From: Patrick McHardy Subject: Re: [bug] __nf_ct_refresh_acct(): WARNING: at lib/list_debug.c:30 __list_add+0x7d/0xad() Date: Wed, 17 Jun 2009 14:36:14 +0200 Message-ID: <4A38E33E.1050006@trash.net> References: <20090615.050449.144947903.davem@davemloft.net> <20090616091538.GA4184@elte.hu> <20090616.034752.226811527.davem@davemloft.net> <20090616105304.GA3579@elte.hu> <20090616122415.GA16630@elte.hu> <20090617092152.GA17449@elte.hu> <4A38C2F3.3000009@gmail.com> <4A38D5BD.2040502@trash.net> <4A38D9BE.3020403@gmail.com> <4A38DAC4.2050902@trash.net> <4A38E2AE.3030106@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Ingo Molnar , David Miller , Thomas Gleixner , torvalds@linux-foundation.org, akpm@linux-foundation.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: Eric Dumazet Return-path: Received: from stinky.trash.net ([213.144.137.162]:33115 "EHLO stinky.trash.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756058AbZFQMgP (ORCPT ); Wed, 17 Jun 2009 08:36:15 -0400 In-Reply-To: <4A38E2AE.3030106@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: Eric Dumazet wrote: > Patrick McHardy a =E9crit : >> Eric Dumazet wrote: >>> Patrick McHardy a =E9crit : >>>> Before the conntrack is confirmed, it is exclusively handled by a >>>> single CPU. I agree that we need to make sure the IPS_CONFIRMED_BI= T >>>> is visible before we add the conntrack to the hash table since the >>>> lookup is lockless, but simply moving the set_bit before the hash >>>> insertion should be fine I think. >>>> >>> >>> Problem is timeout.expires is either a relative or absolute timeout= , >>> and changes happen >>> in __nf_conntrack_confirm() or __nf_ct_refresh_acct(). >>> >>> We must have a synchronization (an barriers), a single bit wont be >>> enough. >> Please have a look at the second patch I just sent. It relies >> on the RCU barriers to make sure all stores are visible before >> other CPUs can find the conntrack. >> >=20 > Sorry, I dont understand how your second patch corrects the problem. >=20 > This (unconfirmed) conntrack is visible by another cpu. No, before it is confirmed, its only visible to the CPU handling the initial packet of a connection. Confirmation is the step that makes it visible to other CPUs. > This other > cpu can call __nf_ct_refresh_acct() while this cpu runs > in __nf_conntrack_confirm() Not for the same conntrack, that would be a seperate bug. Does that explain what I'm trying to do? :) >=20 > @@ -425,7 +425,6 @@ __nf_conntrack_confirm(struct sk_buff *skb) > /* Remove from unconfirmed list */ > hlist_nulls_del_rcu(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode); > =20 > - __nf_conntrack_hash_insert(ct, hash, repl_hash); > /* Timer relative to confirmation time, not original > setting time, otherwise we'd get timer wrap in > weird delay cases. */ > @@ -433,8 +432,15 @@ __nf_conntrack_confirm(struct sk_buff *skb) > add_timer(&ct->timeout); >=20 > <<<< another cpu could here change timeout.expires (thinking its stil= l relative) >>>> >=20 > atomic_inc(&ct->ct_general.use); > set_bit(IPS_CONFIRMED_BIT, &ct->status); > + > + /* Since the lookup is lockless, hash insertion must be after start= ing the > + * timer and setting the CONFIRMED bit. The RCU barriers guarantee = that no > + * other CPU can find the conntrack before the above stores are vis= ible. > + */ > + __nf_conntrack_hash_insert(ct, hash, repl_hash); > NF_CT_STAT_INC(net, insert); > spin_unlock_bh(&nf_conntrack_lock); > help =3D nfct_help(ct); > if (help && help->helper) > nf_conntrack_event_cache(IPCT_HELPER, ct); >=20