext4_fc_cleanup(journal_t *journal, int full)
 	list_splice_init(&sbi->s_fc_dentry_q[FC_Q_STAGING],
 				&sbi->s_fc_dentry_q[FC_Q_MAIN]);
 	list_splice_init(&sbi->s_fc_q[FC_Q_STAGING],
-				&sbi->s_fc_q[FC_Q_STAGING]);
+				&sbi->s_fc_q[FC_Q_MAIN]);
 
 	ext4_clear_mount_flag(sb, EXT4_MF_FC_COMMITTING);
 	ext4_clear_mount_flag(sb, EXT4_MF_FC_INELIGIBLE);
@@ -1318,14 +1318,14 @@ static int ext4_fc_replay_unlink(struct super_block *sb, struct ext4_fc_tl *tl)
 	entry.len = darg.dname_len;
 	inode = ext4_iget(sb, darg.ino, EXT4_IGET_NORMAL);
 
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "Inode %d not found", darg.ino);
 		return 0;
 	}
 
 	old_parent = ext4_iget(sb, darg.parent_ino,
 				EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(old_parent)) {
+	if (IS_ERR(old_parent)) {
 		jbd_debug(1, "Dir with inode  %d not found", darg.parent_ino);
 		iput(inode);
 		return 0;
@@ -1410,7 +1410,7 @@ static int ext4_fc_replay_link(struct super_block *sb, struct ext4_fc_tl *tl)
 			darg.parent_ino, darg.dname_len);
 
 	inode = ext4_iget(sb, darg.ino, EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "Inode not found.");
 		return 0;
 	}
@@ -1466,10 +1466,11 @@ static int ext4_fc_replay_inode(struct super_block *sb, struct ext4_fc_tl *tl)
 	trace_ext4_fc_replay(sb, tag, ino, 0, 0);
 
 	inode = ext4_iget(sb, ino, EXT4_IGET_NORMAL);
-	if (!IS_ERR_OR_NULL(inode)) {
+	if (!IS_ERR(inode)) {
 		ext4_ext_clear_bb(inode);
 		iput(inode);
 	}
+	inode = NULL;
 
 	ext4_fc_record_modified_inode(sb, ino);
 
@@ -1512,7 +1513,7 @@ static int ext4_fc_replay_inode(struct super_block *sb, struct ext4_fc_tl *tl)
 
 	/* Given that we just wrote the inode on disk, this SHOULD succeed. */
 	inode = ext4_iget(sb, ino, EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "Inode not found.");
 		return -EFSCORRUPTED;
 	}
@@ -1564,7 +1565,7 @@ static int ext4_fc_replay_create(struct super_block *sb, struct ext4_fc_tl *tl)
 		goto out;
 
 	inode = ext4_iget(sb, darg.ino, EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "inode %d not found.", darg.ino);
 		inode = NULL;
 		ret = -EINVAL;
@@ -1577,7 +1578,7 @@ static int ext4_fc_replay_create(struct super_block *sb, struct ext4_fc_tl *tl)
 		 * dot and dot dot dirents are setup properly.
 		 */
 		dir = ext4_iget(sb, darg.parent_ino, EXT4_IGET_NORMAL);
-		if (IS_ERR_OR_NULL(dir)) {
+		if (IS_ERR(dir)) {
 			jbd_debug(1, "Dir %d not found.", darg.ino);
 			goto out;
 		}
@@ -1653,7 +1654,7 @@ static int ext4_fc_replay_add_range(struct super_block *sb,
 
 	inode = ext4_iget(sb, le32_to_cpu(fc_add_ex->fc_ino),
 				EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "Inode not found.");
 		return 0;
 	}
@@ -1777,7 +1778,7 @@ ext4_fc_replay_del_range(struct super_block *sb, struct ext4_fc_tl *tl)
 		le32_to_cpu(lrange->fc_ino), cur, remaining);
 
