Merge branch 'master' into for-2.6.35

Conflicts:
	fs/ext3/fsync.c

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>

4 files changed, 63 insertions, 42 deletions

diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c
index a177122a1b25..4a32511f4ded 100644
--- a/fs/ext3/balloc.c
+++ b/fs/ext3/balloc.c
@@ -1584,6 +1584,12 @@ retry_alloc:
 			goto io_error;
 		free_blocks = le16_to_cpu(gdp->bg_free_blocks_count);
 		/*
+		 * skip this group (and avoid loading bitmap) if there
+		 * are no free blocks
+		 */
+		if (!free_blocks)
+			continue;
+		/*
 		 * skip this group if the number of
 		 * free blocks is less than half of the reservation
 		 * window size.
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index 9492f6003ef9..fcf7487734b6 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -48,7 +48,7 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 	struct inode *inode = dentry->d_inode;
 	struct ext3_inode_info *ei = EXT3_I(inode);
 	journal_t *journal = EXT3_SB(inode->i_sb)->s_journal;
-	int ret = 0;
+	int ret, needs_barrier = 0;
 	tid_t commit_tid;
 
 	if (inode->i_sb->s_flags & MS_RDONLY)
@@ -70,29 +70,27 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync)
 	 *  (they were dirtied by commit).  But that's OK - the blocks are
 	 *  safe in-journal, which is all fsync() needs to ensure.
 	 */
-	if (ext3_should_journal_data(inode)) {
-		ret = ext3_force_commit(inode->i_sb);
-		goto out;
-	}
+	if (ext3_should_journal_data(inode))
+		return ext3_force_commit(inode->i_sb);
 
 	if (datasync)
 		commit_tid = atomic_read(&ei->i_datasync_tid);
 	else
 		commit_tid = atomic_read(&ei->i_sync_tid);
 
-	if (log_start_commit(journal, commit_tid)) {
-		log_wait_commit(journal, commit_tid);
-		goto out;
-	}
+	if (test_opt(inode->i_sb, BARRIER) &&
+	    !journal_trans_will_send_data_barrier(journal, commit_tid))
+		needs_barrier = 1;
+	log_start_commit(journal, commit_tid);
+	ret = log_wait_commit(journal, commit_tid);
 
 	/*
 	 * In case we didn't commit a transaction, we have to flush
 	 * disk caches manually so that data really is on persistent
 	 * storage
 	 */
-	if (test_opt(inode->i_sb, BARRIER))
+	if (needs_barrier)
 		blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL,
 				BLKDEV_IFL_WAIT);
-out:
 	return ret;
 }
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index ea33bdf0a300..735f0190ec2a 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -3151,7 +3151,7 @@ int ext3_setattr(struct dentry *dentry, struct iattr *attr)
 	if (error)
 		return error;
 
-	if (ia_valid & ATTR_SIZE)
+	if (is_quota_modification(inode, attr))
 		dquot_initialize(inode);
 	if ((ia_valid & ATTR_UID && attr->ia_uid != inode->i_uid) ||
 		(ia_valid & ATTR_GID && attr->ia_gid != inode->i_gid)) {
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 1bee604cc6cd..0fc1293d0e96 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -653,8 +653,12 @@ static int ext3_show_options(struct seq_file *seq, struct vfsmount *vfs)
 		seq_printf(seq, ",commit=%u",
 			   (unsigned) (sbi->s_commit_interval / HZ));
 	}
-	if (test_opt(sb, BARRIER))
-		seq_puts(seq, ",barrier=1");
+
+	/*
+	 * Always display barrier state so it's clear what the status is.
+	 */
+	seq_puts(seq, ",barrier=");
+	seq_puts(seq, test_opt(sb, BARRIER) ? "1" : "0");
 	if (test_opt(sb, NOBH))
 		seq_puts(seq, ",nobh");
 
@@ -810,8 +814,8 @@ enum {
 	Opt_data_err_abort, Opt_data_err_ignore,
 	Opt_usrjquota, Opt_grpjquota, Opt_offusrjquota, Opt_offgrpjquota,
 	Opt_jqfmt_vfsold, Opt_jqfmt_vfsv0, Opt_jqfmt_vfsv1, Opt_quota,
-	Opt_noquota, Opt_ignore, Opt_barrier, Opt_err, Opt_resize,
-	Opt_usrquota, Opt_grpquota
+	Opt_noquota, Opt_ignore, Opt_barrier, Opt_nobarrier, Opt_err,
+	Opt_resize, Opt_usrquota, Opt_grpquota
 };
 
