Re: [PATCH v3] ptrace: Fix fork event messages across pid namespaces

[Matthew Dempsky]

On Wed, Apr 2, 2014 at 7:58 AM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> On 04/01, Matthew Dempsky wrote:
>>
>> @@ -1605,10 +1605,12 @@ long do_fork(unsigned long clone_flags,
>>        */
>>       if (!IS_ERR(p)) {
>>               struct completion vfork;
>> +             struct pid *pid;
>>
>>               trace_sched_process_fork(current, p);
>>
>> -             nr = task_pid_vnr(p);
>> +             pid = get_task_pid(p, PIDTYPE_PID);
>
> So you decided to use get_pid/put_pid ;) Honestly, I'd prefer to just
> calculate "pid_t trace_pid" before wake_up_new_task(), but I won't
> argue. Plus this way the race window becomes really small, OK.

I was leaning towards that, but then the conditions for trying to
avoid computing the pid_t became complex and I was worried that
waiting for the vfork child to finish could make the race window
arbitrarily large.  Holding a struct pid reference for the duration of
fork seemed like the easiest fix to both of those.

>> +             if (unlikely(trace)) {
>> +                     /*
>> +                      * We want to report the child's pid as seen from the
>> +                      * tracer's pid namespace.
>> +                      * FIXME: We still risk sending a bogus event message if
>> +                      * debuggers from different pid namespaces detach and
>> +                      * reattach between rcu_read_unlock() and ptrace_stop().
>> +                      */
>> +                     unsigned long message;
>> +                     rcu_read_lock();
>> +                     message = pid_nr_ns(pid,
>> +                             task_active_pid_ns(current->parent));
>> +                     rcu_read_unlock();
>> +                     ptrace_event(trace, message);
>> +             }
>>
>>               if (clone_flags & CLONE_VFORK) {
>> -                     if (!wait_for_vfork_done(p, &vfork))
>> -                             ptrace_event(PTRACE_EVENT_VFORK_DONE, nr);
>> +                     if (!wait_for_vfork_done(p, &vfork)) {
>> +                             /* See comment above about pid namespaces. */
>> +                             unsigned long message;
>> +                             rcu_read_lock();
>> +                             message = pid_nr_ns(pid,
>> +                                     task_active_pid_ns(current->parent));
>> +                             rcu_read_unlock();
>> +                             ptrace_event(PTRACE_EVENT_VFORK_DONE, message);
>> +                     }
>
> OK, but may I suggest you to make a helper? Note that the code under
> "if (trace)" and "if (CLONE_VFORK)" is the same. Even the comment above
> equally applies to the CLONE_VFORK branch.

Sure.

> Especially because this code needs a fix. Yes, rcu_read_lock() should
> be enough to ensure that ->parent and its namespace (if !NULL) can not
> go away, but task_active_pid_ns() can return NULL release_task(->parent)
> was already (although this race is pure theoretical). So this helper
> should also check it is !NULL under rcu_read_lock(), afaics.

Does this look right?

    static inline void ptrace_event_pid(int event, struct pid *pid)
    {
        unsigned long message = -1;
        struct pid_namespace *ns;

        rcu_read_lock();
        ns = task_active_pid_ns(rcu_dereference(current->parent));
        if (ns)
            message = pid_nr_ns(pid, ns);
        rcu_read_unlock();

        ptrace_event(event, message);
    }

I'm unsure if the rcu_dereference() is appropriate.  It seems like it
is based on my reading of the RCU documentation and that parent and
real_parent have been marked __rcu since 2011, but they prevailingly
seem to be accessed/mutated without the RCU APIs.

Also, to ensure I understand the race: the issue is that if the parent
were to call do_exit() concurrently with the above RCU critical
section, that parent's call to forget_original_parent() might not yet
be visible when the above code evaluates "current->parent", but a
later call to release_task() (e.g., if autoreap is true in
exit_notify) could detach the task's pids without any intervening
synchronize_rcu() call?

If so, why isn't the fix to have forget_original_parent() call
synchronize_rcu() before returning?  (And probably to use
rcu_assign_pointer() to updater t->real_parent and t->parent.)

Otherwise, it looks like (e.g.) the attempts to get the parent's pid
in fill_prstatus() and tomoyo_sys_getppid() are also theoretical races
of the same kind?

> And I forgot to mention, please send v5 to akpm. We usually route ptrace
> patches via -mm tree.

Will do.

Thanks for being patient with my locking questions! :)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/