 	inode = ext4_iget(sb, le32_to_cpu(lrange->fc_ino), EXT4_IGET_NORMAL);
-	if (IS_ERR_OR_NULL(inode)) {
+	if (IS_ERR(inode)) {
 		jbd_debug(1, "Inode %d not found", le32_to_cpu(lrange->fc_ino));
 		return 0;
 	}
@@ -1832,7 +1833,7 @@ static void ext4_fc_set_bitmaps_and_counters(struct super_block *sb)
 	for (i = 0; i < state->fc_modified_inodes_used; i++) {
 		inode = ext4_iget(sb, state->fc_modified_inodes[i],
 			EXT4_IGET_NORMAL);
-		if (IS_ERR_OR_NULL(inode)) {
+		if (IS_ERR(inode)) {
 			jbd_debug(1, "Inode %d not found.",
 				state->fc_modified_inodes[i]);
 			continue;
@@ -1849,7 +1850,7 @@ static void ext4_fc_set_bitmaps_and_counters(struct super_block *sb)
 
 			if (ret > 0) {
 				path = ext4_find_extent(inode, map.m_lblk, NULL, 0);
-				if (!IS_ERR_OR_NULL(path)) {
+				if (!IS_ERR(path)) {
 					for (j = 0; j < path->p_depth; j++)
 						ext4_mb_mark_bb(inode->i_sb,
 							path[j].p_block, 1, 1);

fs/ext4/file.c

@@ -809,9 +809,12 @@ static int ext4_sample_last_mounted(struct super_block *sb,
 	err = ext4_journal_get_write_access(handle, sbi->s_sbh);
 	if (err)
 		goto out_journal;
-	strlcpy(sbi->s_es->s_last_mounted, cp,
+	lock_buffer(sbi->s_sbh);
+	strncpy(sbi->s_es->s_last_mounted, cp,
 		sizeof(sbi->s_es->s_last_mounted));
-	ext4_handle_dirty_super(handle, sb);
+	ext4_superblock_csum_set(sb);
+	unlock_buffer(sbi->s_sbh);
+	ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 out_journal:
 	ext4_journal_stop(handle);
 out:

fs/ext4/inode.c

@@ -5150,9 +5150,13 @@ static int ext4_do_update_inode(handle_t *handle,
 		err = ext4_journal_get_write_access(handle, EXT4_SB(sb)->s_sbh);
 		if (err)
 			goto out_brelse;
+		lock_buffer(EXT4_SB(sb)->s_sbh);
 		ext4_set_feature_large_file(sb);
+		ext4_superblock_csum_set(sb);
+		unlock_buffer(EXT4_SB(sb)->s_sbh);
 		ext4_handle_sync(handle);
-		err = ext4_handle_dirty_super(handle, sb);
+		err = ext4_handle_dirty_metadata(handle, NULL,
+						 EXT4_SB(sb)->s_sbh);
 	}
 	ext4_update_inode_fsync_trans(handle, inode, need_datasync);
 out_brelse:

fs/ext4/ioctl.c

@@ -1157,7 +1157,10 @@ static long __ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 			err = ext4_journal_get_write_access(handle, sbi->s_sbh);
 			if (err)
 				goto pwsalt_err_journal;
+			lock_buffer(sbi->s_sbh);
 			generate_random_uuid(sbi->s_es->s_encrypt_pw_salt);
+			ext4_superblock_csum_set(sb);
+			unlock_buffer(sbi->s_sbh);
 			err = ext4_handle_dirty_metadata(handle, NULL,
 							 sbi->s_sbh);
 		pwsalt_err_journal:

fs/ext4/namei.c

@@ -2976,14 +2976,17 @@ int ext4_orphan_add(handle_t *handle, struct inode *inode)
 	    (le32_to_cpu(sbi->s_es->s_inodes_count))) {
 		/* Insert this inode at the head of the on-disk orphan list */
 		NEXT_ORPHAN(inode) = le32_to_cpu(sbi->s_es->s_last_orphan);
+		lock_buffer(sbi->s_sbh);
 		sbi->s_es->s_last_orphan = cpu_to_le32(inode->i_ino);
+		ext4_superblock_csum_set(sb);
+		unlock_buffer(sbi->s_sbh);
 		dirty = true;
 	}
 	list_add(&EXT4_I(inode)->i_orphan, &sbi->s_orphan);
 	mutex_unlock(&sbi->s_orphan_lock);
 
