ipc,shm: fix shm_file deletion races
authorGreg Thelen <gthelen@google.com>
Thu, 21 Nov 2013 22:32:00 +0000 (14:32 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 22 Nov 2013 00:42:27 +0000 (16:42 -0800)
commita399b29dfbaaaf91162b2dc5a5875dd51bbfa2a1
treedfff401b36b593b284cd21e4051ad7a23a944845
parent30b0a105d9f7141e4cbf72ae5511832457d89788
ipc,shm: fix shm_file deletion races

When IPC_RMID races with other shm operations there's potential for
use-after-free of the shm object's associated file (shm_file).

Here's the race before this patch:

  TASK 1                     TASK 2
  ------                     ------
  shm_rmid()
    ipc_lock_object()
                             shmctl()
                             shp = shm_obtain_object_check()

    shm_destroy()
      shum_unlock()
      fput(shp->shm_file)
                             ipc_lock_object()
                             shmem_lock(shp->shm_file)
                             <OOPS>

The oops is caused because shm_destroy() calls fput() after dropping the
ipc_lock.  fput() clears the file's f_inode, f_path.dentry, and
f_path.mnt, which causes various NULL pointer references in task 2.  I
reliably see the oops in task 2 if with shmlock, shmu

This patch fixes the races by:
1) set shm_file=NULL in shm_destroy() while holding ipc_object_lock().
2) modify at risk operations to check shm_file while holding
   ipc_object_lock().

Example workloads, which each trigger oops...

Workload 1:
  while true; do
    id=$(shmget 1 4096)
    shm_rmid $id &
    shmlock $id &
    wait
  done

  The oops stack shows accessing NULL f_inode due to racing fput:
    _raw_spin_lock
    shmem_lock
    SyS_shmctl

Workload 2:
  while true; do
    id=$(shmget 1 4096)
    shmat $id 4096 &
    shm_rmid $id &
    wait
  done

  The oops stack is similar to workload 1 due to NULL f_inode:
    touch_atime
    shmem_mmap
    shm_mmap
    mmap_region
    do_mmap_pgoff
    do_shmat
    SyS_shmat

Workload 3:
  while true; do
    id=$(shmget 1 4096)
    shmlock $id
    shm_rmid $id &
    shmunlock $id &
    wait
  done

  The oops stack shows second fput tripping on an NULL f_inode.  The
  first fput() completed via from shm_destroy(), but a racing thread did
  a get_file() and queued this fput():
    locks_remove_flock
    __fput
    ____fput
    task_work_run
    do_notify_resume
    int_signal

Fixes: c2c737a0461e ("ipc,shm: shorten critical region for shmat")
Fixes: 2caacaa82a51 ("ipc,shm: shorten critical region for shmctl")
Signed-off-by: Greg Thelen <gthelen@google.com>
Cc: Davidlohr Bueso <davidlohr@hp.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: <stable@vger.kernel.org> # 3.10.17+ 3.11.6+
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ipc/shm.c