logfs: initialize superblock entries earlier
Linus Torvalds [Tue, 3 May 2011 23:10:25 +0000 (16:10 -0700)]
In particular, s_freeing_list needs to be initialized early, since it is
used on some of the error paths when mounts fail.  The mapping inode,
for example, would be initialized and then free'd on an error path
before s_freeing_list was initialized, but the inode drop operation
needs the s_freeing_list to be set up.

Normally you'd never see this, because not only is logfs fairly rare,
but a successful mount will never have any issues.

Reported-by: werner <w.landgraf@ru.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/logfs/super.c

index 33435e4..ce03a18 100644 (file)
@@ -480,10 +480,6 @@ static int logfs_read_sb(struct super_block *sb, int read_only)
                        !read_only)
                return -EIO;
 
-       mutex_init(&super->s_dirop_mutex);
-       mutex_init(&super->s_object_alias_mutex);
-       INIT_LIST_HEAD(&super->s_freeing_list);
-
        ret = logfs_init_rw(sb);
        if (ret)
                return ret;
@@ -601,6 +597,10 @@ static struct dentry *logfs_mount(struct file_system_type *type, int flags,
        if (!super)
                return ERR_PTR(-ENOMEM);
 
+       mutex_init(&super->s_dirop_mutex);
+       mutex_init(&super->s_object_alias_mutex);
+       INIT_LIST_HEAD(&super->s_freeing_list);
+
        if (!devname)
                err = logfs_get_sb_bdev(super, type, devname);
        else if (strncmp(devname, "mtd", 3))