]> git.baikalelectronics.ru Git - kernel.git/commitdiff
Fix user struct leakage with locked IPC shem segment
authorPavel Emelianov <xemul@openvz.org>
Tue, 31 Jul 2007 07:38:48 +0000 (00:38 -0700)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Tue, 31 Jul 2007 22:39:40 +0000 (15:39 -0700)
When user locks an ipc shmem segmant with SHM_LOCK ctl and the segment is
already locked the shmem_lock() function returns 0.  After this the
subsequent code leaks the existing user struct:

== ipc/shm.c: sys_shmctl() ==
     ...
     err = shmem_lock(shp->shm_file, 1, user);
     if (!err) {
          shp->shm_perm.mode |= SHM_LOCKED;
          shp->mlock_user = user;
     }
     ...
==

Other results of this are:
1. the new shp->mlock_user is not get-ed and will point to freed
   memory when the task dies.
2. the RLIMIT_MEMLOCK is screwed on both user structs.

The exploit looks like this:

==
    id = shmget(...);
    setresuid(uid, 0, 0);
    shmctl(id, SHM_LOCK, NULL);
    setresuid(uid + 1, 0, 0);
    shmctl(id, SHM_LOCK, NULL);
==

My solution is to return 0 to the userspace and do not change the
segment's user.

Signed-off-by: Pavel Emelianov <xemul@openvz.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ipc/shm.c

index d88ac5a07ec72464266227856238964595303825..a86a3a5c8a19589698dc8b9ec555e1d98891466d 100644 (file)
--- a/ipc/shm.c
+++ b/ipc/shm.c
@@ -716,7 +716,7 @@ asmlinkage long sys_shmctl (int shmid, int cmd, struct shmid_ds __user *buf)
                        struct user_struct * user = current->user;
                        if (!is_file_hugepages(shp->shm_file)) {
                                err = shmem_lock(shp->shm_file, 1, user);
-                               if (!err{
+                               if (!err && !(shp->shm_perm.mode & SHM_LOCKED)){
                                        shp->shm_perm.mode |= SHM_LOCKED;
                                        shp->mlock_user = user;
                                }