public inbox for netdev@vger.kernel.org 
 help / color / mirror / Atom feed
From: Patrick McHardy <kaber@trash•net>
To: Changli Gao <xiaosuo@gmail•com>
Cc: hawk@comx•dk, Eric Dumazet <eric.dumazet@gmail•com>,
	Linux Kernel Network Hackers <netdev@vger•kernel.org>,
	netfilter-devel@vger•kernel.org,
	Paul E McKenney <paulmck@linux•vnet.ibm.com>
Subject: Re: DDoS attack causing bad effect on conntrack searches
Date: Thu, 22 Apr 2010 15:17:40 +0200	[thread overview]
Message-ID: <4BD04C74.9020402@trash.net> (raw)
In-Reply-To: <q2h412e6f7f1004220613m488c2ee4r6d24a8d1e65997d4@mail.gmail.com>

Changli Gao wrote:
> On Thu, Apr 22, 2010 at 8:58 PM, Jesper Dangaard Brouer <hawk@comx•dk> wrote:
>> At an unnamed ISP, we experienced a DDoS attack against one of our
>> customers.  This attack also caused problems for one of our Linux
>> based routers.
>>
>> The attack was "only" generating 300 kpps (packets per sec), which
>> usually isn't a problem for this (fairly old) Linux Router.  But the
>> conntracking system chocked and reduced pps processing power to
>> 40kpps.
>>
>> I do extensive RRD/graph monitoring of the machines.  The IP conntrack
>> searches in the period exploded, to a stunning 700.000 searches per
>> sec.
>>
>> http://people.netfilter.org/hawk/DDoS/2010-04-12__001/conntrack_searches001.png
>>
>> First I though it might be caused by bad hashing, but after reading
>> the kernel code (func: __nf_conntrack_find()), I think its caused by
>> the loop restart (goto begin) of the conntrack search, running under
>> local_bh_disable().  These RCU changes to conntrack were introduced in
>> ea781f19 by Eric Dumazet.
>>
>> Code: net/netfilter/nf_conntrack_core.c
>> Func: __nf_conntrack_find()
>>
>> struct nf_conntrack_tuple_hash *
>> __nf_conntrack_find(struct net *net, const struct nf_conntrack_tuple *tuple)
>> {
>>        struct nf_conntrack_tuple_hash *h;
>>        struct hlist_nulls_node *n;
>>        unsigned int hash = hash_conntrack(tuple);
>>
>>        /* Disable BHs the entire time since we normally need to disable them
>>         * at least once for the stats anyway.
>>         */
>>        local_bh_disable();
>> begin:
>>        hlist_nulls_for_each_entry_rcu(h, n, &net->ct.hash[hash], hnnode) {
>>                if (nf_ct_tuple_equal(tuple, &h->tuple)) {
>>                        NF_CT_STAT_INC(net, found);
>>                        local_bh_enable();
>>                        return h;
>>                }
>>                NF_CT_STAT_INC(net, searched);
>>        }
>>        /*
>>         * if the nulls value we got at the end of this lookup is
>>         * not the expected one, we must restart lookup.
>>         * We probably met an item that was moved to another chain.
>>         */
>>        if (get_nulls_value(n) != hash)
>>                goto begin;
>>        local_bh_enable();
>>
> 
> We should add a retry limit there.

We can't do that since that would allow false negatives.

  reply	other threads:[~2010-04-22 13:17 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-22 12:58 DDoS attack causing bad effect on conntrack searches Jesper Dangaard Brouer
2010-04-22 13:13 ` Changli Gao
2010-04-22 13:17   ` Patrick McHardy [this message]
2010-04-22 14:36     ` Eric Dumazet
2010-04-22 14:53       ` Eric Dumazet
2010-04-22 15:51         ` Paul E. McKenney
2010-04-22 16:02           ` Eric Dumazet
2010-04-22 16:34             ` Paul E. McKenney
2010-04-22 20:38             ` Jesper Dangaard Brouer
2010-04-22 21:03               ` Eric Dumazet
2010-04-22 21:14                 ` Eric Dumazet
2010-04-22 23:44                   ` David Miller
2010-04-23  5:44                     ` Eric Dumazet
2010-04-23  8:13                       ` David Miller
2010-04-23  8:18                         ` David Miller
2010-04-23  8:40                           ` Jesper Dangaard Brouer
2010-04-23 10:36                   ` Patrick McHardy
2010-04-23 11:06                     ` Eric Dumazet
2010-04-22 21:28                 ` Jesper Dangaard Brouer
2010-04-23  7:23                   ` Jan Engelhardt
2010-04-23  7:46                     ` Eric Dumazet
2010-04-23  7:55                       ` Jan Engelhardt
2010-04-23  9:23                         ` Eric Dumazet
2010-04-23 10:55                 ` Patrick McHardy
2010-04-23 11:05                   ` Eric Dumazet
2010-04-23 11:06                     ` Patrick McHardy
2010-04-23 20:57               ` Eric Dumazet
2010-04-24 11:11                 ` Jesper Dangaard Brouer
2010-04-24 20:11                   ` Eric Dumazet
2010-04-26 14:36                     ` Jesper Dangaard Brouer
2010-05-31 21:21                       ` Eric Dumazet
2010-06-01  0:28                         ` Changli Gao
2010-06-01  5:05                           ` Eric Dumazet
2010-06-01  5:48                             ` Changli Gao
2010-06-01 10:18                             ` Patrick McHardy
2010-06-01 10:31                               ` Eric Dumazet
2010-06-01 10:41                                 ` Patrick McHardy
2010-06-01 16:20                                   ` [RFC nf-next-2.6] conntrack: per cpu nf_conntrack_untracked Eric Dumazet
2010-06-04 11:40                                     ` Patrick McHardy
2010-06-04 12:10                                       ` Changli Gao
2010-06-04 12:29                                         ` Patrick McHardy
2010-06-04 12:36                                           ` Eric Dumazet
2010-06-04 16:25                                             ` [PATCH nf-next-2.6] conntrack: IPS_UNTRACKED bit Eric Dumazet
2010-06-04 20:15                                               ` [PATCH nf-next-2.6 2/2] conntrack: per_cpu untracking Eric Dumazet
2010-06-08 14:29                                                 ` Patrick McHardy
2010-06-08 14:52                                                   ` Eric Dumazet
2010-06-08 15:12                                                     ` Eric Dumazet
2010-06-09 12:45                                                       ` Patrick McHardy
2010-06-08 14:12                                               ` [PATCH nf-next-2.6] conntrack: IPS_UNTRACKED bit Patrick McHardy
2010-04-23 10:56       ` DDoS attack causing bad effect on conntrack searches Patrick McHardy
2010-04-23 12:45         ` Jesper Dangaard Brouer
2010-04-23 13:57           ` Patrick McHardy
2010-04-22 13:31   ` Jesper Dangaard Brouer
2010-04-23 10:35     ` Patrick McHardy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4BD04C74.9020402@trash.net \
    --to=kaber@trash$(echo .)net \
    --cc=eric.dumazet@gmail$(echo .)com \
    --cc=hawk@comx$(echo .)dk \
    --cc=netdev@vger$(echo .)kernel.org \
    --cc=netfilter-devel@vger$(echo .)kernel.org \
    --cc=paulmck@linux$(echo .)vnet.ibm.com \
    --cc=xiaosuo@gmail$(echo .)com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox