Re: [PATCH] seccomp: do not reject initial filter using TSYNC

[Kees Cook]

On Thu, Jul 17, 2014 at 10:59 AM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote:
> On Thu, Jul 17, 2014 at 10:56 AM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>> On Thu, Jul 17, 2014 at 10:53 AM, Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote:
>>> On Thu, Jul 17, 2014 at 10:49 AM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>>>> There was an unneeded sanity check in the TSYNC code that was causing
>>>> the first filter applied to not allow the TSYNC flag. Additionally,
>>>> this optimizes the thread loops to skip "current". It was harmless, but
>>>> better to not cause problems in the future.
>>>>
>>>> Reported-by: David Drysdale <drysdale@xxxxxxxxxx>
>>>> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
>>>> ---
>>>> This goes on top of the v11 seccomp-tsync series. If I should respin
>>>> as v12, please let me know.
>>>>
>>>> Thanks!
>>>> ---
>>>>  kernel/seccomp.c |   11 ++++++++---
>>>>  1 file changed, 8 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/kernel/seccomp.c b/kernel/seccomp.c
>>>> index 2125b83ccfd4..0e0c6905b81d 100644
>>>> --- a/kernel/seccomp.c
>>>> +++ b/kernel/seccomp.c
>>>> @@ -255,14 +255,15 @@ static inline pid_t seccomp_can_sync_threads(void)
>>>>         BUG_ON(!mutex_is_locked(&current->signal->cred_guard_mutex));
>>>>         BUG_ON(!spin_is_locked(&current->sighand->siglock));
>>>>
>>>> -       if (current->seccomp.mode != SECCOMP_MODE_FILTER)
>>>> -               return -EACCES;
>>>> -
>>>>         /* Validate all threads being eligible for synchronization. */
>>>>         caller = current;
>>>>         for_each_thread(caller, thread) {
>>>>                 pid_t failed;
>>>>
>>>> +               /* Skip current, since it is initiating the sync. */
>>>> +               if (thread == current)
>>>> +                       continue;
>>>> +
>>>
>>> Should that be "thread == caller"?
>>
>> caller shouldn't be changing, correct? Won't it be the same?
>>
>
> I assumed that you loaded caller once as an optimization -- ISTR that,
> at least at some point, accessing current was a slightly expensive.
> Maybe this is moot now.
>
> Anyway, the rest of the code in there is comparing thread to caller,
> using caller seems a bit more consistent.

Fair enough. :)

I'll resend with this folded in for a v12.

Thanks!

-Kees

-- 
Kees Cook
Chrome OS Security
--
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/