 	if (dirty) {
-		err = ext4_handle_dirty_super(handle, sb);
+		err = ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 		rc = ext4_mark_iloc_dirty(handle, inode, &iloc);
 		if (!err)
 			err = rc;
@@ -3059,9 +3062,12 @@ int ext4_orphan_del(handle_t *handle, struct inode *inode)
 			mutex_unlock(&sbi->s_orphan_lock);
 			goto out_brelse;
 		}
+		lock_buffer(sbi->s_sbh);
 		sbi->s_es->s_last_orphan = cpu_to_le32(ino_next);
+		ext4_superblock_csum_set(inode->i_sb);
+		unlock_buffer(sbi->s_sbh);
 		mutex_unlock(&sbi->s_orphan_lock);
-		err = ext4_handle_dirty_super(handle, inode->i_sb);
+		err = ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 	} else {
 		struct ext4_iloc iloc2;
 		struct inode *i_prev =
@@ -3593,9 +3599,6 @@ static int ext4_setent(handle_t *handle, struct ext4_renament *ent,
 			return retval2;
 		}
 	}
-	brelse(ent->bh);
-	ent->bh = NULL;
-
 	return retval;
 }
 
@@ -3794,6 +3797,7 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 		}
 	}
 
+	old_file_type = old.de->file_type;
 	if (IS_DIRSYNC(old.dir) || IS_DIRSYNC(new.dir))
 		ext4_handle_sync(handle);
 
@@ -3821,7 +3825,6 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 	force_reread = (new.dir->i_ino == old.dir->i_ino &&
 			ext4_test_inode_flag(new.dir, EXT4_INODE_INLINE_DATA));
 
-	old_file_type = old.de->file_type;
 	if (whiteout) {
 		/*
 		 * Do this before adding a new entry, so the old entry is sure
@@ -3919,15 +3922,19 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
 	retval = 0;
 
 end_rename:
-	brelse(old.dir_bh);
-	brelse(old.bh);
-	brelse(new.bh);
 	if (whiteout) {
-		if (retval)
+		if (retval) {
+			ext4_setent(handle, &old,
+				old.inode->i_ino, old_file_type);
 			drop_nlink(whiteout);
+		}
 		unlock_new_inode(whiteout);
 		iput(whiteout);
+
 	}
+	brelse(old.dir_bh);
+	brelse(old.bh);
+	brelse(new.bh);
 	if (handle)
 		ext4_journal_stop(handle);
 	return retval;

fs/ext4/resize.c

@@ -899,8 +899,11 @@ static int add_new_gdb(handle_t *handle, struct inode *inode,
 	EXT4_SB(sb)->s_gdb_count++;
 	ext4_kvfree_array_rcu(o_group_desc);
 
+	lock_buffer(EXT4_SB(sb)->s_sbh);
 	le16_add_cpu(&es->s_reserved_gdt_blocks, -1);
-	err = ext4_handle_dirty_super(handle, sb);
+	ext4_superblock_csum_set(sb);
+	unlock_buffer(EXT4_SB(sb)->s_sbh);
+	err = ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 	if (err)
 		ext4_std_error(sb, err);
 	return err;
@@ -1384,6 +1387,7 @@ static void ext4_update_super(struct super_block *sb,
 	reserved_blocks *= blocks_count;
 	do_div(reserved_blocks, 100);
 
