reiserfs_add_entry(): get rid of pointless namelen checks

In all cases namelen is ->d_name.len of some dentry; moreover, a dentry
that has passed ->lookup() without triggering ENAMETOOLONG check there.
The comment next to these checks is either a rudiment of some other
check that used to be there once upon a time, or an attempt to come up
with the possible reason for that check (well, more like "why does
ext3 do it?")

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c
index 994d6e6995ab1967ba5470c92225e366a71b10e4..c5f233b4a27f8b13f15eb923953075e70909b1c1 100644 (file)
--- a/fs/reiserfs/namei.c
+++ b/fs/reiserfs/namei.c
@@ -451,13 +451,6 @@ static int reiserfs_add_entry(struct reiserfs_transaction_handle *th,
 
        BUG_ON(!th->t_trans_id);
 
-       /* cannot allow items to be added into a busy deleted directory */
-       if (!namelen)
-               return -EINVAL;
-
-       if (namelen > REISERFS_MAX_NAME(dir->i_sb->s_blocksize))
-               return -ENAMETOOLONG;
-
        /* each entry has unique key. compose it */
        make_cpu_key(&entry_key, dir,
                     get_third_component(dir->i_sb, name, namelen),