ext4: Allow ext4 to run without a journal

A few weeks ago I posted a patch for discussion that allowed ext4 to run
without a journal.  Since that time I've integrated the excellent
comments from Andreas and fixed several serious bugs.  We're currently
running with this patch and generating some performance numbers against
both ext2 (with backported reservations code) and ext4 with and without
a journal.  It just so happens that running without a journal is
slightly faster for most everything.

We did
	iozone -T -t 4 s 2g -r 256k -T -I -i0 -i1 -i2

which creates 4 threads, each of which create and do reads and writes on
a 2G file, with a buffer size of 256K, using O_DIRECT for all file opens
to bypass the page cache.  Results:

                     ext2        ext4, default   ext4, no journal
  initial writes   13.0 MB/s        15.4 MB/s          15.7 MB/s
  rewrites         13.1 MB/s        15.6 MB/s          15.9 MB/s
  reads            15.2 MB/s        16.9 MB/s          17.2 MB/s
  re-reads         15.3 MB/s        16.9 MB/s          17.2 MB/s
  random readers    5.6 MB/s         5.6 MB/s           5.7 MB/s
  random writers    5.1 MB/s         5.3 MB/s           5.4 MB/s 

So it seems that, so far, this was a useful exercise.

Signed-off-by: Frank Mayhar <fmayhar@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

1 files changed, 140 insertions, 67 deletions

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 8a0ae883f56..9b9076d9c4f 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -136,13 +136,19 @@ handle_t *ext4_journal_start_sb(struct super_block *sb, int nblocks)
 	 * backs (eg. EIO in the commit thread), then we still need to
 	 * take the FS itself readonly cleanly. */
 	journal = EXT4_SB(sb)->s_journal;
-	if (is_journal_aborted(journal)) {
-		ext4_abort(sb, __func__,
-			   "Detected aborted journal");
-		return ERR_PTR(-EROFS);
+	if (journal) {
+		if (is_journal_aborted(journal)) {
+			ext4_abort(sb, __func__,
+				   "Detected aborted journal");
+			return ERR_PTR(-EROFS);
+		}
+		return jbd2_journal_start(journal, nblocks);
 	}
-
-	return jbd2_journal_start(journal, nblocks);
+	/*
+	 * We're not journaling, return the appropriate indication.
+	 */
+	current->journal_info = EXT4_NOJOURNAL_HANDLE;
+	return current->journal_info;
 }
 
 /*
@@ -157,6 +163,14 @@ int __ext4_journal_stop(const char *where, handle_t *handle)
 	int err;
 	int rc;
 
+	if (!ext4_handle_valid(handle)) {
+		/*
+		 * Do this here since we don't call jbd2_journal_stop() in
+		 * no-journal mode.
+		 */
+		current->journal_info = NULL;
+		return 0;
+	}
 	sb = handle->h_transaction->t_journal->j_private;
 	err = handle->h_err;
 	rc = jbd2_journal_stop(handle);
@@ -174,6 +188,8 @@ void ext4_journal_abort_handle(const char *caller, const char *err_fn,
 	char nbuf[16];
 	const char *errstr = ext4_decode_error(NULL, err, nbuf);
 
+	BUG_ON(!ext4_handle_valid(handle));
+
 	if (bh)
 		BUFFER_TRACE(bh, "abort");
 