+	lock_buffer(sbi->s_sbh);
 	ext4_blocks_count_set(es, ext4_blocks_count(es) + blocks_count);
 	ext4_free_blocks_count_set(es, ext4_free_blocks_count(es) + free_blocks);
 	le32_add_cpu(&es->s_inodes_count, EXT4_INODES_PER_GROUP(sb) *
@@ -1421,6 +1425,8 @@ static void ext4_update_super(struct super_block *sb,
 	 * active. */
 	ext4_r_blocks_count_set(es, ext4_r_blocks_count(es) +
 				reserved_blocks);
+	ext4_superblock_csum_set(sb);
+	unlock_buffer(sbi->s_sbh);
 
 	/* Update the free space counts */
 	percpu_counter_add(&sbi->s_freeclusters_counter,
@@ -1515,7 +1521,7 @@ static int ext4_flex_group_add(struct super_block *sb,
 
 	ext4_update_super(sb, flex_gd);
 
-	err = ext4_handle_dirty_super(handle, sb);
+	err = ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 
 exit_journal:
 	err2 = ext4_journal_stop(handle);
@@ -1717,15 +1723,18 @@ static int ext4_group_extend_no_check(struct super_block *sb,
 		goto errout;
 	}
 
+	lock_buffer(EXT4_SB(sb)->s_sbh);
 	ext4_blocks_count_set(es, o_blocks_count + add);
 	ext4_free_blocks_count_set(es, ext4_free_blocks_count(es) + add);
+	ext4_superblock_csum_set(sb);
+	unlock_buffer(EXT4_SB(sb)->s_sbh);
 	ext4_debug("freeing blocks %llu through %llu\n", o_blocks_count,
 		   o_blocks_count + add);
 	/* We add the blocks to the bitmap and set the group need init bit */
 	err = ext4_group_add_blocks(handle, sb, o_blocks_count, add);
 	if (err)
 		goto errout;
-	ext4_handle_dirty_super(handle, sb);
+	ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 	ext4_debug("freed blocks %llu through %llu\n", o_blocks_count,
 		   o_blocks_count + add);
 errout:
@@ -1874,12 +1883,15 @@ static int ext4_convert_meta_bg(struct super_block *sb, struct inode *inode)
 	if (err)
 		goto errout;
 
+	lock_buffer(sbi->s_sbh);
 	ext4_clear_feature_resize_inode(sb);
 	ext4_set_feature_meta_bg(sb);
 	sbi->s_es->s_first_meta_bg =
 		cpu_to_le32(num_desc_blocks(sb, sbi->s_groups_count));
+	ext4_superblock_csum_set(sb);
+	unlock_buffer(sbi->s_sbh);
 
-	err = ext4_handle_dirty_super(handle, sb);
+	err = ext4_handle_dirty_metadata(handle, NULL, sbi->s_sbh);
 	if (err) {
 		ext4_std_error(sb, err);
 		goto errout;

fs/ext4/super.c

@@ -65,7 +65,8 @@ static struct ratelimit_state ext4_mount_msg_ratelimit;
 static int ext4_load_journal(struct super_block *, struct ext4_super_block *,
 			     unsigned long journal_devnum);
 static int ext4_show_options(struct seq_file *seq, struct dentry *root);
-static int ext4_commit_super(struct super_block *sb, int sync);
+static void ext4_update_super(struct super_block *sb);
+static int ext4_commit_super(struct super_block *sb);
 static int ext4_mark_recovery_complete(struct super_block *sb,
 					struct ext4_super_block *es);
 static int ext4_clear_journal_err(struct super_block *sb,
@@ -586,15 +587,12 @@ static int ext4_errno_to_code(int errno)
 	return EXT4_ERR_UNKNOWN;
 }
 
-static void __save_error_info(struct super_block *sb, int error,
-			      __u32 ino, __u64 block,
-			      const char *func, unsigned int line)
+static void save_error_info(struct super_block *sb, int error,
+			    __u32 ino, __u64 block,
+			    const char *func, unsigned int line)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
 