 static const match_table_t tokens = {
@@ -865,6 +869,8 @@ static const match_table_t tokens = {
 	{Opt_quota, "quota"},
 	{Opt_usrquota, "usrquota"},
 	{Opt_barrier, "barrier=%u"},
+	{Opt_barrier, "barrier"},
+	{Opt_nobarrier, "nobarrier"},
 	{Opt_resize, "resize"},
 	{Opt_err, NULL},
 };
@@ -967,7 +973,11 @@ static int parse_options (char *options, struct super_block *sb,
 		int token;
 		if (!*p)
 			continue;
-
+		/*
+		 * Initialize args struct so we know whether arg was
+		 * found; some options take optional arguments.
+		 */
+		args[0].to = args[0].from = 0;
 		token = match_token(p, tokens, args);
 		switch (token) {
 		case Opt_bsd_df:
@@ -1215,9 +1225,15 @@ set_qf_format:
 		case Opt_abort:
 			set_opt(sbi->s_mount_opt, ABORT);
 			break;
+		case Opt_nobarrier:
+			clear_opt(sbi->s_mount_opt, BARRIER);
+			break;
 		case Opt_barrier:
-			if (match_int(&args[0], &option))
-				return 0;
+			if (args[0].from) {
+				if (match_int(&args[0], &option))
+					return 0;
+			} else
+				option = 1;	/* No argument, default to 1 */
 			if (option)
 				set_opt(sbi->s_mount_opt, BARRIER);
 			else
@@ -1890,21 +1906,6 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	get_random_bytes(&sbi->s_next_generation, sizeof(u32));
 	spin_lock_init(&sbi->s_next_gen_lock);
 
-	err = percpu_counter_init(&sbi->s_freeblocks_counter,
-			ext3_count_free_blocks(sb));
-	if (!err) {
-		err = percpu_counter_init(&sbi->s_freeinodes_counter,
-				ext3_count_free_inodes(sb));
-	}
-	if (!err) {
-		err = percpu_counter_init(&sbi->s_dirs_counter,
-				ext3_count_dirs(sb));
-	}
-	if (err) {
-		ext3_msg(sb, KERN_ERR, "error: insufficient memory");
-		goto failed_mount3;
-	}
-
 	/* per fileystem reservation list head & lock */
 	spin_lock_init(&sbi->s_rsv_window_lock);
 	sbi->s_rsv_window_root = RB_ROOT;
@@ -1945,15 +1946,29 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	if (!test_opt(sb, NOLOAD) &&
 	    EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_HAS_JOURNAL)) {
 		if (ext3_load_journal(sb, es, journal_devnum))
-			goto failed_mount3;
+			goto failed_mount2;
 	} else if (journal_inum) {
 		if (ext3_create_journal(sb, es, journal_inum))
-			goto failed_mount3;
+			goto failed_mount2;
 	} else {
 		if (!silent)
 			ext3_msg(sb, KERN_ERR,
 				"error: no journal found. "
 				"mounting ext3 over ext2?");
+		goto failed_mount2;
+	}
+	err = percpu_counter_init(&sbi->s_freeblocks_counter,
+			ext3_count_free_blocks(sb));
+	if (!err) {
+		err = percpu_counter_init(&sbi->s_freeinodes_counter,
+				ext3_count_free_inodes(sb));
+	}
+	if (!err) {
+		err = percpu_counter_init(&sbi->s_dirs_counter,
+				ext3_count_dirs(sb));
+	}
+	if (err) {
+		ext3_msg(sb, KERN_ERR, "error: insufficient memory");
 		goto failed_mount3;
 	}
 
@@ -1978,7 +1993,7 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 			ext3_msg(sb, KERN_ERR,
 				"error: journal does not support "
 				"requested data journaling mode");
-			goto failed_mount4;
+			goto failed_mount3;
 		}
 	default:
 		break;
@@ -2001,19 +2016,19 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	if (IS_ERR(root)) {
 		ext3_msg(sb, KERN_ERR, "error: get root inode failed");
 		ret = PTR_ERR(root);
-		goto failed_mount4;
+		goto failed_mount3;
 	}
 	if (!S_ISDIR(root->i_mode) || !root->i_blocks || !root->i_size) {
 		iput(root);
 		ext3_msg(sb, KERN_ERR, "error: corrupt root inode, run e2fsck");
-		goto failed_mount4;
+		goto failed_mount3;
 	}
 	sb->s_root = d_alloc_root(root);
 	if (!sb->s_root) {
 		ext3_msg(sb, KERN_ERR, "error: get root dentry failed");
 		iput(root);
 		ret = -ENOMEM;
-		goto failed_mount4;
+		goto failed_mount3;
 	}
 
 	ext3_setup_super (sb, es, sb->s_flags & MS_RDONLY);
@@ -2039,12 +2054,11 @@ cantfind_ext3:
 		       sb->s_id);
 	goto failed_mount;
 
-failed_mount4:
-	journal_destroy(sbi->s_journal);
 failed_mount3:
 	percpu_counter_destroy(&sbi->s_freeblocks_counter);
 	percpu_counter_destroy(&sbi->s_freeinodes_counter);
 	percpu_counter_destroy(&sbi->s_dirs_counter);
+	journal_destroy(sbi->s_journal);
 failed_mount2:
 	for (i = 0; i < db_count; i++)
 		brelse(sbi->s_group_desc[i]);
@@ -2278,6 +2292,9 @@ static int ext3_load_journal(struct super_block *sb,
 			return -EINVAL;
 	}
 
+	if (!(journal->j_flags & JFS_BARRIER))
+		printk(KERN_INFO "EXT3-fs: barriers not enabled\n");
+
 	if (!really_read_only && test_opt(sb, UPDATE_JOURNAL)) {
 		err = journal_update_format(journal);
 		if (err)  {