@@ -448,11 +464,13 @@ static void ext4_put_super(struct super_block *sb)
 	ext4_mb_release(sb);
 	ext4_ext_release(sb);
 	ext4_xattr_put_super(sb);
-	err = jbd2_journal_destroy(sbi->s_journal);
-	sbi->s_journal = NULL;
-	if (err < 0)
-		ext4_abort(sb, __func__, "Couldn't clean up the journal");
-
+	if (sbi->s_journal) {
+		err = jbd2_journal_destroy(sbi->s_journal);
+		sbi->s_journal = NULL;
+		if (err < 0)
+			ext4_abort(sb, __func__,
+				   "Couldn't clean up the journal");
+	}
 	if (!(sb->s_flags & MS_RDONLY)) {
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 		es->s_state = cpu_to_le16(sbi->s_mount_state);
@@ -522,6 +540,11 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
 	memset(&ei->i_cached_extent, 0, sizeof(struct ext4_ext_cache));
 	INIT_LIST_HEAD(&ei->i_prealloc_list);
 	spin_lock_init(&ei->i_prealloc_lock);
+	/*
+	 * Note:  We can be called before EXT4_SB(sb)->s_journal is set,
+	 * therefore it can be null here.  Don't check it, just initialize
+	 * jinode.
+	 */
 	jbd2_journal_init_jbd_inode(&ei->jinode, &ei->vfs_inode);
 	ei->i_reserved_data_blocks = 0;
 	ei->i_reserved_meta_blocks = 0;
@@ -588,7 +611,8 @@ static void ext4_clear_inode(struct inode *inode)
 	}
 #endif
 	ext4_discard_preallocations(inode);
-	jbd2_journal_release_jbd_inode(EXT4_SB(inode->i_sb)->s_journal,
+	if (EXT4_JOURNAL(inode))
+		jbd2_journal_release_jbd_inode(EXT4_SB(inode->i_sb)->s_journal,
 				       &EXT4_I(inode)->jinode);
 }
 
@@ -1406,20 +1430,15 @@ static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
 		printk(KERN_WARNING
 		       "EXT4-fs warning: checktime reached, "
 		       "running e2fsck is recommended\n");
-#if 0
-		/* @@@ We _will_ want to clear the valid bit if we find
-		 * inconsistencies, to force a fsck at reboot.  But for
-		 * a plain journaled filesystem we can keep it set as
-		 * valid forever! :)
-		 */
-	es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
-#endif
+	if (!sbi->s_journal) 
+		es->s_state &= cpu_to_le16(~EXT4_VALID_FS);
 	if (!(__s16) le16_to_cpu(es->s_max_mnt_count))
 		es->s_max_mnt_count = cpu_to_le16(EXT4_DFL_MAX_MNT_COUNT);
 	le16_add_cpu(&es->s_mnt_count, 1);
 	es->s_mtime = cpu_to_le32(get_seconds());
 	ext4_update_dynamic_rev(sb);
-	EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
+	if (sbi->s_journal)
+		EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
 
 	ext4_commit_super(sb, es, 1);
 	if (test_opt(sb, DEBUG))
@@ -1431,9 +1450,13 @@ static int ext4_setup_super(struct super_block *sb, struct ext4_super_block *es,
 			EXT4_INODES_PER_GROUP(sb),
 			sbi->s_mount_opt);
 
-	printk(KERN_INFO "EXT4 FS on %s, %s journal on %s\n",
-	       sb->s_id, EXT4_SB(sb)->s_journal->j_inode ? "internal" :
-	       "external", EXT4_SB(sb)->s_journal->j_devname);
+	if (EXT4_SB(sb)->s_journal) {
+		printk(KERN_INFO "EXT4 FS on %s, %s journal on %s\n",
+		       sb->s_id, EXT4_SB(sb)->s_journal->j_inode ? "internal" :
+		       "external", EXT4_SB(sb)->s_journal->j_devname);
+	} else {
+		printk(KERN_INFO "EXT4 FS on %s, no journal\n", sb->s_id);
+	}
 	return res;
 }
 