-	EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
-	if (bdev_read_only(sb->s_bdev))
-		return;
 	/* We default to EFSCORRUPTED error... */
 	if (error == 0)
 		error = EFSCORRUPTED;
@@ -618,15 +616,6 @@ static void __save_error_info(struct super_block *sb, int error,
 	spin_unlock(&sbi->s_error_lock);
 }
 
-static void save_error_info(struct super_block *sb, int error,
-			    __u32 ino, __u64 block,
-			    const char *func, unsigned int line)
-{
-	__save_error_info(sb, error, ino, block, func, line);
-	if (!bdev_read_only(sb->s_bdev))
-		ext4_commit_super(sb, 1);
-}
-
 /* Deal with the reporting of failure conditions on a filesystem such as
  * inconsistencies detected or read IO failures.
  *
@@ -647,19 +636,40 @@ static void save_error_info(struct super_block *sb, int error,
  * used to deal with unrecoverable failures such as journal IO errors or ENOMEM
  * at a critical moment in log management.
  */
-static void ext4_handle_error(struct super_block *sb, bool force_ro)
+static void ext4_handle_error(struct super_block *sb, bool force_ro, int error,
+			      __u32 ino, __u64 block,
+			      const char *func, unsigned int line)
 {
 	journal_t *journal = EXT4_SB(sb)->s_journal;
+	bool continue_fs = !force_ro && test_opt(sb, ERRORS_CONT);
 
+	EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
 	if (test_opt(sb, WARN_ON_ERROR))
 		WARN_ON_ONCE(1);
 
-	if (sb_rdonly(sb) || (!force_ro && test_opt(sb, ERRORS_CONT)))
+	if (!continue_fs && !sb_rdonly(sb)) {
+		ext4_set_mount_flag(sb, EXT4_MF_FS_ABORTED);
+		if (journal)
+			jbd2_journal_abort(journal, -EIO);
+	}
+
+	if (!bdev_read_only(sb->s_bdev)) {
+		save_error_info(sb, error, ino, block, func, line);
+		/*
+		 * In case the fs should keep running, we need to writeout
+		 * superblock through the journal. Due to lock ordering
+		 * constraints, it may not be safe to do it right here so we
+		 * defer superblock flushing to a workqueue.
+		 */
+		if (continue_fs)
+			schedule_work(&EXT4_SB(sb)->s_error_work);
+		else
+			ext4_commit_super(sb);
+	}
+
+	if (sb_rdonly(sb) || continue_fs)
 		return;
 
-	ext4_set_mount_flag(sb, EXT4_MF_FS_ABORTED);
-	if (journal)
-		jbd2_journal_abort(journal, -EIO);
 	/*
 	 * We force ERRORS_RO behavior when system is rebooting. Otherwise we
 	 * could panic during 'reboot -f' as the underlying device got already
@@ -682,8 +692,39 @@ static void flush_stashed_error_work(struct work_struct *work)
 {
 	struct ext4_sb_info *sbi = container_of(work, struct ext4_sb_info,
 						s_error_work);
+	journal_t *journal = sbi->s_journal;
+	handle_t *handle;
 
-	ext4_commit_super(sbi->s_sb, 1);
+	/*
+	 * If the journal is still running, we have to write out superblock
+	 * through the journal to avoid collisions of other journalled sb
+	 * updates.
+	 *
+	 * We use directly jbd2 functions here to avoid recursing back into
+	 * ext4 error handling code during handling of previous errors.
+	 */
+	if (!sb_rdonly(sbi->s_sb) && journal) {
+		handle = jbd2_journal_start(journal, 1);
+		if (IS_ERR(handle))
+			goto write_directly;
+		if (jbd2_journal_get_write_access(handle, sbi->s_sbh)) {
+			jbd2_journal_stop(handle);
+			goto write_directly;
+		}
+		ext4_update_super(sbi->s_sb);
+		if (jbd2_journal_dirty_metadata(handle, sbi->s_sbh)) {
+			jbd2_journal_stop(handle);
+			goto write_directly;
+		}
+		jbd2_journal_stop(handle);
+		return;
+	}
+write_directly:
+	/*
+	 * Write through journal failed. Write sb directly to get error info
+	 * out and hope for the best.
+	 */
+	ext4_commit_super(sbi->s_sb);
 }
 
