btrfs: qgroup: Fix possible leak in btrfs_add_qgroup_relation()
authorChristian Engelmayer <cengelma@gmx.at>
Sat, 2 May 2015 15:19:55 +0000 (17:19 +0200)
committerChris Mason <clm@fb.com>
Wed, 3 Jun 2015 02:34:35 +0000 (19:34 -0700)
Commit 9c8b35b1ba21 ("btrfs: quota: Automatically update related qgroups or
mark INCONSISTENT flags when assigning/deleting a qgroup relations.")
introduced the allocation of a temporary ulist in function
btrfs_add_qgroup_relation() and added the corresponding cleanup to the out
path. However, the allocation was introduced before the src/dst level check
that directly returns. Fix the possible leakage of the ulist by moving the
allocation after the input validation. Detected by Coverity CID 1295988.

Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Reviewed-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Chris Mason <clm@fb.com>
fs/btrfs/qgroup.c

index 3d6546581bb9fd4f778f2c7536c54f7367717f53..842ff86d4ae880fb37398120c552b7eefdd42df1 100644 (file)
@@ -1115,14 +1115,14 @@ int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans,
        struct ulist *tmp;
        int ret = 0;
 
-       tmp = ulist_alloc(GFP_NOFS);
-       if (!tmp)
-               return -ENOMEM;
-
        /* Check the level of src and dst first */
        if (btrfs_qgroup_level(src) >= btrfs_qgroup_level(dst))
                return -EINVAL;
 
+       tmp = ulist_alloc(GFP_NOFS);
+       if (!tmp)
+               return -ENOMEM;
+
        mutex_lock(&fs_info->qgroup_ioctl_lock);
        quota_root = fs_info->quota_root;
        if (!quota_root) {