fs/proc/task_mmu.c: don't use task->mm in m_start() and show_*map()
authorOleg Nesterov <oleg@redhat.com>
Thu, 9 Oct 2014 22:25:19 +0000 (15:25 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 10 Oct 2014 02:25:48 +0000 (22:25 -0400)
commit46c298cf69d0e9a27d33ff992a81bd7b441c7933
tree5bf50f1166996eb08fd7303ba93cc8bf8c5df7d9
parentb1a8de1f534337b398c7778578a56ec4f018cb27
fs/proc/task_mmu.c: don't use task->mm in m_start() and show_*map()

get_gate_vma(priv->task->mm) looks ugly and wrong, task->mm can be NULL or
it can changed by exec right after mm_access().

And in theory this race is not harmless, the task can exec and then later
exit and free the new mm_struct.  In this case get_task_mm(oldmm) can't
help, get_gate_vma(task->mm) can read the freed/unmapped memory.

I think that priv->task should simply die and hold_task_mempolicy() logic
can be simplified.  tail_vma logic asks for cleanups too.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/proc/task_mmu.c