@@ -1867,6 +1890,7 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	unsigned long def_mount_opts;
 	struct inode *root;
 	char *cp;
+	const char *descr;
 	int ret = -EINVAL;
 	int blocksize;
 	int db_count;
@@ -2278,21 +2302,23 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 				EXT4_SB(sb)->s_mount_state |= EXT4_ERROR_FS;
 				es->s_state |= cpu_to_le16(EXT4_ERROR_FS);
 				ext4_commit_super(sb, es, 1);
-				printk(KERN_CRIT
-				       "EXT4-fs (device %s): mount failed\n",
-				      sb->s_id);
 				goto failed_mount4;
 			}
 		}
 	} else if (journal_inum) {
 		if (ext4_create_journal(sb, es, journal_inum))
 			goto failed_mount3;
+	} else if (test_opt(sb, NOLOAD) && !(sb->s_flags & MS_RDONLY) &&
+	      EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER)) {
+		printk(KERN_ERR "EXT4-fs: required journal recovery "
+		       "suppressed and not mounted read-only\n");
+		goto failed_mount4;
 	} else {
-		if (!silent)
-			printk(KERN_ERR
-			       "ext4: No journal on filesystem on %s\n",
-			       sb->s_id);
-		goto failed_mount3;
+		clear_opt(sbi->s_mount_opt, DATA_FLAGS);
+		set_opt(sbi->s_mount_opt, WRITEBACK_DATA);
+		sbi->s_journal = NULL;
+		needs_recovery = 0;
+		goto no_journal;
 	}
 
 	if (ext4_blocks_count(es) > 0xffffffffULL &&
@@ -2344,6 +2370,8 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 		break;
 	}
 
+no_journal:
+
 	if (test_opt(sb, NOBH)) {
 		if (!(test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_WRITEBACK_DATA)) {
 			printk(KERN_WARNING "EXT4-fs: Ignoring nobh option - "
@@ -2428,13 +2456,22 @@ static int ext4_fill_super(struct super_block *sb, void *data, int silent)
 	EXT4_SB(sb)->s_mount_state |= EXT4_ORPHAN_FS;
 	ext4_orphan_cleanup(sb, es);
 	EXT4_SB(sb)->s_mount_state &= ~EXT4_ORPHAN_FS;
-	if (needs_recovery)
+	if (needs_recovery) {
 		printk(KERN_INFO "EXT4-fs: recovery complete.\n");
-	ext4_mark_recovery_complete(sb, es);
-	printk(KERN_INFO "EXT4-fs: mounted filesystem with %s data mode.\n",
-	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA ? "journal":
-	       test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA ? "ordered":
-	       "writeback");
+		ext4_mark_recovery_complete(sb, es);
+	}
+	if (EXT4_SB(sb)->s_journal) {
+		if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA)
+			descr = " journalled data mode";
+		else if (test_opt(sb, DATA_FLAGS) == EXT4_MOUNT_ORDERED_DATA)
+			descr = " ordered data mode";
+		else
+			descr = " writeback data mode";
+	} else
+		descr = "out journal";
+
+	printk(KERN_INFO "EXT4-fs: mounted filesystem %s with%s\n",
+	       sb->s_id, descr);
 
 	lock_kernel();
 	return 0;
@@ -2446,8 +2483,11 @@ cantfind_ext4:
 	goto failed_mount;
 
 failed_mount4:
-	jbd2_journal_destroy(sbi->s_journal);
-	sbi->s_journal = NULL;
+	printk(KERN_ERR "EXT4-fs (device %s): mount failed\n", sb->s_id);
+	if (sbi->s_journal) {
+		jbd2_journal_destroy(sbi->s_journal);
+		sbi->s_journal = NULL;
+	}
 failed_mount3:
 	percpu_counter_destroy(&sbi->s_freeblocks_counter);
 	percpu_counter_destroy(&sbi->s_freeinodes_counter);
@@ -2508,6 +2548,8 @@ static journal_t *ext4_get_journal(struct super_block *sb,
 	struct inode *journal_inode;
 	journal_t *journal;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	/* First, test for the existence of a valid inode on disk.  Bad
 	 * things happen if we iget() an unused inode, as the subsequent
 	 * iput() will try to delete it. */
@@ -2556,6 +2598,8 @@ static journal_t *ext4_get_dev_journal(struct super_block *sb,
 	struct ext4_super_block *es;
 	struct block_device *bdev;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	bdev = ext4_blkdev_get(j_dev);
 	if (bdev == NULL)
 		return NULL;
@@ -2643,6 +2687,8 @@ static int ext4_load_journal(struct super_block *sb,
 	int err = 0;
 	int really_read_only;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	if (journal_devnum &&
 	    journal_devnum != le32_to_cpu(es->s_journal_dev)) {
 		printk(KERN_INFO "EXT4-fs: external journal device major/minor "
@@ -2817,6 +2863,10 @@ static void ext4_mark_recovery_complete(struct super_block *sb,
 {
 	journal_t *journal = EXT4_SB(sb)->s_journal;
 
