Why is it that p->mm it's checked inside oom_kill_task and again in
__oom_kill_task? Cause __oom_kill_task it's called in case p->mm not null and
not &init_mm, otherwise it just returns NULL, this has been bothering me,
I've patched with the last rc5 and it's still there, why?
Is there a chance a process can grow an mm area between one call and another?
Vicente.