Re: [PATCH net] dccp/tcp: fix routing redirect race

From: Jonathan Maxwell
Date: Wed Mar 08 2017 - 23:51:27 EST


On Thu, Mar 9, 2017 at 3:40 PM, Eric Dumazet <eric.dumazet@xxxxxxxxx> wrote:
> On Thu, 2017-03-09 at 14:42 +1100, Jonathan Maxwell wrote:
>> Sorry let me resend in plain text mode.
>>
>> On Thu, Mar 9, 2017 at 1:10 PM, Eric Dumazet <eric.dumazet@xxxxxxxxx> wrote:
>> > On Thu, 2017-03-09 at 12:15 +1100, Jon Maxwell wrote:
>> >> We have seen a few incidents lately where a dst_enty has been freed
>> >> with a dangling TCP socket reference (sk->sk_dst_cache) pointing to that
>> >> dst_entry. If the conditions/timings are right a crash then ensues when the
>> >> freed dst_entry is referenced later on. A Common crashing back trace is:
>> >
>> > Very nice catch !
>> >
>>
>> Thanks Eric.
>>
>> > Don't we have a similar issue for IPv6 ?
>> >
>> >
>>
>> Good point.
>>
>> We checked and as far as we can tell IPv6 does not invalidate the route.
>> So it should be safer.
>
> Simply doing :
>
> __sk_dst_check(sk, np->dst_cookie);
>
> is racy, even before calling dst->ops->redirect(dst, sk, skb);
>
> (if socket is owned by user)
>
>
>

Okay, I will add a similar patch for IPv6 to also protect from that.