+	if (!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL)) {
+		BUG_ON(journal != NULL);
+		return;
+	}
 	jbd2_journal_lock_updates(journal);
 	if (jbd2_journal_flush(journal) < 0)
 		goto out;
@@ -2846,6 +2896,8 @@ static void ext4_clear_journal_err(struct super_block *sb,
 	int j_errno;
 	const char *errstr;
 
+	BUG_ON(!EXT4_HAS_COMPAT_FEATURE(sb, EXT4_FEATURE_COMPAT_HAS_JOURNAL));
+
 	journal = EXT4_SB(sb)->s_journal;
 
 	/*
@@ -2878,14 +2930,17 @@ static void ext4_clear_journal_err(struct super_block *sb,
 int ext4_force_commit(struct super_block *sb)
 {
 	journal_t *journal;
-	int ret;
+	int ret = 0;
 
 	if (sb->s_flags & MS_RDONLY)
 		return 0;
 
 	journal = EXT4_SB(sb)->s_journal;
-	sb->s_dirt = 0;
-	ret = ext4_journal_force_commit(journal);
+	if (journal) {
+		sb->s_dirt = 0;
+		ret = ext4_journal_force_commit(journal);
+	}
+
 	return ret;
 }
 
@@ -2897,9 +2952,13 @@ int ext4_force_commit(struct super_block *sb)
  */
 static void ext4_write_super(struct super_block *sb)
 {
-	if (mutex_trylock(&sb->s_lock) != 0)
-		BUG();
-	sb->s_dirt = 0;
+	if (EXT4_SB(sb)->s_journal) {
+		if (mutex_trylock(&sb->s_lock) != 0)
+			BUG();
+		sb->s_dirt = 0;
+	} else {
+		ext4_commit_super(sb, EXT4_SB(sb)->s_es, 1);
+	}
 }
 
 static int ext4_sync_fs(struct super_block *sb, int wait)
@@ -2908,10 +2967,14 @@ static int ext4_sync_fs(struct super_block *sb, int wait)
 
 	trace_mark(ext4_sync_fs, "dev %s wait %d", sb->s_id, wait);
 	sb->s_dirt = 0;
-	if (wait)
-		ret = ext4_force_commit(sb);
-	else
-		jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
+	if (EXT4_SB(sb)->s_journal) {
+		if (wait)
+			ret = ext4_force_commit(sb);
+		else
+ 			jbd2_journal_start_commit(EXT4_SB(sb)->s_journal, NULL);
+	} else {
+		ext4_commit_super(sb, EXT4_SB(sb)->s_es, wait);
+	}
 	return ret;
 }
 
@@ -2926,15 +2989,17 @@ static void ext4_write_super_lockfs(struct super_block *sb)
 	if (!(sb->s_flags & MS_RDONLY)) {
 		journal_t *journal = EXT4_SB(sb)->s_journal;
 
-		/* Now we set up the journal barrier. */
-		jbd2_journal_lock_updates(journal);
+		if (journal) {
+			/* Now we set up the journal barrier. */
+			jbd2_journal_lock_updates(journal);
 
-		/*
-		 * We don't want to clear needs_recovery flag when we failed
-		 * to flush the journal.
-		 */
-		if (jbd2_journal_flush(journal) < 0)
-			return;
+			/*
+			 * We don't want to clear needs_recovery flag when we
+			 * failed to flush the journal.
+			 */
+			if (jbd2_journal_flush(journal) < 0)
+				return;
+		}
 