 #define ext4_error_ratelimit(sb)					\
@@ -710,8 +751,7 @@ void __ext4_error(struct super_block *sb, const char *function,
 		       sb->s_id, function, line, current->comm, &vaf);
 		va_end(args);
 	}
-	save_error_info(sb, error, 0, block, function, line);
-	ext4_handle_error(sb, force_ro);
+	ext4_handle_error(sb, force_ro, error, 0, block, function, line);
 }
 
 void __ext4_error_inode(struct inode *inode, const char *function,
@@ -741,9 +781,8 @@ void __ext4_error_inode(struct inode *inode, const char *function,
 			       current->comm, &vaf);
 		va_end(args);
 	}
-	save_error_info(inode->i_sb, error, inode->i_ino, block,
-			function, line);
-	ext4_handle_error(inode->i_sb, false);
+	ext4_handle_error(inode->i_sb, false, error, inode->i_ino, block,
+			  function, line);
 }
 
 void __ext4_error_file(struct file *file, const char *function,
@@ -780,9 +819,8 @@ void __ext4_error_file(struct file *file, const char *function,
 			       current->comm, path, &vaf);
 		va_end(args);
 	}
-	save_error_info(inode->i_sb, EFSCORRUPTED, inode->i_ino, block,
-			function, line);
-	ext4_handle_error(inode->i_sb, false);
+	ext4_handle_error(inode->i_sb, false, EFSCORRUPTED, inode->i_ino, block,
+			  function, line);
 }
 
 const char *ext4_decode_error(struct super_block *sb, int errno,
@@ -849,8 +887,7 @@ void __ext4_std_error(struct super_block *sb, const char *function,
 		       sb->s_id, function, line, errstr);
 	}
 
-	save_error_info(sb, -errno, 0, 0, function, line);
-	ext4_handle_error(sb, false);
+	ext4_handle_error(sb, false, -errno, 0, 0, function, line);
 }
 
 void __ext4_msg(struct super_block *sb,
@@ -944,13 +981,16 @@ __acquires(bitlock)
 	if (test_opt(sb, ERRORS_CONT)) {
 		if (test_opt(sb, WARN_ON_ERROR))
 			WARN_ON_ONCE(1);
-		__save_error_info(sb, EFSCORRUPTED, ino, block, function, line);
-		schedule_work(&EXT4_SB(sb)->s_error_work);
+		EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
+		if (!bdev_read_only(sb->s_bdev)) {
+			save_error_info(sb, EFSCORRUPTED, ino, block, function,
+					line);
+			schedule_work(&EXT4_SB(sb)->s_error_work);
+		}
 		return;
 	}
 	ext4_unlock_group(sb, grp);
-	save_error_info(sb, EFSCORRUPTED, ino, block, function, line);
-	ext4_handle_error(sb, false);
+	ext4_handle_error(sb, false, EFSCORRUPTED, ino, block, function, line);
 	/*
 	 * We only get here in the ERRORS_RO case; relocking the group
 	 * may be dangerous, but nothing bad will happen since the
@@ -1152,7 +1192,7 @@ static void ext4_put_super(struct super_block *sb)
 		es->s_state = cpu_to_le16(sbi->s_mount_state);
 	}
 	if (!sb_rdonly(sb))
-		ext4_commit_super(sb, 1);
+		ext4_commit_super(sb);
 
 	rcu_read_lock();
 	group_desc = rcu_dereference(sbi->s_group_desc);
@@ -2642,7 +2682,7 @@ static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
 	if (sbi->s_journal)
 		ext4_set_feature_journal_needs_recovery(sb);
 
