ext4: make sure quota gets properly shutdown on error
authorJan Kara <jack@suse.cz>
Thu, 7 Oct 2021 15:53:35 +0000 (17:53 +0200)
committerTheodore Ts'o <tytso@mit.edu>
Mon, 10 Jan 2022 18:25:54 +0000 (13:25 -0500)
When we hit an error when enabling quotas and setting inode flags, we do
not properly shutdown quota subsystem despite returning error from
Q_QUOTAON quotactl. This can lead to some odd situations like kernel
using quota file while it is still writeable for userspace. Make sure we
properly cleanup the quota subsystem in case of error.

Signed-off-by: Jan Kara <jack@suse.cz>
Cc: stable@kernel.org
Link: https://lore.kernel.org/r/20211007155336.12493-2-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
fs/ext4/super.c

index b72f8f6084e48b88c03374fbf9be38e80a898dfc..863a3eae505a3e245c9c828eb9defbca48ba6869 100644 (file)
@@ -6749,10 +6749,7 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
 
        lockdep_set_quota_inode(path->dentry->d_inode, I_DATA_SEM_QUOTA);
        err = dquot_quota_on(sb, type, format_id, path);
-       if (err) {
-               lockdep_set_quota_inode(path->dentry->d_inode,
-                                            I_DATA_SEM_NORMAL);
-       } else {
+       if (!err) {
                struct inode *inode = d_inode(path->dentry);
                handle_t *handle;
 
@@ -6772,7 +6769,12 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
                ext4_journal_stop(handle);
        unlock_inode:
                inode_unlock(inode);
+               if (err)
+                       dquot_quota_off(sb, type);
        }
+       if (err)
+               lockdep_set_quota_inode(path->dentry->d_inode,
+                                            I_DATA_SEM_NORMAL);
        return err;
 }