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()

1 files changed, 118 insertions, 72 deletions

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 21121787c874..9a6f9875aa34 100644
--- a/fs/ext4/super.c
+++ b/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 @@ no_journal:
 	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 @@ err_out:
 	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;
 	}