 		/* Journal blocked and flushed, clear needs_recovery flag. */
 		EXT4_CLEAR_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
@@ -2948,7 +3013,7 @@ static void ext4_write_super_lockfs(struct super_block *sb)
  */
 static void ext4_unlockfs(struct super_block *sb)
 {
-	if (!(sb->s_flags & MS_RDONLY)) {
+	if (EXT4_SB(sb)->s_journal && !(sb->s_flags & MS_RDONLY)) {
 		lock_super(sb);
 		/* Reser the needs_recovery flag before the fs is unlocked. */
 		EXT4_SET_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_RECOVER);
@@ -2999,7 +3064,8 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 
 	es = sbi->s_es;
 
-	ext4_init_journal_params(sb, sbi->s_journal);
+	if (sbi->s_journal)
+		ext4_init_journal_params(sb, sbi->s_journal);
 
 	if ((*flags & MS_RDONLY) != (sb->s_flags & MS_RDONLY) ||
 		n_blocks_count > ext4_blocks_count(es)) {
@@ -3028,9 +3094,11 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			 * We have to unlock super so that we can wait for
 			 * transactions.
 			 */
-			unlock_super(sb);
-			ext4_mark_recovery_complete(sb, es);
-			lock_super(sb);
+			if (sbi->s_journal) {
+				unlock_super(sb);
+				ext4_mark_recovery_complete(sb, es);
+				lock_super(sb);
+			}
 		} else {
 			__le32 ret;
 			if ((ret = EXT4_HAS_RO_COMPAT_FEATURE(sb,
@@ -3084,7 +3152,8 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 			 * been changed by e2fsck since we originally mounted
 			 * the partition.)
 			 */
-			ext4_clear_journal_err(sb, es);
+			if (sbi->s_journal)
+				ext4_clear_journal_err(sb, es);
 			sbi->s_mount_state = le16_to_cpu(es->s_state);
 			if ((err = ext4_group_extend(sb, es, n_blocks_count)))
 				goto restore_opts;
@@ -3092,6 +3161,9 @@ static int ext4_remount(struct super_block *sb, int *flags, char *data)
 				sb->s_flags &= ~MS_RDONLY;
 		}
 	}
+	if (sbi->s_journal == NULL)
+		ext4_commit_super(sb, es, 1);
+
 #ifdef CONFIG_QUOTA
 	/* Release old quota file names */
 	for (i = 0; i < MAXQUOTAS; i++)
@@ -3368,7 +3440,8 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
 	 * When we journal data on quota file, we have to flush journal to see
 	 * all updates to the file when we bypass pagecache...
 	 */
-	if (ext4_should_journal_data(path.dentry->d_inode)) {
+	if (EXT4_SB(sb)->s_journal &&
+	    ext4_should_journal_data(path.dentry->d_inode)) {
 		/*
 		 * We don't need to lock updates but journal_flush() could
 		 * otherwise be livelocked...
@@ -3442,7 +3515,7 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type,
 	struct buffer_head *bh;
 	handle_t *handle = journal_current_handle();
 
-	if (!handle) {
+	if (EXT4_SB(sb)->s_journal && !handle) {
 		printk(KERN_WARNING "EXT4-fs: Quota write (off=%llu, len=%llu)"
 			" cancelled because transaction is not started.\n",
 			(unsigned long long)off, (unsigned long long)len);
@@ -3467,7 +3540,7 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type,
 		flush_dcache_page(bh->b_page);
 		unlock_buffer(bh);
 		if (journal_quota)
-			err = ext4_journal_dirty_metadata(handle, bh);
+			err = ext4_handle_dirty_metadata(handle, NULL, bh);
 		else {
 			/* Always do at least ordered writes for quotas */
 			err = ext4_jbd2_file_inode(handle, inode);