kernfs: fix off by one error.
authorRichard Cochran <richardcochran@gmail.com>
Wed, 5 Mar 2014 16:10:52 +0000 (17:10 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 27 Apr 2014 00:19:03 +0000 (17:19 -0700)
commit 88391d49abb7d8dee91d405f96bd9e003cb6798d upstream.

The hash values 0 and 1 are reserved for magic directory entries, but
the code only prevents names hashing to 0. This patch fixes the test
to also prevent hash value 1.

Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/kernfs/dir.c

index bd6e18b..39c0143 100644 (file)
@@ -37,7 +37,7 @@ static unsigned int kernfs_name_hash(const char *name, const void *ns)
        hash = (end_name_hash(hash) ^ hash_ptr((void *)ns, 31));
        hash &= 0x7fffffffU;
        /* Reserve hash numbers 0, 1 and INT_MAX for magic directory entries */
-       if (hash < 1)
+       if (hash < 2)
                hash += 2;
        if (hash >= INT_MAX)
                hash = INT_MAX - 1;