Path: csiph.com!aioe.org!bofh.it!news.nic.it!robomod
From: Cong Wang <xiyou.wangcong@gmail.com>
Newsgroups: linux.kernel
Subject: Re: net: deadlock between ip_expire/sch_direct_xmit
Date: Tue, 14 Mar 2017 17:50:01 +0100
Message-ID: <tkXO1-2Py-13@gated-at.bofh.it>
References: <tkUQb-Mo-53@gated-at.bofh.it> <tkVVU-1uM-17@gated-at.bofh.it> <tkW5z-1Ad-3@gated-at.bofh.it> <tkW5z-1Ad-1@gated-at.bofh.it>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Sender: robomod@news.nic.it
Approved: robomod@news.nic.it
Lines: 31
Organization: linux.* mail to news gateway
Xref: csiph.com linux.kernel:1600687

On Tue, Mar 14, 2017 at 7:56 AM, Eric Dumazet <edumazet@google.com> wrote:
> On Tue, Mar 14, 2017 at 7:46 AM, Dmitry Vyukov <dvyukov@google.com> wrote:
>
>> I am confused. Lockdep has observed both of these stacks:
>>
>>        CPU0                    CPU1
>>        ----                    ----
>>   lock(&(&q->lock)->rlock);
>>                                lock(_xmit_ETHER#2);
>>                                lock(&(&q->lock)->rlock);
>>   lock(_xmit_ETHER#2);
>>
>>
>> So it somehow happened. Or what do you mean?
>>
>
> Lockdep said " possible circular locking dependency detected " .
> It is not an actual deadlock, but lockdep machinery firing.
>
> For a dead lock to happen, this would require that he ICMP message
> sent by ip_expire() is itself fragmented and reassembled.
> This cannot be, because ICMP messages are not candidates for
> fragmentation, but lockdep can not know that of course...

It doesn't have to be ICMP, as long as get the same hash for
the inet_frag_queue, we will need to take the same lock and
deadlock will happen.

        hash = ipqhashfn(iph->id, iph->saddr, iph->daddr, iph->protocol);

So it is really up to this hash function.