isofs: Fix unchecked printing of ER records
authorJan Kara <jack@suse.cz>
Thu, 18 Dec 2014 16:26:10 +0000 (17:26 +0100)
committerBen Hutchings <ben@decadent.org.uk>
Fri, 20 Feb 2015 00:49:32 +0000 (00:49 +0000)
commit 4e2024624e678f0ebb916e6192bd23c1f9fdf696 upstream.

We didn't check length of rock ridge ER records before printing them.
Thus corrupted isofs image can cause us to access and print some memory
behind the buffer with obvious consequences.

Reported-and-tested-by: Carl Henrik Lunde <chlunde@ping.uio.no>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
fs/isofs/rock.c

index 26859de..1780949 100644 (file)
@@ -362,6 +362,9 @@ repeat:
                        rs.cont_size = isonum_733(rr->u.CE.size);
                        break;
                case SIG('E', 'R'):
+                       /* Invalid length of ER tag id? */
+                       if (rr->u.ER.len_id + offsetof(struct rock_ridge, u.ER.data) > rr->len)
+                               goto out;
                        ISOFS_SB(inode->i_sb)->s_rock = 1;
                        printk(KERN_DEBUG "ISO 9660 Extensions: ");
                        {