-	err = ext4_commit_super(sb, 1);
+	err = ext4_commit_super(sb);
 done:
 	if (test_opt(sb, DEBUG))
 		printk(KERN_INFO "[EXT4 FS bs=%lu, gc=%u, "
@@ -4868,7 +4908,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	if (DUMMY_ENCRYPTION_ENABLED(sbi) && !sb_rdonly(sb) &&
 	    !ext4_has_feature_encrypt(sb)) {
 		ext4_set_feature_encrypt(sb);
-		ext4_commit_super(sb, 1);
+		ext4_commit_super(sb);
 	}
 
 	/*
@@ -5418,7 +5458,7 @@ static int ext4_load_journal(struct super_block *sb,
 		es->s_journal_dev = cpu_to_le32(journal_devnum);
 
 		/* Make sure we flush the recovery flag to disk. */
-		ext4_commit_super(sb, 1);
+		ext4_commit_super(sb);
 	}
 
 	return 0;
@@ -5428,16 +5468,14 @@ static int ext4_load_journal(struct super_block *sb,
 	return err;
 }
 
-static int ext4_commit_super(struct super_block *sb, int sync)
+/* Copy state of EXT4_SB(sb) into buffer for on-disk superblock */
+static void ext4_update_super(struct super_block *sb)
 {
 	struct ext4_sb_info *sbi = EXT4_SB(sb);
-	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
-	struct buffer_head *sbh = EXT4_SB(sb)->s_sbh;
-	int error = 0;
-
-	if (!sbh || block_device_ejected(sb))
-		return error;
+	struct ext4_super_block *es = sbi->s_es;
+	struct buffer_head *sbh = sbi->s_sbh;
 
+	lock_buffer(sbh);
 	/*
 	 * If the file system is mounted read-only, don't update the
 	 * superblock write time.  This avoids updating the superblock
@@ -5451,17 +5489,17 @@ static int ext4_commit_super(struct super_block *sb, int sync)
 	if (!(sb->s_flags & SB_RDONLY))
 		ext4_update_tstamp(es, s_wtime);
 	es->s_kbytes_written =
-		cpu_to_le64(EXT4_SB(sb)->s_kbytes_written +
+		cpu_to_le64(sbi->s_kbytes_written +
 		    ((part_stat_read(sb->s_bdev, sectors[STAT_WRITE]) -
-		      EXT4_SB(sb)->s_sectors_written_start) >> 1));
-	if (percpu_counter_initialized(&EXT4_SB(sb)->s_freeclusters_counter))
+		      sbi->s_sectors_written_start) >> 1));
+	if (percpu_counter_initialized(&sbi->s_freeclusters_counter))
 		ext4_free_blocks_count_set(es,
-			EXT4_C2B(EXT4_SB(sb), percpu_counter_sum_positive(
-				&EXT4_SB(sb)->s_freeclusters_counter)));
-	if (percpu_counter_initialized(&EXT4_SB(sb)->s_freeinodes_counter))
+			EXT4_C2B(sbi, percpu_counter_sum_positive(
+				&sbi->s_freeclusters_counter)));
+	if (percpu_counter_initialized(&sbi->s_freeinodes_counter))
 		es->s_free_inodes_count =
 			cpu_to_le32(percpu_counter_sum_positive(
-				&EXT4_SB(sb)->s_freeinodes_counter));
+				&sbi->s_freeinodes_counter));
 	/* Copy error information to the on-disk superblock */
 	spin_lock(&sbi->s_error_lock);
 	if (sbi->s_add_error_count > 0) {
@@ -5502,10 +5540,20 @@ static int ext4_commit_super(struct super_block *sb, int sync)
 	}
 	spin_unlock(&sbi->s_error_lock);
 
