ext4: check for directory entries too close to block end
authorJan Kara <jack@suse.cz>
Mon, 2 Dec 2019 17:02:13 +0000 (18:02 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 31 Dec 2019 15:46:17 +0000 (16:46 +0100)
commit 109ba779d6cca2d519c5dd624a3276d03e21948e upstream.

ext4_check_dir_entry() currently does not catch a case when a directory
entry ends so close to the block end that the header of the next
directory entry would not fit in the remaining space. This can lead to
directory iteration code trying to access address beyond end of current
buffer head leading to oops.

CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191202170213.4761-3-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
fs/ext4/dir.c

index 9fdd2b2..6305d5e 100644 (file)
@@ -81,6 +81,11 @@ int __ext4_check_dir_entry(const char *function, unsigned int line,
                error_msg = "rec_len is too small for name_len";
        else if (unlikely(((char *) de - buf) + rlen > size))
                error_msg = "directory entry overrun";
+       else if (unlikely(((char *) de - buf) + rlen >
+                         size - EXT4_DIR_REC_LEN(1) &&
+                         ((char *) de - buf) + rlen != size)) {
+               error_msg = "directory entry too close to block end";
+       }
        else if (unlikely(le32_to_cpu(de->inode) >
                        le32_to_cpu(EXT4_SB(dir->i_sb)->s_es->s_inodes_count)))
                error_msg = "inode out of bounds";