-	BUFFER_TRACE(sbh, "marking dirty");
 	ext4_superblock_csum_set(sb);
-	if (sync)
-		lock_buffer(sbh);
+	unlock_buffer(sbh);
+}
+
+static int ext4_commit_super(struct super_block *sb)
+{
+	struct buffer_head *sbh = EXT4_SB(sb)->s_sbh;
+	int error = 0;
+
+	if (!sbh || block_device_ejected(sb))
+		return error;
+
+	ext4_update_super(sb);
+
 	if (buffer_write_io_error(sbh) || !buffer_uptodate(sbh)) {
 		/*
 		 * Oh, dear.  A previous attempt to write the
@@ -5520,17 +5568,15 @@ static int ext4_commit_super(struct super_block *sb, int sync)
 		clear_buffer_write_io_error(sbh);
 		set_buffer_uptodate(sbh);
 	}
+	BUFFER_TRACE(sbh, "marking dirty");
 	mark_buffer_dirty(sbh);
-	if (sync) {
-		unlock_buffer(sbh);
-		error = __sync_dirty_buffer(sbh,
-			REQ_SYNC | (test_opt(sb, BARRIER) ? REQ_FUA : 0));
-		if (buffer_write_io_error(sbh)) {
-			ext4_msg(sb, KERN_ERR, "I/O error while writing "
-			       "superblock");
-			clear_buffer_write_io_error(sbh);
-			set_buffer_uptodate(sbh);
-		}
+	error = __sync_dirty_buffer(sbh,
+		REQ_SYNC | (test_opt(sb, BARRIER) ? REQ_FUA : 0));
+	if (buffer_write_io_error(sbh)) {
+		ext4_msg(sb, KERN_ERR, "I/O error while writing "
+		       "superblock");
+		clear_buffer_write_io_error(sbh);
+		set_buffer_uptodate(sbh);
 	}
 	return error;
 }
@@ -5561,7 +5607,7 @@ static int ext4_mark_recovery_complete(struct super_block *sb,
 
 	if (ext4_has_feature_journal_needs_recovery(sb) && sb_rdonly(sb)) {
 		ext4_clear_feature_journal_needs_recovery(sb);
-		ext4_commit_super(sb, 1);
+		ext4_commit_super(sb);
 	}
 out:
 	jbd2_journal_unlock_updates(journal);
@@ -5603,7 +5649,7 @@ static int ext4_clear_journal_err(struct super_block *sb,
 
 		EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
 		es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
-		ext4_commit_super(sb, 1);
+		ext4_commit_super(sb);
 
 		jbd2_journal_clear_err(journal);
 		jbd2_journal_update_sb_errno(journal);
@@ -5705,7 +5751,7 @@ static int ext4_freeze(struct super_block *sb)
 		ext4_clear_feature_journal_needs_recovery(sb);
 	}
 
-	error = ext4_commit_super(sb, 1);
+	error = ext4_commit_super(sb);
 out:
 	if (journal)
 		/* we rely on upper layer to stop further updates */
@@ -5727,7 +5773,7 @@ static int ext4_unfreeze(struct super_block *sb)
 		ext4_set_feature_journal_needs_recovery(sb);
 	}
 
-	ext4_commit_super(sb, 1);
+	ext4_commit_super(sb);
 	return 0;
 }
 
@@ -5987,7 +6033,7 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 	}
 
 	if (sbi->s_journal == NULL && !(old_sb_flags & SB_RDONLY)) {
-		err = ext4_commit_super(sb, 1);
+		err = ext4_commit_super(sb);
 		if (err)
 			goto restore_opts;
 	}

fs/ext4/xattr.c

@@ -792,8 +792,11 @@ static void ext4_xattr_update_super_block(handle_t *handle,
 
 	BUFFER_TRACE(EXT4_SB(sb)->s_sbh, "get_write_access");
 	if (ext4_journal_get_write_access(handle, EXT4_SB(sb)->s_sbh) == 0) {
+		lock_buffer(EXT4_SB(sb)->s_sbh);
 		ext4_set_feature_xattr(sb);
-		ext4_handle_dirty_super(handle, sb);
+		ext4_superblock_csum_set(sb);
+		unlock_buffer(EXT4_SB(sb)->s_sbh);
+		ext4_handle_dirty_metadata(handle, NULL, EXT4_SB(sb)->s_sbh);
 	}
 }